Ffmpeg-devel-irc

Download

ffmpeg-devel-irc@ffmpeg.org

February 2016

1 participants
58 discussions

ffmpeg-devel.log.20160213
by burek 14 Feb '16

14 Feb '16

[01:18:23 CET] <Timothy_Gu> TIL sed \| is a GNU extension [01:23:12 CET] <cone-518> ffmpeg 03Mats Peterson 07master:bbb45773b043: avcodec/rawdec: Initialize default mono palette only for bits_per_coded_sample == 1 [01:23:12 CET] <cone-518> ffmpeg 03Mats Peterson 07master:efcea7799685: avcodec/rawdec: Fix nut pal8 test [01:23:12 CET] <cone-518> ffmpeg 03Mats Peterson 07master:4e5db87174ed: avcodec/rawdec: Print stride and packet size at debug level [02:16:56 CET] <cone-518> ffmpeg 03Michael Niedermayer 07master:588e2e338717: avcodec/h264: Fix memleak of a53_caption [12:20:24 CET] Action: wm4 stares at the thread with 90% mats posts [12:33:30 CET] <nevcairiel> someone should rate limit him on the ML [12:33:32 CET] <nevcairiel> one per day or so [13:21:14 CET] <durandal_1707> this guy needs some help [13:25:31 CET] <cone-438> ffmpeg 03Rostislav Pehlivanov 07master:0fe0e213c042: aacenc: temporarily disable Mid/Side coding with multichannel files [13:34:02 CET] <durandal_1707> i will just make gmail to move his mails to spam [13:36:32 CET] <wm4> hm a snide reply on my post [13:36:35 CET] <wm4> into the spam filter he goes [13:58:48 CET] <nevcairiel> remember that his first post to the community was to call us all assholes on trac [14:00:16 CET] <nevcairiel> noone should have ever engaged him [14:22:50 CET] <rcombs> to be fair, we _are_ all assholes [14:23:06 CET] <wm4> D: [14:23:47 CET] <rcombs> it's a necessary defense mechanism for dealing with media bullshit [14:27:40 CET] <fritsch> we are all sitting in shit, but in different deepness [14:27:44 CET] <fritsch> :-) [14:45:49 CET] <BBB> is this about mats? [14:46:16 CET] <BBB> hes been asked several times to not post 10x/min in a single thread& I dont think thats an unreasonable request [14:48:35 CET] <nevcairiel> he has serious self control issues as well, everytime you say anything he snaps at you and apologizes 5 mails later [14:49:32 CET] <rcombs> holy shit, reading through this stuff [14:49:39 CET] <JEEB> don't [14:49:49 CET] <JEEB> it's not fun and it's just a sad piece of text [14:49:58 CET] <rcombs> someone should write a script [14:50:17 CET] <nevcairiel> i wish i could setup digest delivery for specific people only =p [14:50:20 CET] <rcombs> when he sends a mail, don't immediately accept it but start a timer for 10 minutes or something [14:50:42 CET] <rcombs> coalesce all his mails on that thread, resetting the timer period on each one [14:50:50 CET] <rcombs> when the timer expires, send as a single message [14:50:59 CET] <rcombs> like digest but for input [14:52:35 CET] <BBB> I want the ML software to do that for me [14:52:46 CET] <BBB> not each client should have to do that @ output manually [14:52:58 CET] <rcombs> yeah, that's what I mean [14:56:23 CET] <BBB> ah I see, sorry [14:56:28 CET] <rcombs> np [14:56:31 CET] <BBB> yes that would be amazing and productivity-enhancing [14:58:46 CET] <nevcairiel> i just ignore all these threads entirely [14:58:55 CET] <nevcairiel> its also productivity-enhancing and sanity-preserving [14:59:45 CET] <JEEB> true that [15:00:29 CET] <J_Darnley> Crap. I wrote a comment saying "the 'Q' command in sed is a GNU extension". Useful(!) I can't remember *why* I was using it. [15:05:59 CET] <J_Darnley> Oh "This command is the same as `q', but will not print the contents of pattern space." (What does q do again?) [15:06:04 CET] <J_Darnley> :) [15:06:31 CET] <J_Darnley> Oh it exits [16:17:01 CET] <cone-438> ffmpeg 03Michael Niedermayer 07master:50a960673db7: doc/APIchanges: Fill in some missing values [16:17:47 CET] <FearThe1337> I think I noticed a typo in the source code, could anyone verify to make sure it's not just me thinking wrongly? [16:20:15 CET] <rcombs> that would not be surprising; post the line in question? [16:20:15 CET] <FearThe1337> https://github.com/FFmpeg/FFmpeg/blob/master/libavdevice/dshow.c#L242-L244 [16:20:25 CET] <FearThe1337> link above ;) [16:20:36 CET] <rcombs> yup, quite right [16:21:01 CET] <FearThe1337> Should most likely be !=, since if (a = 0) is always false [16:21:56 CET] <rcombs> yup [16:22:11 CET] <rcombs> should be a compiler warning [16:22:41 CET] <rcombs> patch welcome [16:23:08 CET] <FearThe1337> Hmm yes, decided to try and compile it myself and noticed that as one of the million warnings [16:23:55 CET] <rcombs> (patches welcome for the other ones too) [16:35:13 CET] <cone-438> ffmpeg 03Michael Niedermayer 07master:dab493c89ec3: doc/APIchanges: fill in more missing things [16:35:14 CET] <cone-438> ffmpeg 03Michael Niedermayer 07master:d18df8beb0a0: avutil/imgutils: do not ignore align parameter for pseudo-pal in av_image_get_buffer_size() [16:38:24 CET] <Daemon404> omg the mats patches finally went in [16:52:57 CET] <cone-438> ffmpeg 03Michael Niedermayer 07master:a9b81bfd1c7a: avutil/avstring: add a "ALL" entry and the possibility to negate matches to av_match_name() [16:59:09 CET] <cone-438> ffmpeg 03Timothy Gu 07master:94a1c7491bb3: fate/source: Do not use GNU extensions in sed [17:12:52 CET] <BBB> are the tech specs for ffv1 finished? [17:12:57 CET] <BBB> or are they still being worked on? [17:40:54 CET] <kierank> BBB: still going [17:41:33 CET] <kierank> BBB: https://mailarchive.ietf.org/arch/search/?email_list=cellar [17:55:37 CET] <michaelni> BBB, can the vp9 & hwaccel-mt patches be applied or is there a reason to wait ? [17:55:50 CET] <BBB> no reason to wait, can be applied [17:55:59 CET] <BBB> sorry, I havent kept track very closely of that patch [17:56:09 CET] <BBB> I can apply if you want, but its probably faster if you merge for me [17:56:40 CET] <BBB> I believe with that patch, we can revert the previous patch that prevented mt+hwaccel, as andreas requested [17:56:56 CET] <BBB> it will still not work for dxva2, but thats unrelated to vp9 since it fails for h264 also [17:59:10 CET] <michaelni> i can apply the patchs, i just dont want to step on anyones toes, the vp9 one seems clear but which is the 2nd one, the one from anreas that changes it to a warning or revert 31741ae ? [19:01:09 CET] <Daemon404> kierank, "tech free saturday" failed i see [19:01:27 CET] <kierank> watching skyfall [19:01:29 CET] <kierank> and playing on my phone [19:52:12 CET] <cone-438> ffmpeg 03Ronald S. Bultje 07master:5307adadeb22: vp9: only call ff_get_format on stream format changes. [19:52:13 CET] <cone-438> ffmpeg 03Michael Niedermayer 07master:4c920ce777a1: avutil/imgutils: only align the palette in av_image_copy_to_buffer() if there is enough space [21:08:25 CET] <cone-438> ffmpeg 03Michael Niedermayer 07master:4099e4a77d2d: avutil/frame: Free destination qp_table_buf in frame_copy_props() [22:07:00 CET] <JEEB> mateo`: should this be enough during app initialization? http://up-cat.net/p/fc18d971 [22:07:43 CET] <JEEB> because even with that I'm still getting http://up-cat.net/p/d5afaa7c [22:51:08 CET] <JEEB> mateo`: ok, got it to work by moving the init inside libmpv [23:58:08 CET] <xyz> and then moving it back to where it belongs [23:58:23 CET] <JEEB> yup [23:59:03 CET] <xyz> I'll post my error here too I guess [23:59:39 CET] <xyz> mateo`: getting spammed with "E/C2DColorConvert( 339): unknown format passed for luma alignment number" when using your mediacodec branch [00:00:00 CET] --- Sun Feb 14 2016

1 0

ffmpeg.log.20160213
by burek 14 Feb '16

14 Feb '16

[00:00:08 CET] <J_Darnley> Let ffmpeg detect your cpu features. [00:00:14 CET] <J_Darnley> It knows better then you. [00:00:34 CET] <J_Darnley> For example it knows intels have mmx, sse, avx, not the arm crap. [00:01:18 CET] <J_Darnley> Back to your first question: if it was x86 then I would immediately say "64 bit is faster" [00:17:56 CET] <mattf000> thanks.. but back to the 2nd one quick... it's not 100% clear to me that ffmpeg knows it is sending this video to an older android device [00:18:48 CET] <mattf000> of course it always notes the cpu flags it leverages when i run the program ... i'm just not sure if that is more about helping transcode faster ... but would it negatively impact the decoding time on the droid? [00:18:53 CET] <J_Darnley> So what? [00:19:02 CET] <mattf000> i only ask because i'm trying to trim latency right now... i'm at about 180 [00:19:04 CET] <mattf000> ms [00:19:48 CET] <J_Darnley> I don't understand your thought process at all. [00:20:49 CET] <J_Darnley> The hardware that matters is the one that is running ffmpeg. [00:21:20 CET] <J_Darnley> The decoder could be a set-top box, bluray player, mobile phone, desktop computer. [00:21:32 CET] <J_Darnley> ffmpeg doesn't know and doesn't care [00:21:33 CET] <mattf000> i guess i just have more confidence in the quad core intel to do its job [00:21:48 CET] <mattf000> i'm more worried about the armv7a android being used by a surgeon [00:22:19 CET] <mattf000> obviously any latency is an issue in this application [00:22:54 CET] <mattf000> the server will be in the room with him but it's going over 2.4gz 802.11nba[ [00:23:27 CET] <J_Darnley> I am still lost. [00:23:28 CET] <mattf000> it wasn't hard getting to sub 200ms but getting it lower now is way harder than i imagined [00:23:40 CET] <J_Darnley> Are you running ffmpeg on the mobile device or the server? [00:23:58 CET] <t4nk533> Hi. I can't find nowhere how to extract one image per FIELD from an interlaced video steam? Normally I get a single image per frame, with two fields merged (at least no deinterlacing is used) [00:24:09 CET] <mattf000> server .. i just didn't know if there were ways to make the decoding easier on the client [00:24:23 CET] <J_Darnley> Are you using libx264? [00:24:31 CET] <mattf000> on the server but not the client [00:24:43 CET] <J_Darnley> (of course. it is not a decoder) [00:24:43 CET] <mattf000> client has hw decoding tho via flash [00:25:16 CET] <J_Darnley> I will point you to the fastdecode and zerolatency tunings [00:26:00 CET] <mattf000> yea.. got that [00:26:34 CET] <mattf000> -tune zerolatency,fastdecode [00:26:36 CET] <furq> t4nk533: -vf separatefields [00:26:58 CET] <furq> https://ffmpeg.org/ffmpeg-filters.html#separatefields [00:27:39 CET] <mattf000> this would be better if i were using udp ... is there reliable way for libx264 to tell a client that it needs to throw out frames that are falling beghind? [00:27:39 CET] <t4nk533> Thanks furq. Will that give me images of half height? [00:27:44 CET] <furq> yes [00:28:43 CET] <furq> mattf000: using udp over wifi sounds like a bad idea [00:29:01 CET] <furq> especially if surgery is involved [00:29:11 CET] <J_Darnley> yes libx264 has features for that [00:29:23 CET] <J_Darnley> I don't think they are exposed through libavcodec though [00:29:25 CET] <mattf000> do you recommend a transport over RTMP? (which is what i'm using) [00:29:58 CET] <furq> i recommend it for streaming videos over the internet, which is about where my knowledge ends [00:30:32 CET] <derekprestegard> hey guys - any advice on converting openexr files into an 4:2:0 8bpc video? [00:30:52 CET] <J_Darnley> ffmpeg -i INPUT OUTPUT? [00:30:53 CET] <derekprestegard> simple -i file.exr -pix_fmt yuv420p out.mp4 works, but the image is extremely dark [00:31:05 CET] <mattf000> i'm basically going from a capture card to smart glasses over wifi [00:31:31 CET] <J_Darnley> derekprestegard: perhaps you need to correct gamma so something [00:31:32 CET] <mattf000> the glasses are locked down on 4.0.4 tho [00:31:34 CET] <derekprestegard> the openexr format is huge - either 16 or 32 bpc, so I probably need to somehow tell ffmpeg how to do the conversion [00:31:39 CET] <derekprestegard> J_Darnley: yeah thats what I was thinking [00:32:02 CET] <t4nk533> Why are the fields merged by default though? As far as I know, an interlaced video stream is basically encoded in the same way as a progressive one, the fields are handled as frames with half height? [00:34:56 CET] <t4nk533> Or not? furq? [00:35:43 CET] <Guiri> I'm trying to troubleshoot a libfdk-aac error message: http://fpaste.org/322056/20118145/. I did a build on Ubuntu 14.04 following these instructions: https://trac.ffmpeg.org/wiki/CompilationGuide/Ubuntu [00:36:14 CET] <J_Darnley> What's so hard to understand about "No such file or directory"? [00:36:41 CET] <Guiri> Because I compiled the libfdk_aac library in ~/ffmpeg_sources and there were no errors on make or make install [00:36:58 CET] <J_Darnley> Does your system know to find libraries in that directory? [00:37:29 CET] <J_Darnley> try LD_LIBRARY_PATH or stop building shared libraries [00:38:04 CET] <Guiri> `./configure --prefix="$HOME/ffmpeg_build" --disable-shared` should've disabled the shared library build [00:38:45 CET] <J_Darnley> Then Idon't know how you managed to link with a shared aac library. [00:39:49 CET] <J_Darnley> as I said "try LD_LIBRARY_PATH" [00:50:37 CET] <c_14> --disable-shared doesn't disable linking against shared libraries [00:50:51 CET] <c_14> It just disables building of libav* shared libraries [00:51:15 CET] <c_14> So configure probably found a shared libfdk_aac somewhere and linked against that [02:56:13 CET] <Guiri> ls [04:39:37 CET] <VelusUniverseSys> what does this mean notification: Speex header too small [04:39:49 CET] <VelusUniverseSys> and WARNING: Can't write keyframe-seek-index into non-seekable output stream! Writing Skeleton3 track [04:46:25 CET] <J_Darnley> The second sounds like you're writing to a pipe or perhaps an http stream [04:47:03 CET] <J_Darnley> The first sounds like a broken file. [04:58:39 CET] <VelusUniverseSys> ok it shouldnt be a broken file lol but ok, i will check with another file [06:21:00 CET] <Melchior> Unexpected decoder output format Planar 420P 10-bit little-endian [06:21:25 CET] <Melchior> What in the world is that error from/for (?) [06:21:37 CET] <Melchior> I know its related to 10bit colour space [06:21:55 CET] <Melchior> Yet.... I'm using newish smplayer / mplayer [06:25:31 CET] <Melchior> Gstreamer plays these fine; so its a mplayer ffmpeg type issue [06:25:38 CET] <Melchior> ^^ what gives?! [07:28:08 CET] <C0nundrum> When recording an hls stream, if there is a low framerate in the video could it be because of the connection ? [16:41:13 CET] <k_sze> Is there some sensible parameter to remix TrueHD to ac3? [16:44:58 CET] <JEEB> truehd tracks don't contain an ac3 track [16:45:03 CET] <JEEB> you'd have to re-encode [16:45:26 CET] <JEEB> with DTS-HD MA it's all an extension on top of base DTS so you can "extract" the base DTS thing in theory [16:45:28 CET] <k_sze> JEEB: of course [16:45:39 CET] <k_sze> I know I have to re-encode [16:46:19 CET] <k_sze> So what I really mean is whether there is some sensible ffmpeg options I should use to re-encode and remix TrueHD to Dolby Digital 5.1. [16:46:44 CET] <k_sze> I mean options that some of you may have tried and produced reasonably good results. [16:49:03 CET] <rcombs> -acodec ac3 [16:54:38 CET] <k_sze> I mean bitrate/crf/q and mixing ratios. [16:55:08 CET] <k_sze> (Is it even possible to specify mixing ratios from TrueHD to Dolby 5.1? [17:38:34 CET] <someguy234> is it possible to change filter parameters over time with ffmpeg? [17:39:11 CET] <someguy234> http://ffmpeg.org/ffmpeg-filters.html#Timeline-editing this only allows for conditional enabling/disabling of given filter [17:40:12 CET] <someguy234> what I mean is to for example use the "rotate" filter and have the rotation increase by a couple of degrees every frame indefinitely [17:57:52 CET] <xeons> I'm trying to combine several dozen videos (h.264/mov) with concat and noticing an increasing lag in audio sync. All the videos are the same (from the same camera) and re-coding fixes the image but any kind of -c copy is a problem [17:58:25 CET] <xeons> *fixes the video sync issue, but coping the codec results in this discrepancy [17:59:25 CET] <xeons> I've tried -auto_convert 1, -fflags +genpts, -async 1, -copyts, and several other flags as hours of googling have detailed [17:59:49 CET] <xeons> ffmpeg -f concat -fflags +genpts -i ./segments.txt -copyts -c copy test8.mov [18:17:36 CET] <rocktop> Hello is it possible to add logo to while video and add video intro and background in same command line ? [18:18:08 CET] <rocktop> s/while/whole [18:21:58 CET] <rocktop> anyidea ? [18:32:31 CET] <groupers> Hi, I have a video production device that will stream to a RTMP server, is there some way to get ffmpeg to listen for incoming connections as if it were adobe media server [18:33:14 CET] <Mavrik> ffmpeg isn't a streaming server, it's a video processing tool [18:33:34 CET] <Mavrik> So you'll need to find another streaming server (a lot of them can invoke ffmpeg if needed). [18:33:37 CET] <groupers> the other options are windows media pull and Windows media push. if I use the pull option it gives me an http address that will play in windows media player or vlc but not ffmpeg, could I play in vlc and stream to ffmpeg? [18:33:47 CET] <groupers> the end goal is to transcoder to [18:34:25 CET] <groupers> transcode to a format that can be played in a flash applet or embedded player on a webpage [18:34:37 CET] <groupers> and send to a multicast address [18:36:09 CET] <furq> groupers: https://github.com/arut/nginx-rtmp-module [18:36:48 CET] <furq> specifically https://github.com/arut/nginx-rtmp-module/wiki/Directives#exec_push [18:38:05 CET] <furq> also if your device streams to rtmp then it'll already in a format which can be played by flash [18:38:09 CET] <furq> +be [18:39:43 CET] <rocktop> anyidea ? [18:41:38 CET] <groupers> furq great. so the nginx module will listen for incoming rtmp connections just like flash media server [19:19:13 CET] <conkis> is there a way to set ffmpeg INPUT codec explicitly, without probing/analyzing ? [19:22:15 CET] <c_14> ffmpeg -c:a codec -i blah should work [19:22:26 CET] <c_14> Well, -c:[stream identifier] [19:24:11 CET] <conkis> will tray, thank you c_14 [19:24:16 CET] <conkis> try* [19:32:40 CET] <rocktop> is it possible to merge different video size for example 1280x720 with 180x320 ? [19:37:28 CET] <rocktop> anyidea ? [19:39:00 CET] <jkqxz> rocktop: Probably. You'll need to more precisely define what you mean by "merge", though. [19:41:04 CET] <rocktop> jkqxz: I have video format 1280x720 and other one with 180x320 in its format I would liketo merge theme to be one video [19:43:20 CET] <jkqxz> You want them to play one after the other, from the same file? [19:45:51 CET] <jkqxz> Putting different resolution videos in the same container is unfortunately a can with many worms in. It is easiest if you are happy to up/downscale one of the videos to match the other. [23:53:39 CET] <rocktop> Hello , how can I merge 2 video with different format the firstone 1280x728 the secnd format 180x320 ? [23:56:01 CET] <J_Darnley> What do you mean by "merge" [23:56:15 CET] <J_Darnley> Several very different oprations might be "merge" [23:56:34 CET] <J_Darnley> I seem to recall you were asked this before. [23:57:17 CET] <J_Darnley> Should the second video play after the first? [23:57:37 CET] <J_Darnley> Should they play at the same time but be stacked next to each other? [23:57:49 CET] <J_Darnley> Should there be two video streams? [23:58:30 CET] <rocktop> J_Darnley, the first one is the intro and second play after the first one [23:58:52 CET] <J_Darnley> Then you should look at "concat" [23:59:17 CET] <rocktop> J_Darnley: I already try but I failed [23:59:24 CET] <rocktop> can you help please [23:59:30 CET] <J_Darnley> What exactly failed? [23:59:43 CET] <furq> they're different resolutions [23:59:49 CET] <rocktop> J_Darnley: I can't build right syntax [23:59:52 CET] <furq> you need to scale one before you can concat them [00:00:00 CET] --- Sun Feb 14 2016

1 0

ffmpeg-devel.log.20160212
by burek 13 Feb '16

13 Feb '16

[00:34:21 CET] <cone-576> ffmpeg 03Michael Niedermayer 07master:9313fb6a3e90: avfilter/vf_zoompan: Remove dead store [00:44:31 CET] <J_Darnley> Oh lord. That's a messy patch (gdi mouse cursor). [00:44:53 CET] <J_Darnley> tabs, whitespace changes, unrelated changes. [00:45:23 CET] <nevcairiel> corporate junior level developer spotted? :d [00:48:29 CET] <drv> it also deletes the call to GetIconInfo() but still uses the data that it was supposed to fill out? [00:48:48 CET] <J_Darnley> I think so [00:49:04 CET] <J_Darnley> I was just applying the patch to see what the real changes are [00:51:20 CET] <J_Darnley> I decided I shouldn't send a complaint-only email [00:51:41 CET] <nevcairiel> i posted a mail about the style problems [00:51:41 CET] <wm4> isn't there a much better capture method with d3d [00:51:47 CET] <nevcairiel> there is [00:52:02 CET] <furkan> i'm trying to re-build omxplayer and whenever i try "make ffmpeg" after modifying something, it tells me "make: 'ffmpeg' is up to date." [00:52:03 CET] <nevcairiel> i even implemented it once for another project [00:52:08 CET] <furkan> does anybody have any tips as to why that might be? [00:53:16 CET] <drv> i think the actual change is probably bogus too, unless windows only allows integer scaling factors [00:53:35 CET] <nevcairiel> i think the scaling things are in dpi [00:54:06 CET] <drv> in 'pos.x = pos.x * desktopvertres / vertres;', the two vertres things are ints [00:54:23 CET] <nevcairiel> it might overflow, but if it doesnt it should be accurate [00:54:45 CET] <drv> hmm, yeah, i guess that is OK since the multiply is done first [00:55:54 CET] <J_Darnley> My spacebar is getting worse. [00:55:59 CET] <jamrial> huh, fate-source went bananas [00:56:34 CET] <jamrial> "sed: 1: "/\/\|^ff/!{s/$.*$/ffm ...": bad flag in substitute command: '}'" [00:56:36 CET] <nevcairiel> jamrial: i bet that new check for inclusion guards just doesnt work everywhere [00:56:46 CET] <jamrial> fails with every single header now [00:56:46 CET] <nevcairiel> f'ing source checks [00:56:47 CET] <cone-576> ffmpeg 03Michael Niedermayer 07master:43bf15d1a425: avfilter/avf_showfreqs: assert that variables are initialized by switch() [00:56:48 CET] <cone-576> ffmpeg 03Michael Niedermayer 07master:156013111a33: avfilter/avf_ahistogram: assert that variables are initialized by switch() [00:57:41 CET] <jamrial> Timothy_Gu: ^ [00:58:56 CET] <jamrial> seems to be free/net/openbsd only [01:53:20 CET] <cone-576> ffmpeg 03Michael Niedermayer 07master:bbc4d069d0ef: avcodec/cfhd: Check the number of tag/value pairs [02:00:55 CET] <cone-576> ffmpeg 03Kieran Kunhya 07master:1e2fd57fc096: avcodec/cfhd: Fix errors with valid files. [02:10:21 CET] <michaelni> rcombs, what should be done about the sidx regressions caused by your commit ? (https://trac.ffmpeg.org/ticket/5090 https://trac.ffmpeg.org/ticket/5216) [02:10:50 CET] <rcombs> oh yeah, I keep forgetting about that [02:10:54 CET] <rcombs> lemme take another look [02:11:25 CET] <michaelni> please do and thanks [03:55:37 CET] <furkan> is it possible to disable pthreads without recompiling ffmeg? [03:55:44 CET] <furkan> *ffmpeg, of course [03:56:28 CET] <J_Darnley> Set -threads maybe? [03:56:42 CET] <furkan> it's on a raspberry pi so compiling takes ages heh [03:56:46 CET] <J_Darnley> Where the fuck did my "1" go? [03:57:05 CET] <J_Darnley> POS keyboard [03:57:23 CET] <furkan> any function i can use in the omxplayer source similar to how i used av_dict_set_int to set the buffer size? [03:58:09 CET] <furkan> since i can't pass any command-line arguments [03:59:26 CET] <jamrial> you should cross compile from a desktop [03:59:27 CET] <furkan> looks like i should probably be able to do av_dict_set_int(&d, "threads", 1, 0) [04:00:02 CET] <furkan> jamrial: i suppose i could [04:02:18 CET] <furkan> hmm i think -threads is only for encoding [04:05:18 CET] <furkan> kierank: i'm 99% sure that the circular buffer w/ pthreads enabled is what's causing my problem... i just need to rebuild with pthreads disabled to confirm [04:05:46 CET] <furkan> earlier today it was working fine with pthreads disabled, but then i forgot to disable it when rebuilding again [04:15:09 CET] <furkan> michaelni: seems like you're the one who implemented the circular buffer, any ideas why that might be happening? [04:15:34 CET] <furkan> i'll get a screenshot hold on [04:16:55 CET] <furkan> michaelni: this is what happens https://www.dropbox.com/s/8s6m1ssi1wowbky/IMAG0096.jpg?dl=0 [04:17:23 CET] <furkan> with pthreads disabled and by increasing my buffer_size, the corruption is solved [04:17:38 CET] <furkan> but with pthreads enabled i can't seem to find any way to fix it [04:18:23 CET] <furkan> i've tried increasing the kernel's UDP receive buffer size, increasing both buffer_size and the fifo buffer size [04:18:48 CET] <furkan> but basically what's happening is that the kernel is dropping UDP packets because the application isn't processing them fast enough [04:20:26 CET] <furkan> i confirmed that via the "RcvbufErrors" row with netstat -s [04:27:37 CET] <bbfan> Hi there! I have a ffmpeg installation issue. Can anybody please help me? [04:37:10 CET] <J_Darnley> No because you haven't said anything specific [04:37:46 CET] <J_Darnley> Also you should be asking thatsort of question in #ffmpeg [04:41:06 CET] <bbfan> Nobody's answering at #ffmpeg. I just re-run the ./configure with the --enable-libass option, it ends with no errors but when i execute ffmpeg there is no libass compiled, so i guess it's the old version. How do i update it? [04:42:06 CET] <J_Darnley> After you run configure you need to run make [04:42:31 CET] <cone-576> ffmpeg 03Michael Niedermayer 07master:6c6f2e49e40a: avcodec/h264_slice: remove duplicate unconditional picture_structure setting code [04:42:32 CET] <cone-576> ffmpeg 03Michael Niedermayer 07master:873158fd76f3: avcodec/h264_slice: assert relation between current_slice ans slice_ctx [04:42:33 CET] <cone-576> ffmpeg 03Michael Niedermayer 07master:69738466189a: avcodec/h264_slice: Fix dequant table init with field pictures [04:43:23 CET] <bbfan> yes, i run make, make install, make distclean and hash -r, but it happens what i told. [04:43:47 CET] <J_Darnley> Why can't you just run it out the the directory you built it in? [04:44:04 CET] <bbfan> I executed find / -name ffmpeg and the new ffmpeg binary is at /root/binary (my build root is /root/ffmpeg_build/) [04:44:12 CET] <bbfan> i don't know, that's the question [04:44:21 CET] <J_Darnley> Just run ./ffmpeg and be done with it. [04:46:15 CET] <bbfan> not found [04:46:37 CET] <J_Darnley> Oh yes. You got rid of it by running make distclean [04:46:45 CET] <J_Darnley> Why would you do that? [04:47:04 CET] <bbfan> ok, let me try [04:58:59 CET] <bbfan> Now this is the error [04:59:00 CET] <bbfan> ./ffmpeg: error while loading shared libraries: libass.so.5: cannot open shared object file: No such file or directory [04:59:32 CET] <bbfan> i looked for it and it's located on /root/ffmpeg_build/lib/ [05:01:57 CET] <J_Darnley> Then either set LD_LIBRARY_PATH, move it, or stop building shared libraries. [05:02:08 CET] Action: J_Darnley sleeps [06:06:29 CET] <cone-576> ffmpeg 03James Almer 07master:0abdf7030eeb: avcodec/vc2enc_dwt: add missing header [07:09:41 CET] <Timothy_Gu> damn sed [07:28:42 CET] <Timothy_Gu> anybody with an *bsd set up i could use to test the patch I just sent on sed usage? [07:52:08 CET] <relaxed> Timothy_Gu: there's http://sdf.lonestar.org/ [08:41:07 CET] <furkan> michaelni: yeah i've confirmed, i built 2 copies of ffmpeg, one with pthreads enabled and one with disabled. with the copy that has pthreads enabled, i can't seem to do anything to eliminate the corruption [08:41:36 CET] <furkan> but of course CPU load is much higher with pthreads disabled [08:44:01 CET] <furkan> michaelni: and i only seem to have this problem on the raspberry pi... seems to work fine on my laptop with ffplay. if there's any further tests that i could run to help isolate the issue i'd be very interested in trying them out since i'd really like to solve this issue [09:11:41 CET] <furkan> i will run some more tests tomorrow and report back [11:25:08 CET] <nevcairiel> AVFilter is crazy, it has a init_opaque function which can take a struct to initialize the filter, and it also defines such a struct in the header file .... but the public init function and this init_opaque function do not actually forward said init structure [11:31:32 CET] <durandal21707> nevcairiel: what init structure? [11:31:55 CET] <nevcairiel> AVBufferSinkParams and AVABufferSinkParams [11:32:10 CET] <nevcairiel> they cant effectively be used because no public function properly accepts them [11:34:57 CET] <durandal21707> it is just hack around to give list of pix_fmts/sample_fmts/... [11:36:12 CET] <wm4> so the struct and av_abuffersink_params_alloc are useless? [11:36:21 CET] <nevcairiel> pretty much [11:36:26 CET] <wm4> fascinating [11:36:26 CET] <nevcairiel> you cant pass it to the filter [11:36:45 CET] <nevcairiel> from what i gather, it supposed to be passed to avfilter_graph_create_filter in the opaque parameter [11:36:50 CET] <nevcairiel> but that function doesnt do anything with it [11:37:38 CET] <durandal21707> report it to saste [12:17:52 CET] <atomnuker> so in what order are the entries in doc/encoders.texi arranged? [12:18:09 CET] <atomnuker> doesn't seem to be alphabetical [12:20:48 CET] <rcombs> the order in which people put them [12:21:35 CET] <rcombs> tbh I don't think doc/encoders, doc/filters, etc& should be single monolithic files [12:22:28 CET] <rcombs> have an autogenerated file a la ffmpeg-all that shows all the <insert components>' docs, sure, but having one big source file is awkward [12:44:35 CET] <wm4> why does Mats have to ALWAYS reply to himself at least a few times [12:47:33 CET] <kierank> https://forums.xilinx.com/t5/Xcell-Daily-Blog/One-minute-shows-you-how-to-i… [12:47:37 CET] <kierank> fpga accelerated ffmpeg [13:08:42 CET] <wm4> jkqxz: so I guess we wait until Libav pushes the hwcontext patchset? [13:09:59 CET] <jkqxz> It seems easier to just work on libav. (Also stops there being any nasty collision later.) [13:12:24 CET] <jkqxz> Unless this is already colliding in some way which I don't know about, and will not all be merged? [13:13:26 CET] <wm4> so, that should be fine and would make it easier for everyone [13:13:41 CET] <wm4> s/so/no/ [13:16:35 CET] <iive> or just don't have about Libav [13:16:49 CET] <iive> pushing patches there is no fun. [13:17:20 CET] <durandal21707> you must vertically align everything [13:17:23 CET] <iive> ops... don't care .. [13:17:43 CET] <iive> aligning and sorting alphabetically is easy... [14:40:53 CET] <wm4> nevcairiel: so about this main10 dxva patch (I'm already getting afraid of asking) [14:57:32 CET] <Timothy_Gu> relaxed: do you have an account on sdf? [15:04:40 CET] <Compn> sdf still around? :) [15:17:25 CET] <durandal21707> rcombs: why you rescale loop_start in brstm/bfstm? [15:18:07 CET] <rcombs> durandal21707: because it's set on the container, not the stream [15:18:25 CET] <durandal21707> but its sample accurate [15:18:58 CET] <durandal21707> it sets exact sample where to seek [15:19:25 CET] <rcombs> the container doesn't have a time base [15:20:24 CET] <durandal21707> container time base is sample rate [15:25:21 CET] <rcombs> durandal21707: hmm? That's on the stream [15:25:45 CET] <rcombs> in the actual brstm format, yes, the time base is the sample rate [15:26:06 CET] <rcombs> but AVFormatContext's time base is AV_TIME_BASE [15:46:16 CET] <durandal21707> no yet candidates from outreachy? [15:52:33 CET] <relaxed> Timothy_Gu: no, but I know it's free and they use netbsd [16:00:16 CET] <cone-415> ffmpeg 03Michael Niedermayer 07master:03b26a2514c5: avfilter/vf_zoompan: Fix use of uninitialized variables [16:00:17 CET] <cone-415> ffmpeg 03Michael Niedermayer 07master:f9f684c0fb8c: avfilter/vf_zoompan: Initialize ret [16:00:18 CET] <cone-415> ffmpeg 03Michael Niedermayer 07master:817bf8483e13: avfilter/af_anequalizer: Avoid loss of precision when calculating nyquist frequency [16:59:06 CET] <JEEB> ok, so how much stream selection logic can there be in filter_complex and how simply can one put conditionals there? I know that -map supports conditionals at least [16:59:11 CET] <JEEB> (by appending ?) [16:59:22 CET] <JEEB> but if I have to do filter_complex I can't use -map IIRC [17:00:15 CET] <JEEB> like, "pick swedish if exists, otherwise something lavf outputs as '???' otherwise anything else" [17:01:02 CET] <ubitux> lavfi doesn't have much knowledge on stream information unfortunately [17:01:14 CET] <JEEB> well you can do [0:v:0] etc [17:01:26 CET] <JEEB> so I guess the mapping way of language data should be similar to how -map works? [17:01:38 CET] <ubitux> yeah, but unfortunately, iirc the lavfi source doesn't import the metadata of the stream [17:01:43 CET] <JEEB> ok [17:01:47 CET] <ubitux> so you won't for instance have a meta on the lang [17:01:49 CET] <nevcairiel> i dont think lavfi gets any information beyond the stream numbers [17:01:56 CET] <ubitux> i'd love to have stream info injected in lavfi btw [17:02:22 CET] <JEEB> ok... so that's one thing that is yet to have been done [17:02:37 CET] <JEEB> then there's the question of the conditional stream selection [17:02:50 CET] <JEEB> which -map seems to work with, but -map + filter_complex doesn't seem to work together [17:02:58 CET] <JEEB> unless I did something fabulously wrong [17:03:26 CET] <wm4> in theory ffmpeg.c could do the mapping? [17:03:47 CET] <JEEB> I guess so [17:06:17 CET] <Daemon404> "It works here with a repo I found on my hard disc" [17:06:20 CET] Action: Daemon404 sighs [17:07:55 CET] <durandal21707> libavcodec/x86/h264_qpel.c:388:1: warning: unused function 'ff_avg_h264_qpel8or16_hv1_lowpass_mmxext' [-Wunused-function] [17:16:26 CET] <cone-415> ffmpeg 03Paul B Mahol 07master:13ea7c978057: avfilter/avf_showspectrum: improve pts handling [17:50:16 CET] <cone-415> ffmpeg 03Rostislav Pehlivanov 07master:0384e3d5dfe1: doc/encoders: add documentation for the VC-2 encoder [17:58:08 CET] <BBB> Daemon404: wbs: do you guys want to review the c99toc89 patch that came in? [17:58:27 CET] <BBB> (I cant believe people use that stuff, I mean, isnt msvs 2015 out?) [17:59:39 CET] <nevcairiel> some people are even so crazy to use it for other projects than ffmpeg/libav [18:01:12 CET] <JEEB> BBB: has been for a while now [18:02:05 CET] <BBB> nevcairiel: well, thats not unthinkable, although the reasons for its existence were indeed kinda highly politically related to ffmpeg specifics :-p [18:02:47 CET] <Daemon404> BBB, yeah saw [18:02:52 CET] <Daemon404> wbs wants a bit of time to try it out [18:02:59 CET] <BBB> ah okiedokie, cool [18:03:02 CET] <wm4> I'm glad you made c99toc89, thank you [18:03:05 CET] <BBB> just wanted to make sure it wouldnt get lost [18:29:12 CET] <jamrial> why do we still support msvc 2012? it even miscompiles ffmpeg [18:42:46 CET] <jamrial> "No code is (ever) committed that we believe isn't fit to release" mmh... [18:43:04 CET] <nevcairiel> and yet we let him commit things [18:43:05 CET] <nevcairiel> :D [18:43:13 CET] <kierank> let who? [18:44:27 CET] <jamrial> cehoyos sure likes to talk as if he was the project's spokeman [18:44:36 CET] <j-b> lol [18:45:12 CET] <wm4> cehoyos is pretending a tricycle with broken wheels is a sports car herew [18:45:45 CET] <wm4> even the guy who tried to maintain the lib completely agrees that it's a lost cause [18:55:03 CET] <RiCON> atomnuker: tasts? [18:57:59 CET] <Daemon404> i have an easy fix for myself... [18:58:08 CET] <cone-415> ffmpeg 03Derek Buitenhuis 07master:5da6bed1e025: MAINTAINERS: Remove myself as libutvideo maintainer [18:58:35 CET] <wm4> why didn't you replace it with cehoyos' name [18:58:45 CET] <wm4> blunder [18:58:51 CET] <Daemon404> that would be trolling [18:58:56 CET] <Daemon404> whereas nobody can argue agains this [18:59:05 CET] <Daemon404> i dont need approval to NOT maintain something [18:59:38 CET] <durandal21707> yea [19:00:57 CET] <jamrial> just remove it. nobody except him wants that thing in the tree [19:01:08 CET] <Daemon404> btw [19:01:09 CET] <Daemon404> libx264.c Mans Rullgard, Jason Garrett-Glaser [19:01:12 CET] <Daemon404> this is a gem [19:01:15 CET] <Daemon404> one of them is long gone [19:01:19 CET] <Daemon404> and one of them literally no longer exists [19:01:46 CET] <durandal21707> michaelni: is there a way to disable rescaling when input resolution changes? [19:02:11 CET] <durandal21707> ffmpeg will auto-insert scale filter to rescale to previous resolution [19:02:13 CET] <wm4> jamrial: still kind of needs a civil resolution I guess [19:02:24 CET] <RiCON> atomnuker: https://j.fsbn.eu/r7FO.txt could only find these two [19:02:38 CET] <wm4> durandal21707: lavfi doesn't support dynamic changes, even if you pretend it does [19:03:00 CET] <durandal21707> wm4: lavfi auto-inserts scale filter [19:03:17 CET] <Daemon404> even via api? [19:19:35 CET] <JEEB> wow [19:19:41 CET] <JEEB> the "discussion" regarding libutvideo [19:19:48 CET] <Daemon404> tl;dr car [19:19:48 CET] <Daemon404> l [19:19:49 CET] <JEEB> aka everyone vs carl [19:20:37 CET] <Daemon404> maybe i should add 10bit to the native decoder. but im so lazy. [19:20:44 CET] <Daemon404> and then itll still be "but it's faster" [19:20:51 CET] <JEEB> except it isn't [19:22:13 CET] <jamrial> Daemon404: it's used http://lucy.pkh.me/ffmpeg-coverage-snapshots/1455242957/src/libswscale/slic… [19:22:30 CET] <Daemon404> by what [19:24:12 CET] <kierank> Daemon404: just push it [19:24:14 CET] <kierank> fuck carl [19:24:17 CET] <jamrial> no idea [19:24:36 CET] <Daemon404> kierank, no im not, because i: a) value sanity b) am not going to start flame/revert wars [19:24:52 CET] <wm4> so uh let's use this vote thingy? [19:26:04 CET] <Daemon404> jamrial, ill wait for the author [19:26:10 CET] <Daemon404> because i really dont get how it worked [19:26:35 CET] <jamrial> wm4: sure. start a new thread and call for a vote. give it a week deadline or so [19:26:59 CET] <Daemon404> er [19:27:12 CET] <Daemon404> wasnt carl one of the people who didnt agree to teh entire concept of voting [19:27:21 CET] <nevcairiel> who cares what he thinks [19:27:24 CET] <Daemon404> anyway [19:27:27 CET] <Daemon404> ENOTMYPROBLEM [19:27:29 CET] <Daemon404> no longer maintainer [19:30:33 CET] <jamrial> Daemon404: too bad for him, the rest agree and that's what was decided in the last meeting [19:33:58 CET] <atomnuker> RiCON: thanks [19:36:48 CET] <durandal_1707> I will start thread, no libutvideo maintainer - removal [19:42:38 CET] <cone-415> ffmpeg 03Rostislav Pehlivanov 07master:d119268ed23b: doc/encoders: fix typos in the vc2 encoder documentation [19:42:39 CET] <cone-415> ffmpeg 03Rostislav Pehlivanov 07master:f0a8212436c4: aacenc: make a better estimate for the audio bitrate if not provided [19:58:46 CET] <cone-415> ffmpeg 03Lou Logan 07master:0b0121d7f209: doc/filters: add geq diagonal split screen example [20:08:26 CET] <wm4> so should I create a drama... I mean vote thread about the libutvideo issue, or do we just forget about it [20:14:19 CET] <wm4> durandal_1707: couldn't filters have nicer names than "Parsed_asplit_0" [20:14:59 CET] <nevcairiel> you can add them explicitly in API and then name them =p [20:15:03 CET] <durandal_1707> what you propose? [20:15:55 CET] <wm4> nevcairiel: this is about graph strings [20:16:07 CET] <wm4> durandal_1707: dunno, maybe just "asplit", and the second "asplit_1" etc. [20:36:44 CET] <jamrial> wm4: create it [20:56:06 CET] <JEEB> anyone else can replicate issues with https://kuroko.fushizen.eu/videos/sppl_01_demo_640x360.wmv ? [21:06:55 CET] <wm4> what issues? [21:12:46 CET] <JEEB> mpv on windows at least seems to get a lot of various artifacts [21:13:10 CET] <JEEB> and on ffmpeg I see tons of timestamp errors [21:14:10 CET] <wm4> huh why does ffprobe -show_frames not show any audio frames [21:14:55 CET] <JEEB> probably because of "Audio: WMA Lossless 44100Hz stereo" [21:15:04 CET] <Compn> lossless ,,, [21:15:11 CET] <Compn> JEEB : does new asf demuxer work on it ? [21:15:17 CET] <Compn> asf_o or whatnot [21:15:18 CET] <JEEB> is that under another name? [21:15:19 CET] <JEEB> oh [21:15:29 CET] <wm4> JEEB: most of the time it seems to skip audio frame timestamps, and when it doesn't they seem to be inconsistent [21:15:29 CET] <JEEB> let me see [21:15:33 CET] Action: Compn downloading [21:16:20 CET] <JEEB> hmm [21:16:31 CET] <Compn> if not, also report bug to them ;p [21:16:57 CET] <JEEB> yea, just checking the switch to mpv to switch demuxer [21:17:19 CET] <wm4> JEEB: that's complicated [21:17:30 CET] <Compn> in mplayer it would be -demuxer lavf -lavfdopts format=asf_o [21:17:34 CET] <Compn> but mpv? no clue... [21:17:38 CET] <wm4> --demuxer-lavf-format=asf_o [21:17:46 CET] <JEEB> dänks [21:17:49 CET] <durandal_1707> why lines looks like but deinterlacer was used? [21:18:03 CET] <durandal_1707> *bad [21:18:16 CET] <JEEB> downscaling by a random games company [21:18:17 CET] <JEEB> most probably [21:18:30 CET] <JEEB> meanwhile using wma lossless :D [21:18:32 CET] <wm4> doesn't look like asf_o works better [21:18:45 CET] <wm4> this is probably a parser problem of some sort anyway [21:18:46 CET] <Compn> do other wma lossless files work (regression?) [21:18:53 CET] <Compn> yes, curious about parsing... [21:19:25 CET] <JEEB> seems to work after I seek to 1:25 [21:19:48 CET] <Compn> http://samples.ffmpeg.org/A-codecs/lossless/luckynight.wma [21:19:50 CET] <JEEB> so it could have something to do with the automagical VFR that MS did [21:20:03 CET] <JEEB> since the first pictures have little actual motion [21:20:41 CET] <Compn> seems to work ok with mplayer asf demuxer [21:20:52 CET] <Compn> and binary audio decoder :P [21:21:51 CET] <Compn> not seeing artifacts with ffplay version N-68788-g37b35fe Copyright (c) 2003-2014 [21:22:36 CET] <wm4> JEEB: actually mpv does some gross shit on pts resets, this could cause the video corruption [21:22:45 CET] <JEEB> oh, ok [21:23:50 CET] <durandal_1707> I see no corruptions [21:26:17 CET] <Compn> did you enable framedropping ? [21:28:48 CET] <durandal_1707> disabling dropping fixes it [21:29:04 CET] <durandal_1707> sync issues [22:22:08 CET] <cone-415> ffmpeg 03Paul B Mahol 07master:f80b4750b9e7: doc/filters: add one more blend example [00:00:00 CET] --- Sat Feb 13 2016

1 0

ffmpeg.log.20160212
by burek 13 Feb '16

13 Feb '16

[00:36:05 CET] <onyx> hey guys so Im using this command to get a image every minute ffmpeg -i myvideo.avi -vf fps=1/60 img%03d.jpg [00:36:45 CET] <onyx> its working well, the only issue I have it that since there is movement most of the images come out some what blurry [00:37:12 CET] <onyx> I tried using the "select='eq(pict_type,PICT_TYPE_I)'" to see if selecting I frames would do better [00:38:52 CET] <onyx> but it errors saying nable to find a suitable output format for 'select='eq(pict_type,PICT_TYPE_I)'' [00:38:52 CET] <onyx> select='eq(pict_type,PICT_TYPE_I)': Invalid argument [00:39:06 CET] <onyx> any idea of how can I get the most clear pictures possible? [00:43:05 CET] <anewinckname> I am trying to make a script that when ran will execute a ffmpeg command.I have the script mostly made,minus one issue.I need to output the encoded video in to a new folder(which is made at the start of the script).Im having trouble finding out how to output to a folder. Here is the script: http://pastebin.com/DDeCjmwH [00:51:25 CET] <thesurferdude1> Anyone have any tips on how to run ffmpeg with a HTTP network stream as the input? I have a HDHomerun Prime but whenever I try to specify the url directly via command line but the connection cuts out almost immediately [00:51:53 CET] <thesurferdude1> works fine if I pull some video first and then point ffmpeg at the file instead [01:26:53 CET] <llogan> anewinckname: this isn't a scripting help resource. anyway, you should quote the done in your mkdir command to make it literal, otherwise your shell may confuse it with done as in a loop [02:27:57 CET] <anewinckname> llogan:This is an ffmpeg issue.It is not taking the given output and outputing in to the folder given at the "done/$file" the file being the name given in the input and as well as the subtitle field [02:31:07 CET] <anewinckname> llogan:Running the command without the script, it still does not output the file in to the done folder [02:52:59 CET] <anewinckname> llogan: http://pastebin.com/p8q2atQT [02:53:39 CET] <llogan> NICE CAPS [02:54:10 CET] <llogan> not enough info here to provide a solution [02:55:15 CET] <anewinckname> It is exactly what you asked for. [02:55:46 CET] <llogan> you didn't include the console output [02:56:01 CET] <anewinckname> You want 3 thousand lines of frame spam? [02:56:10 CET] <anewinckname> There is no error [02:56:10 CET] <llogan> you may trim multiple repeating lines [03:05:22 CET] <MarkusDB1> Can ffmpeg use nvidias hardware accelleration of h.264 encoding? (The thing nvidia has implemented in windows as "nvidia experience, shadowplay"? [03:06:36 CET] <kepstin> yeah, via "nvenv"; see https://trac.ffmpeg.org/wiki/HWAccelIntro [03:06:43 CET] <kepstin> apparently it's a bit tricky to set up? [03:06:50 CET] <MarkusDB1> kepstin: thanks! [03:06:59 CET] <MarkusDB1> kepstin: imo this technology is SOOO cool [03:07:01 CET] <kepstin> unless you need low-cpu-usage high-speed encoding, I'd still recommend x264 [03:07:28 CET] <kepstin> nvenc* [03:07:30 CET] <MarkusDB1> I want that realtime stuff that the kids got for gaming... but for my linux screen capturing.. [03:07:38 CET] <MarkusDB1> to use my VPS'es as desktops =) [03:07:55 CET] <MarkusDB1> or.. well... [03:07:59 CET] <furq> isn't that what vnc is for [03:08:05 CET] <MarkusDB1> dedicated servers with nvidia gpus as desktops =) [03:08:14 CET] <MarkusDB1> furq: vnc is slow as.. snails [03:08:36 CET] <MarkusDB1> furq: google makes much magic with their "remote desktop" plugin for chrome though [03:08:41 CET] <MarkusDB1> fastest thing I've ever seen.. [03:08:47 CET] <MarkusDB1> using no special hardware [03:08:49 CET] <furq> i'd be surprised if nvenc was low enough latency to make this work well [03:08:53 CET] <furq> can't say i've tried it though [03:08:59 CET] <MarkusDB1> furq: it's amazing [03:09:31 CET] <MarkusDB1> I bet with the ffmpeg DIY settings.. nvidia might not implement any.. "in app" limitations.. [03:10:09 CET] <MarkusDB1> on the gamer market they will only go for very safe settings... since else the kids complain [03:10:20 CET] <kepstin> the nvidia hardware encoder is used by stuff like the steam remote play with decently low latency, so the hardware should be capable of it [03:10:37 CET] <MarkusDB1> if you got your own setup.. and own datacenter and good gigabit uplinks, you might be able to tweak the settings more with the ffmpeg [03:10:53 CET] <MarkusDB1> kepstin: yeah sure.. No doubt the hardware can do it =) [03:11:05 CET] <kepstin> note that as used in ffmpeg, it can't encode directly out of the screen scanout buffer like I assume the windows stuff does, so it'll be slower due to some extra copies. [03:11:12 CET] <MarkusDB1> kepstin: I just want it scriptable in CLI and not some stupid "kid gamer UI" [03:12:09 CET] <MarkusDB1> ah is there is gotchas? [03:12:18 CET] <MarkusDB1> *so there is.. [03:26:07 CET] <MarkusDB1> kepstin: btw, x264 that is intel "quicksync" based accelleration... correct? [03:26:28 CET] <kepstin> no, intel qsv is a hardware encoder like nvidia's [03:26:33 CET] <kepstin> x264 is a software encoder [03:26:37 CET] <MarkusDB1> I see [03:26:43 CET] <MarkusDB1> so more latency for my application I guess [03:27:07 CET] <MarkusDB1> For what I'm about to do.. I like.. extremely low latency.. quality is not really so picky [03:27:30 CET] <kepstin> x264 can do really low latency [03:28:00 CET] <MarkusDB1> how low are we talking? [03:28:11 CET] <MarkusDB1> I need.. < 200ms or something [03:28:40 CET] <kepstin> in fact, it has a built-in tuning called 'zerolatency' that makes it have have no frames buffered internally, you get a frame out for each frame in limited only by encoding speed; and it can be tweaked to run /really/ fast. [03:28:43 CET] <MarkusDB1> I want to use it for remote desktop. [03:29:20 CET] <kepstin> the main benefit of using a hardware encoding block like nvidia or intel qsv for this sort of thing is that they don't use the cpu. [03:29:23 CET] <MarkusDB1> kepstin: you know of any.. "screen sharing" applications (any os is fine) that uses x264 low latency streaming currently? [03:29:31 CET] <kepstin> which leaves more cpu for apps [03:29:54 CET] <MarkusDB1> https://www.youtube.com/watch?v=NQ73aVnWzhA <- what I'm trying to achieve [03:30:11 CET] <VelusUniverseSys> hello is anyone about? [03:30:33 CET] <DeHackEd> just state your question [03:36:36 CET] <MarkusDB1> kepstin: Almost thinking about installing a "gaming" pc.. now, since that stuff works so darn well =) [04:08:53 CET] <anewinckname> ffmpeg will not send a finished file to the output folder given.command:http://pastebin.com/TedSKTVT output:http://pastebin.com/vJ1Sb5xK [04:16:17 CET] <kepstin> anewinckname: I don't see any problems in your command-line or output... [04:21:17 CET] <anewinckname> kepstin: I checked permissions aswell as the "done" folder exists before the command is executed [04:21:48 CET] <kepstin> it looks like it's all working; the command would have failed with an error if it couldn't write to the output file. [04:24:15 CET] <bbfan> Hi there. Can someone help me? Is a ffmpeg install issue [04:40:00 CET] <bbfan> Hi there! Can anybody help me? I have an issue with ffmpeg installation [04:59:13 CET] <Guiri> I'm following this guide to compile ffmpeg on Ubuntu 14.04, but pkg-config can't seem to find librtmp-dev, which I installed: https://trac.ffmpeg.org/wiki/CompilationGuide/Ubuntu Is there a way to add it to my path or pkgconfig? [05:56:15 CET] <tombbraindamage> Hi I'm tying to hardcode subtitles from a MKV, I tried: ffmpeg -i video.mkv -vf subtitles=video.mkv out.avi from the official docs but this doesn't work [05:59:20 CET] <tombbraindamage> version 2.5.10 and no, they are text-based subs. [06:25:24 CET] <tombbraindamage> Its an internal ASS sub: but the idea is the same: http://pastebin.com/mk2s4GcT [06:26:46 CET] <relaxed> tombbraindamage: your ffmpeg may be too old. Try, http://johnvansickle.com/ffmpeg/ [06:36:25 CET] <relaxed> tombbraindamage: for higher quality video add -q:v 3 after the input [06:51:59 CET] <mzbotr> hi, I would like to know how to compute the filesize of a mpeg2 file converted into a DVD element... how could I do this? [06:52:29 CET] <mzbotr> i'm using dvdauthor to make the DVD tree, I'm writing a script to generate XML files from a list. the script needs to know how many files will fit into the disc project. [07:00:57 CET] <kepstin> mzbotr: you do it the other way around; you figure out how many files you want, then use that to determine the file sizes to use and then set the bitrate to get the desired size [07:01:11 CET] <kepstin> it's just a matter of picking tradeoff of amount of video vs quality [07:01:33 CET] <kepstin> (although dvd does have a max bitrate, so there is a point past which the files stop getting bigger) [10:17:15 CET] <Ccdc_DuckZ> 'morning [10:31:33 CET] <Ccdc_DuckZ> I'm checking the docs, but I can't see how to convert the picture format from one type to the other [10:31:40 CET] <Ccdc_DuckZ> anybody's got a hint? [10:48:45 CET] <durandal_1707> Ccdc_DuckZ: -pix_fmt [10:58:23 CET] <BtbN> format as in jpeg to png? [11:23:37 CET] <Ccdc_DuckZ> BtbN: no, as in whatever colour format I get from loading a png/jpg to paletted 8 bit (I'm trying to join individual files into an animated gif) [11:25:00 CET] <BtbN> there is a filter that generates an optimal gif palette from a given set of input images [12:16:15 CET] <momomo> guys, what is the best way to serve a hls stream to the client? [12:16:29 CET] <momomo> there are so many overkill solutions it appears [12:16:48 CET] <momomo> wowza and red5 which both costs money and comes with tousand tools [12:17:00 CET] <momomo> i am on java [13:41:57 CET] <DHE> momomo: I'm using nginx (apache also works) and ffmpeg [13:42:19 CET] <DHE> ffmpeg can do the whole thing. you just need to specify the right command-line. mine is fairly complex [14:04:03 CET] <momomo> DHE, so if I use nginx, it comes as a default configuration option? [14:06:37 CET] <momomo> I am totally at loss on how to proceed and there is little info online on the various options that one needs to use. I will look into nginx now .. I am using a simple ffmpeg command line right now to convert a strream into hls ... [14:06:51 CET] <DHE> no no no. static content service [14:06:53 CET] <momomo> ffmpeg -i http://url.path/stream/channel/fdccb829c6e768d4643ae226c6e3b282 out.m3u8 [14:07:12 CET] <momomo> static content service ? [14:07:25 CET] <DHE> Live content or pre-rendered movies? [14:07:29 CET] <momomo> live content [14:08:43 CET] <DHE> So something like: ffmpeg -i [source] [codec-options] -crf 23 -f hls -hls_list_size 4 -hls_time 5 -hls_flags delete_segments /var/www/html/out.m3u8 [14:08:52 CET] <DHE> oops, -crf 23 is a codec-option [14:11:25 CET] <momomo> this will generate the hls files, right? and possibly delete them when they get old? ... but I am guessing it does not resolve the issue of serving them to the client [14:11:26 CET] <momomo> ? [14:11:47 CET] <furq> any webserver can serve them to the client [14:11:53 CET] <DHE> as I said, nginx or apache. On my centos server /var/www/html is the default root for the default web host [14:12:00 CET] <furq> the main reason to use nginx is if you want an rtmp fallback [14:12:00 CET] <DHE> (for apache) [14:12:12 CET] <furq> nginx has an rtmp module which will also automatically mux and serve hls for you [14:12:23 CET] <DHE> my main reason is that nginx just scales better if you're serving to slow clients (eg: cell phones) [14:12:42 CET] <furq> i meant nginx-rtmp over ffmpeg + any httpd [14:13:07 CET] <DHE> yes, but if you're using HLS and ffmpeg then you don't really need nginx-rtmp [14:13:11 CET] <momomo> ook ... so I install nginx then this rtmp module and I place the files in /var/www/html/ [14:13:33 CET] <momomo> ook forget about that module then [14:14:03 CET] <momomo> but wait ... nginx + rtmp and there is no need for ffmpeg ? [14:14:12 CET] <furq> no there is a need for ffmpeg [14:14:24 CET] <DHE> no, you still need something to process and feed the video. nginx then becomes the distributor [14:14:36 CET] <furq> but instead of having ffmpeg output hls, you have ffmpeg output to rtmp [14:15:05 CET] <DHE> HLS' main advantage is being HTTP-based, allowing you to use standard software. bigger companies will use services like Akamai for distribution if they will be doing it on a large scale [14:15:36 CET] <furq> momomo: if you decide to go the nginx-rtmp route then https://gist.github.com/fur-q/d7028f51c38f7d0bb56e [14:15:43 CET] <furq> you can ignore the lua bit, i just use that for auth [14:16:24 CET] <momomo> holy crap ... i think I am over my head here ... you guys seem to know this pretty well [14:16:50 CET] <furq> like i said, that's mostly useful if you want a flash fallback [14:17:18 CET] <furq> `ffmpeg -i src /var/www/out.m3u8` will give you a directory full of hls stuff that you can serve however you want [14:17:34 CET] <furq> or probably /var/www/hls/out.m3u8 [14:17:34 CET] <DHE> momomo: in all seriousness, did you try the comamnd I gave you? [14:17:48 CET] <furq> oh yeah DHE has a better command [14:18:15 CET] <DHE> for out-of-my-ass codec options, -c:v libx264 -crf 23 -c:a aac -b:a 128k -ac 2 [14:21:59 CET] <momomo> DHE, i think the original one was lacking something [14:22:44 CET] <momomo> this is what I have: [14:22:52 CET] <momomo> ffmpeg -i http://theurl.com/stream/channel/fdccb829c6e768d4643ae226c6e3b288 [codec-options] -crf 23 -f hls -hls_list_size 4 -hls_time 5 -hls_flags delete_segments /var/www/html/out.m3u8 [14:23:01 CET] <momomo> what to replace [codec-options] with? [14:23:13 CET] <DHE> <DHE> for out-of-my-ass codec options, -c:v libx264 -crf 23 -c:a aac -b:a 128k -ac 2 [14:24:20 CET] <momomo> ffmpeg -i http://85.2132321.9.206:9981/stream/channel/fdccb829c6e768d4643ae226c6e3b288 -c:v libx264 -crf 23 -c:a aac -b:a 128k -ac 2 -f hls -hls_list_size 4 -hls_time 5 -hls_flags delete_segments /momomo/Generated/Tv/aaa.m3u8 [14:24:43 CET] <momomo> resulted in: The encoder 'aac' is experimental but experimental codecs are not enabled, add '-strict -2' if you want to use it. [14:24:43 CET] <momomo> [aac @ 0x21f6aa0] Alternatively use the non experimental encoder 'libvo_aacenc'. [14:25:31 CET] <DHE> ah, yeah. if libvo_aacenc works in place of aac then try that instead. otherwise add "-strict -2" before the output filename [14:28:13 CET] <momomo> DHE, yes, libvo_aacenc seesm to worked [14:29:34 CET] <momomo> I have to figure out some day what this command means and it's options [14:29:55 CET] <momomo> i like that it is deleting old files [14:29:57 CET] <momomo> great! [14:30:07 CET] <momomo> now I have to install nginx [14:48:09 CET] <momomo> this is my nginx config file: http://hastebin.com/ebadamizox.nginx [14:48:26 CET] <momomo> it's the default one i believe [14:55:01 CET] <momomo> DHE, is it <video height="100%" autoplay controls> [14:55:01 CET] <momomo> <source src="http://localhost/out.m3u8"> [14:55:01 CET] <momomo> </video> then for nginx ? provided that the files is in /var/www/html ? [14:56:40 CET] <DHE> and the nginx.conf says "root /var/www/html;" [14:57:17 CET] <DHE> but I can't help on the web browser front. I deal with cellphones and apps therein [14:58:52 CET] <momomo> DHE, yes, I get that .. [14:59:08 CET] <momomo> did you see my nginx conf file ? http://hastebin.com/ebadamizox.nginx [14:59:18 CET] <momomo> it seems to state a different path for root [14:59:28 CET] <momomo> for user [14:59:38 CET] <momomo> i can't see root [15:01:08 CET] <DHE> I can't. That site won't work without javascript [15:01:27 CET] <momomo> ook [15:02:43 CET] <momomo> here is a pastebin, maybe better: http://pastebin.com/RQyqgKNu [15:03:01 CET] <momomo> raw: http://pastebin.com/raw/RQyqgKNu [15:09:56 CET] <DHE> follow the includes. /etc/nginx/conf.d/*.conf and /etc/nginx/sites-enabled/* [15:10:18 CET] <DHE> if you have apache experience maybe you should stick to that [15:13:48 CET] <momomo> unfortunately not much more .. i am more from a javaee background ... not had much use of nginx or apache there [17:04:43 CET] <shincodex> avcodec_find_decoder... Do i need to free the return value if not null? [17:09:54 CET] <kepstin> shincodex: I don't *think* so, I think the AVCodec structures is statically allocated. Should get someone else to confirm tho. [17:14:31 CET] <kepstin> (you also shouldn't ever just call 'free' on something from ffmpeg, but rather use the 'avwhatever_unref_blah' or 'avwhatever_free_blah' functions as appropriate) [17:26:56 CET] <m3gab0y> ffmpeg stream @ 500 kbps on reasonable SD quality in 1 pass? [17:27:17 CET] <shincodex> Well looking in find_decoder i dont see a reference to malloc at first glance. [17:27:26 CET] <shincodex> When I say free I mean what you say. [17:28:41 CET] <m3gab0y> I need to make a similar quality stream http://www.flashvideofactory.com/test/DEMO720_Heima_H264_500K.html [17:31:52 CET] Action: DHE rips the video [17:36:13 CET] <J_Darnley> m3gab0y: depends on the content [17:36:32 CET] <J_Darnley> low motion video like your example will probably be fine [17:36:44 CET] <J_Darnley> or at least "tolerable" [17:37:10 CET] <J_Darnley> Although those fast cuts are now starting to hurt quality [17:37:32 CET] <J_Darnley> As I said: depends on the content [17:39:51 CET] <explodes> Lets say my video is at 10 seconds and I seek to 4 minutes. I will sometimes keep decoding a few frames from the 10 second mark before I get to 4 minutes [17:40:09 CET] <explodes> Is there some way to abort any frames that are currently being decoded? [17:40:55 CET] <J_Darnley> I assume with threads and/or B-frames that you will need to flush the decoder in some manner. [17:41:17 CET] <explodes> Cool, do you know off the top of your head a function or family of functions that does just that? [17:41:43 CET] <J_Darnley> No [17:42:17 CET] <J_Darnley> Don't you know somehow which are "old" frames, using pts or something? [17:43:47 CET] <explodes> sort of. I could discard frames where the pts is out of a certain range [17:46:24 CET] <explodes> that feels dirty and prone to errors [17:51:18 CET] <ValdikSS> Can anyone please give attention to this? http://article.gmane.org/gmane.comp.video.ffmpeg.user/61055 [18:26:41 CET] <mzbotr> are any SoC on the market good for video transcoding? [18:27:25 CET] <mzbotr> i want to use v4l2 and a uvc dongle to grab some VHS and hopefully encode the videos [18:34:56 CET] <kepstin> usb2 is fast enough to do raw video at SD rates, and software encoders running on the pc would be better quality... [18:35:13 CET] <kepstin> not that "quality" matters with vhs, but still :) [18:50:15 CET] <mzbotr> kepstin: maybe use gstreamer to transport the raw video over the network to ffmpeg running elsewhere? no weird hardware artifacts should show up that way, right? [19:09:19 CET] <derekprestegard> with vhs the vtr and the capture card are by far the most important [19:09:36 CET] <derekprestegard> I just got a whole VHS capture rig at home :D [20:28:37 CET] <podman> Is it possible to generate a progressive jpeg using ffmpeg? [20:33:37 CET] <durandal21707> podman: afaik nope [20:34:16 CET] <podman> durandal21707: cool, i'll just use imagemagick then [20:51:20 CET] <lmat> It looks like what we need is a format that exactly mirrors blu-ray capabilities (menus and other features), but doesn't have encryption. [20:51:38 CET] <lmat> So that blu-rays can be decrypted, but lose none of their features (e.g. subtitles, angles, etc.) [20:51:42 CET] <lmat> Does something like that exist? [20:52:05 CET] <BtbN> no [20:52:12 CET] <BtbN> Not even for DVDs [20:52:32 CET] <J_Darnley> Who wants a goddamn menu in their video? [20:53:21 CET] <podman> lol [20:58:08 CET] <DHE> one that players will enforce you watch through in its entirety before you can actually click "play movie"... which is followed by an outro of course [20:58:57 CET] <lmat> By menu, I mean options, like a scene index, etc. [20:59:05 CET] <lmat> not necessarily an animated menu ^_^ [20:59:26 CET] <JEEB> that's just called chapters? [20:59:34 CET] <JEEB> which your player lets you access [20:59:34 CET] Action: J_Darnley was bout to say that. [20:59:49 CET] <J_Darnley> My player has a menu for direct scene access [21:00:02 CET] <lmat> So, if I rip one of my Blu-rays, I can access the scene index, etc., if I find the correct chapter? [21:00:29 CET] <JEEB> blu-rays contain the timestamps at least, pretty sure the "names" aren't there [21:00:42 CET] <JEEB> so as long as you rip correctly you will have playlists that contain chapters [21:00:46 CET] <lmat> And in what format is a ripped blu-ray? [21:00:51 CET] <JEEB> uhh [21:01:03 CET] <JEEB> a rip of a blu-ray by definition is the same thing that's on the disc just without the DRM [21:01:19 CET] <JEEB> a remux or re-encode is something that is often done out of rips and often (incorrectly) called rips [21:02:01 CET] <JEEB> on blu-rays there are special playlist files that contain chapters, which are the things the menus actually play [21:02:17 CET] <lmat> JEEB: Ahh, I thought that blu-ray (by definition) requires all the AACS or BD+... or both [21:02:39 CET] <lmat> "done out of rips" what does that mean? [21:03:01 CET] <JEEB> a rip is the same data without the copy protection, and that accessible data can be used in various ways [21:05:28 CET] <lmat> What software exists to create a bluray ISO without DRM from an original bluray? [21:05:40 CET] <lmat> I thought that blurays are converted to avi or mkv or some other "plain" video format. [21:05:51 CET] <JEEB> they are in "plain" video format on the disc already (MPEG-TS) [21:06:11 CET] <TD-Linux> there were some people trying to put menus into matroska at one point [21:06:13 CET] <JEEB> you then have playlists and other additional things of which the other parts consist [21:06:20 CET] <TD-Linux> and there are horrors like MMTP [21:06:25 CET] <JEEB> TD-Linux: looking at matroska-devel they're still trying... [21:06:34 CET] <JEEB> someone just brought it up as the standardization effort is picking up [21:06:52 CET] <TD-Linux> JEEB, yeah I know, I doubt it will make it in though just because of the work required [21:06:58 CET] <JEEB> yeah [21:07:52 CET] <JEEB> lmat: anyways there's not really a shortage of software capable of removing DRM from blu-rays, although due to the crypto being somewhat better than with DVDs it's a constant cat and mouse game and "oracle" based solutions tend to support things quicker [21:08:20 CET] <JEEB> open source there's libaacs and such [21:08:27 CET] <JEEB> (requires a database of keys) [21:08:50 CET] <lmat> Yes, I have read a good deal about the encryption and "cat and mouse", etc. https://wiki.archlinux.org/index.php/Blu-ray [21:09:07 CET] <lmat> JEEB: So libaacs is used (with the database of keys) to create a new bluray ISO without the DRM? [21:09:11 CET] <JEEB> uhh [21:09:28 CET] <JEEB> I have no idea why you specifically mention ISO, I mean it's just directories on the disc [21:09:45 CET] <JEEB> so the usual ways of getting the data from the disc is basically a copy of that directory structure [21:09:50 CET] <JEEB> you can make an ISO of it, sure [21:10:07 CET] <lmat> I see [21:10:20 CET] <JEEB> then there's anydvd hd, dvdfab hd and makemkv (last of which also has a mode for a simple rip mode, unlike its name) [21:10:32 CET] <lmat> JEEB: So mount /dev/bluray1 /mnt/bluray; cp -r /mnt/bluray ~/myFirstMovie; ? [21:10:39 CET] <JEEB> yup [21:11:03 CET] <lmat> So the magic is in mount? It detects that the source device is bluray, and automagically invokes libaacs? [21:11:10 CET] <JEEB> no [21:11:22 CET] <JEEB> on linux you have to use software that supports it and you have to have the keys around [21:11:36 CET] <lmat> JEEB: oh, so that mount then cp command chain doesn't work? [21:11:37 CET] <JEEB> anydvd hd is IIRC the only thing that has a transparent mode [21:11:41 CET] <JEEB> and that's on windows [21:11:59 CET] <lmat> JEEB: Is transparent mode the mode where the files are copied "straight"? [21:12:15 CET] <JEEB> transparent mode is just transparent decryption while you access the drive [21:12:27 CET] <JEEB> so like your mount point example [21:12:41 CET] <lmat> Oh, right [21:12:59 CET] <lmat> JEEB: I don't care about that . just as good would be mkv /dev/bluray ~/myFirstMovie; ^_^ [21:13:19 CET] <lmat> or vlc -i /dev/bluray -o ~/myFirstMovie; (I'm making up syntax here) [21:13:31 CET] <JEEB> the vlc one is closest since vlc has support for libaacs [21:13:33 CET] <||arifaX> http://www.howtogeek.com/240636/everything-you-need-to-know-about-watching-… [21:13:46 CET] <JEEB> that said, you still need the key database [21:14:07 CET] <lmat> JEEB: Of course, that's trivial. [21:14:12 CET] <lmat> ||arifaX: Thanks [21:14:22 CET] <lmat> Aww, what a cute penguin [21:16:53 CET] <lmat> That is *not* everything I need to know.. [21:17:32 CET] <lmat> I guess I should go get a bluray player and get started on this stuff. [21:17:55 CET] <lmat> It looks like there are 3d titles on bluray, but not DVD. [21:18:05 CET] <lmat> Is that because the resolution would be unacceptably low on DVD/ [21:18:21 CET] <J_Darnley> Of course. DVD is a better format! [21:18:32 CET] <J_Darnley> Poorer encryption. [21:18:39 CET] <J_Darnley> No 3D [21:21:44 CET] <DHE> this stuff needs to be standardized. DVD spec is ~704x480 max resolution, and about 9800 kilobit MPEG-2 for optimal quality [21:22:04 CET] <DHE> and think pre-Y2K [21:22:27 CET] <lmat> ^_^ [21:22:33 CET] <J_Darnley> The good old days [23:13:04 CET] <VelusUniverseSys> hello does the ffserver take a lot of cpu and ram? [23:13:44 CET] <JEEB> even if it doesn't, you still won't get much support for it if at all [23:13:57 CET] <JEEB> so I'd recommend looking at a solution that uses ffmpeg itself [23:14:05 CET] <JEEB> (and some streaming server if required) [23:17:07 CET] <DHE> ffserver is a hack. if anything else is available, we suggest using that instead. [23:19:51 CET] <VelusUniverseSys> the only thing i can find well it dosnt work proeperly lol im thinking of changing things around now lol [23:51:59 CET] <mattf000> quick question.. for a specific assignment i know that the player will always be coming from an armv7 neon android. will i see any difference in performance if install 32bit or 64bit ubuntu? [23:59:24 CET] <mattf000> should i be setting the -cpuflags to armv7a? since i'm transcoding on an intel? [23:59:48 CET] <J_Darnley> What? [00:00:00 CET] --- Sat Feb 13 2016

1 0

ffmpeg-devel.log.20160211
by burek 12 Feb '16

12 Feb '16

[00:26:14 CET] <llogan> excellent. more users asking *-owner for help. [01:02:18 CET] <cone-566> ffmpeg 03Rostislav Pehlivanov 07master:ac5a4deb3dd0: MAINTAINERS: add my name to the parts I'm maintaining [01:19:34 CET] <JEEB> was anyone working on a mediacodec hwaccel/avcodec lately? [01:42:21 CET] <michaelni> durandal_1707, "avfilter/dualinput: use pts provided by framesync" breaks fate-sub2video [01:44:38 CET] <michaelni> durandal_17: http://fate.ffmpeg.org/report.cgi?time=20160209134651&slot=x86_32-debian-kf… [01:56:26 CET] <michaelni> durandal_17, i guess it just needs a update to the reference [02:12:56 CET] <cone-566> ffmpeg 03Michael Niedermayer 07master:8836b8c2a430: fate: Update fate-sub2video for 5486d7fa91f7233e6d9aa9ba8de7fe9287402637 [02:32:07 CET] <cone-566> ffmpeg 03Michael Niedermayer 07master:ac3a275d1e0d: avfilter/vf_nnedi: Simplify away some trivial mem*() [07:15:48 CET] <tmm1> ubitux: finally got around to implementing charsets in the ccaption decoder [10:35:50 CET] <wm4> av_assert0(av_buffer_get_ref_count(sl->ref_list[list][index].parent->f->buf[0]) > 0); [10:35:52 CET] <wm4> wtf [10:36:08 CET] <wm4> (this crashes for me... but possibly not as assert, but as segfault) [10:38:39 CET] <nevcairiel> buf[0] is NULL? [10:38:54 CET] <nevcairiel> one shall not forget to set the buf [10:39:06 CET] <wm4> it happens when videotoolbox returns an error [10:39:21 CET] <wm4> it plays weird tricks with buf, so maybe that's the cause [10:39:31 CET] <wm4> I've looked at it before and concluded it's fucked [10:40:36 CET] <nevcairiel> why would it do such things [10:40:58 CET] <nevcairiel> let me guess, its because of this async crap? [10:41:21 CET] <wm4> no [10:41:30 CET] <wm4> because videotoolbox allocates a surface [10:41:37 CET] <wm4> instead of the API user/get_buffer [10:42:03 CET] <wm4> it sets a dummy buffer in ff_videotoolbox_alloc_frame [10:42:17 CET] <wm4> and the proper one in ff_videotoolbox_buffer_create [10:42:36 CET] <nevcairiel> sounds evil [10:42:37 CET] <wm4> maybe it shouldn't free the dummy buffer on error and leave buf NULL [10:42:49 CET] <wm4> I think videotoolbox shouldn't be a hwaccel at all [10:48:09 CET] <wm4> maybe it was caused by 9747219958060d8c4f697df62e7f172c2a77e6c7, my previous attempt to deal with this crap [10:49:15 CET] <nevcairiel> personally I would oppose any further h264*.c hacks to make it not crash on these hacky buffers [10:49:31 CET] <jkqxz> What if you change _alloc_frame to make buf[0] be empty but sized for sizeof(vtctx->frame), then just copy the right thing into place in _buffer_create? [10:49:52 CET] <nevcairiel> changing the buffer around later is high grade evil [10:50:19 CET] <wm4> nevcairiel: is the crashing code correct in the first place? [10:50:28 CET] <nevcairiel> sure [10:50:51 CET] <nevcairiel> why wouldnt it be [10:50:59 CET] <wm4> jkqxz: leaving the buf set on error will make it think the frame was decoded correctly, and return it to the user [10:52:59 CET] <jkqxz> What error is going to cause trouble here? [10:54:03 CET] <wm4> if hwaccel->end_frame returns failure [10:54:25 CET] <nevcairiel> end_frame failure should result in the decoder bailing out, shouldnt it [10:54:44 CET] <wm4> well it crashes the next time it enters the decoder [10:54:56 CET] <nevcairiel> although due to re-ordering it will probably still have said broken frame somewhere in the DPB [10:55:09 CET] <nevcairiel> this is why we have a central alloc function [10:55:18 CET] <nevcairiel> that one can return a failure, and then the frame is not touched again [10:55:26 CET] <nevcairiel> why vt doesnt use that is not something i want to understand [10:55:40 CET] <nevcairiel> ie. i consider it broken by design and will never use it [10:55:43 CET] <wm4> it does use it? [10:55:53 CET] <nevcairiel> but not for the actual frame allocation [10:55:54 CET] <wm4> it just returns a dummy [10:55:57 CET] <nevcairiel> just forsome dummy that never fails [10:56:15 CET] <nevcairiel> the actual allocation fails much much later, when avcodec assumes your frame is already allocated and all is fine [10:56:54 CET] <nevcairiel> ie broken by design =p [10:58:34 CET] <wm4> wtf the frame is returned by a callback [10:58:41 CET] <wm4> even though it's asynchronous [10:58:43 CET] <wm4> err [10:58:45 CET] <wm4> even though it's synchronous [10:58:52 CET] <jkqxz> Right, the actual allocation happens in end_frame from the asynchronous callback you're waiting for iff decoding succeeds. So if you get an error there you're just screwed. [10:59:19 CET] <wm4> indeed [10:59:47 CET] <nevcairiel> interestingly our mac users have not yet really asked for hwaccel decoding, so i'm staying away from that subject [10:59:54 CET] <jkqxz> end_frame could always make an empty frame to stop it crashing? [11:00:17 CET] <nevcairiel> there is no 100% guarantee you can allocate an empty frame if other things failed [11:00:19 CET] <wm4> jkqxz: an empty frame is invalid [11:00:20 CET] <jkqxz> On error, that is. (Also that needs to not be able to fail.) [11:00:31 CET] <wm4> nevcairiel: you could preallocate it [11:00:45 CET] <wm4> jkqxz: unless you mean a real frame... but you don't know the format [11:01:01 CET] <jkqxz> What do you mean by "invalid"? I was meaning a real frame with (possibly with zeroes in it). [11:01:18 CET] <wm4> anyway, I feel like this should be a real decoder, not a hwaccel; I'm just not sure how to use it [11:01:41 CET] <nevcairiel> definitely sounds like it doesnt really fit the hwaccel model [11:02:09 CET] <wm4> nevcairiel: yet it can deal with slice input... [11:02:27 CET] <nevcairiel> the real question is if it can also deal with full bitstream input without a lot of user magic [11:02:39 CET] <nevcairiel> if yes then making it a decoder would probably be easy [11:02:41 CET] <nevcairiel> if no, good luck [11:02:55 CET] <BtbN> doesn't the hwaccel for vt reconstruct the full bitstream? I remember some crazy hack like that somewhere. [11:03:28 CET] <wm4> it concatenates the data it gets from decode_slice [11:03:47 CET] <nevcairiel> well most hwaccels do that really in some form or another [11:13:13 CET] <jkqxz> Given that it takes NAL units rather than bytestream, what even happens if you only give it (slices forming) half a frame? [11:14:07 CET] <wm4> it's quite eager to return decoding errors instead of frames on broken streams [11:14:46 CET] <jkqxz> (I assume it can't return an error from the first submission, because that would entail reading the entire bitstream and defeats the point of asynchronicity, and it can't give you a callback yet because you might be giving it more NAL units later.) [11:15:17 CET] <wm4> not sure how it works in the async case [11:16:02 CET] <wm4> seems it's more like input is synchronous, while output is asynchronous [11:16:50 CET] <mateo`> Hello guys, speaking of hw accel api, i'm still working on the mediacodec stuff; got something working for h264 and android >= 4 (in "production" at work). I still need to clean a lot of stuff related to jni~ mostly. I decided to support android 4 as there are still too many of that kind of devices in the wild, meaning there are jni crap involved (even if I've only supported android 5 and using the C [11:16:52 CET] <mateo`> api, it would have involved some jni stuff as the C api lacks some features like having access to the decoder name which is mandatory to work around crappy implementations). [11:17:22 CET] <wm4> heh, it's probably unavoidable [11:21:43 CET] <jkqxz> ((H.264 slices would have been much more helpful for weird hardware cases if the header also included last_mb_in_slice or similar. I've been annoyed by this in the past.)) [13:10:47 CET] <atomnuker> j-b: did you know that installing a single KDE application (e.g. Kate) on debian without any KDE stuff will also install VLC [13:11:46 CET] <atomnuker> it's also mandatory so no amount of --no-install-recommends will save you from having VLC if you just wanted a nice text editor [13:12:25 CET] <atomnuker> VLC is truly everywhere now [13:12:51 CET] <nevcairiel> i would probably go troll debian about their silly dependency setup [13:13:02 CET] <durandal21707> does VLC supports utvideo? [13:15:19 CET] <atomnuker> nevcairiel: blame the KDE Phonon guys who insist on having libVLC as the "best" backend [13:15:23 CET] <atomnuker> and all they ever do with it [13:15:29 CET] <atomnuker> is play wav files [13:15:35 CET] <nevcairiel> phonon is such a crapshow [13:15:43 CET] <nevcairiel> its like gstreamer for kde [13:15:43 CET] <nevcairiel> =p [13:16:11 CET] <atomnuker> gstreamer is at least a more okayish dependency because many music players (e.g. quodlibet) depend on it [13:16:49 CET] <atomnuker> but still phonon could basically be rewritten in 1 line: "paplay <file>" [13:17:16 CET] <wm4> so --enable-memalign-hack is cargo-culting these days? [13:17:38 CET] <nevcairiel> yes [13:17:46 CET] <nevcairiel> configure should auto-enable it when needed [13:17:50 CET] <nevcairiel> when it doesnt, file a bug [13:18:11 CET] <wm4> not aware of such a situation (unless it can happen on windows) [13:20:16 CET] <nevcairiel> the OS provided functions probably just do something similar to the hack [13:20:22 CET] <nevcairiel> at least the mingw emulation of them does [13:20:47 CET] <nevcairiel> didnt bother to look into MS libc [13:29:17 CET] <J_Darnley> atomnuker: I would love to see how that dependancy chain works kate->???->vlc [13:30:16 CET] <atomnuker> ??? = KDE libs [13:30:34 CET] <wm4> kde depends on stuff just for the heck of it [13:30:46 CET] <wm4> there's even some code that makes _all_ kde apps exit if there's no dbus running [13:30:53 CET] <wm4> (comment it and you can run most stuff without dbus) [13:32:03 CET] <atomnuker> Daemon404: output.flac - 49576; output.wv - 49292; highest compression level for both [13:32:06 CET] <atomnuker> pretty decent [13:34:37 CET] <atomnuker> though I still one day plan to write a lossless codec with so much prediction, vector quantization system and real entropy encoding [13:35:00 CET] <atomnuker> just need to find <alien technology> to use to encode the per coefficient errors [13:41:23 CET] <nevcairiel> if its as fast as the flac alien algorithm mode, then good luck =p [13:42:09 CET] <atomnuker> the flac alien mode did result in worse compression actually [13:42:41 CET] <nevcairiel> did you borrow a supercomputer or how did you find out? :) [13:42:58 CET] <atomnuker> I let it run overnight to encode a 1 second file [13:43:13 CET] <atomnuker> was disappointing to see it being like 60ish bytes larger [13:43:51 CET] <nevcairiel> and people say vp9 is slow [13:44:27 CET] <atomnuker> it really is slow, but they've mislabed the quality setting [13:44:32 CET] <atomnuker> normal = normal but fast [13:44:43 CET] <atomnuker> good = SUPER PLACEBO ULTRA SLOW MODE [13:45:49 CET] <atomnuker> that's probably how they earn their ultra high PSNR score (but probably at the cost of visual quality) [13:45:53 CET] <nevcairiel> its "good" and "best", but yes [13:46:08 CET] <nevcairiel> everyone including the devs will generally recommend against using best [13:46:35 CET] <nevcairiel> in other news, Microsoft made the format of their pdb debug information open source on GitHub [13:46:43 CET] <nevcairiel> hopefully more tool support now [13:50:04 CET] <wm4> nevcairiel: wasn't that a while ago [13:50:55 CET] <nevcairiel> they started in october or something, but only recently did they push a whole lot of the missing info [13:52:17 CET] <nevcairiel> would be awesome if clang and/or gcc could eventually generate pdb compatible debug info [13:52:39 CET] <wm4> they did this for clang, didn't they [13:52:43 CET] <nevcairiel> yes [13:52:53 CET] <nevcairiel> so clang will likely be the first [13:53:27 CET] <wm4> as long as the clang dev working on this doesn't kill himself (must be some quite terrible fiddling...) [13:54:01 CET] <nevcairiel> microsoft seems to be quite willing to help, so its probably not too bad [13:54:13 CET] <nevcairiel> getting rid of mingw would be a day to celebrate [13:54:30 CET] <wm4> indeed [13:55:26 CET] <nevcairiel> that reminds me that I wanted to investigate why msvc builds crash in hevc sometimes [13:55:33 CET] <nevcairiel> oddly enough not on any fate samples [14:14:23 CET] <wm4> wow, if you seek to the start of an avi file, the first packet timestamp you get is different from when you start an avi file without seek? [14:14:26 CET] <wm4> how does this work [14:17:25 CET] <nevcairiel> magic [14:17:39 CET] <nevcairiel> avi has no timestamps, so possibly a bug [14:17:57 CET] <wm4> it's an avi with h264 and b-frames [14:18:04 CET] <wm4> so lots of crap could happen [14:18:28 CET] <wm4> and why can't ffprobe seek [14:20:29 CET] <wm4> actually it can [14:20:58 CET] <wm4> and it confirms my findings [14:21:14 CET] <wm4> a seek offsets it by 2 frames [14:21:24 CET] <wm4> with is also the b-frame delay in this file [14:22:11 CET] <wm4> encoder=MEncoder SVN-r23784 [14:26:28 CET] <wm4> /* XXX: How to handle B-frames in AVI? */ [14:26:29 CET] <wm4> hmmm [14:28:16 CET] <Daemon404> atomnuker, there are closed source lossless audio codecs which compress better than either [14:28:25 CET] <Daemon404> nobody uses/re'd them really though [14:31:03 CET] <Compn> wm4 : h264 in avi is invalid [14:31:13 CET] <Compn> double invalid when made with mencoder... [14:31:17 CET] <BBB> atomnuker: omg I cant believe you tried that [14:32:03 CET] <Compn> wm4 : it only plays back correctly (e.g. b-frames are in correct order) because of ffmpeg magic. [14:32:23 CET] <Compn> some hardware decoders will play it out of order and it jerks back and forth [14:32:28 CET] <wm4> Compn: doens't matter, thanks to ffmpeg eating this crap (and doing it incorrectly), I'll have to deal with it anyway [14:32:54 CET] <Compn> if h264 in avi > ret error "you suck" [14:32:56 CET] <Compn> ehe [14:33:25 CET] <Daemon404> there;s quite a bit of h264 in avi in the wild [14:33:45 CET] <Daemon404> during the xvid->x264 transition, there was a very large following of x264's vfw codec [14:33:56 CET] <Daemon404> i think it was even in x264's git at one point [14:34:09 CET] <Compn> nothing wrong with a vfw x264 encoder [14:34:28 CET] <Compn> mkv muxer can be added to dshow/vfw chain [14:34:32 CET] <Daemon404> "hitler did nothing wrong" - Compn, 2016 [14:35:00 CET] <Daemon404> [13:34] <@Compn> mkv muxer can be added to dshow/vfw chain <-- not vfw [14:35:05 CET] <Daemon404> unless you cound the vfw compat in mkv [14:35:11 CET] <Daemon404> which is basically an avi muxed into mkv [14:35:17 CET] <Daemon404> count* [14:35:39 CET] <wm4> the funniest thing is that vc-1 in mkv can be muxed only using vfw compat (i.e. it behaves like in avi) [14:35:44 CET] <wm4> even though vc-1 has b-frames [14:35:59 CET] <Compn> since x264 vfw is non standard, i'm counting ffdshow mkv output or modified vdub or other tool ,even avisynth or graphedit or mencoder [14:36:01 CET] <Daemon404> i was pretty sure mkvtoolnix had a way to mux normal vc1 [14:36:03 CET] <Daemon404> from an elementary stream [14:36:48 CET] <Daemon404> Compn, iirc modified vdub also output mkv with the vfw compat layer, graphedit doesnt do vfw, avisynth doesnt output anything directly, and ive never heard offdshow outputting directly [14:37:02 CET] <wm4> well neither ffmpeg's nor mpv's demuxer have such vc-1 support [14:37:21 CET] <Daemon404> wm4, chicken/egg im sure [14:37:37 CET] <Daemon404> more like chicken corpse / smashed egg [14:37:42 CET] <Daemon404> because vc1. [14:39:25 CET] <Compn> graphedit does vfw, wtf are you on about ? [14:39:32 CET] Action: Compn afk [14:40:53 CET] <Daemon404> Compn, by way of dshow wrapper [14:41:19 CET] <Daemon404> the vfw codec should still output the same (packed) packets [14:42:22 CET] <Daemon404> (and im not even sure if that still exists in modern windows versions) [14:47:47 CET] <rcombs> heh, AVERROR_GOFUCKYOURSELF [14:48:58 CET] <Daemon404> if we wanted to screw with people we could have AVERROR_PATENTSREQUIRED [14:49:35 CET] <Daemon404> based on timezone, like vlc did [14:49:36 CET] Action: Daemon404 runs [14:54:41 CET] <Daemon404> ... hevc hls? [14:57:05 CET] <wm4> hm I declare avidec.c indecipherable, I guess I'll just tell the users to fuck off [14:58:27 CET] <Daemon404> with those exact works? you're a nice guy, man. [14:59:05 CET] <wm4> "exact works" wat [14:59:52 CET] <Daemon404> words [15:00:12 CET] <Daemon404> you need to work on your ESP [15:11:27 CET] <durandal21707> Daemon404: going to REd MagicYUV? [15:17:23 CET] <endle> Excuse me, I'm reading ffmpeg's code now, and faced a problem [15:17:29 CET] <endle> The coverage page(http://coverage.ffmpeg.org/) shows that libswscale got 54.7% coverage, but I can't find the test code for it [15:17:53 CET] <endle> So where is the test code for libswscale? The code in libswscale are tested indirectly? [15:19:37 CET] <nevcairiel> swscale is tested indirectly through various other format conversion tests [15:19:51 CET] <Daemon404> i thought there was some filter tests that used it [15:19:54 CET] <Daemon404> kinda directly [15:20:05 CET] <nevcairiel> i suppose [15:20:10 CET] <nevcairiel> but those are through avfilter then [15:20:14 CET] <nevcairiel> not explicit swscale tests as such [15:20:37 CET] <endle> Thanks [15:21:03 CET] <durandal21707> michaelni: when is new release planned? who can make new releases? [15:24:16 CET] <endle> If I want to send my first patch to ffmpeg, is there any suggestion? [15:24:33 CET] <durandal21707> there is docs to read written just for you [15:24:57 CET] <endle> Let me have a check? [15:25:05 CET] <durandal21707> http://ffmpeg.org/developer.html#toc-Contributing [15:26:00 CET] <Daemon404> round v8 of mats replying to himself... [15:26:03 CET] <Daemon404> holy crap. [15:27:18 CET] <nevcairiel> he must never be bored, always someone to talk to [15:29:58 CET] <endle> I'm fixing a compile error in libswscale, and I'm afraid that I might break something without without direct test suite [15:30:46 CET] <durandal21707> well post patch [15:31:37 CET] <endle> not finished yet... Can I just send my patch to mail list, and claiming "Draft" in the title? [15:32:03 CET] <Daemon404> what are you fixing? [15:33:16 CET] <endle> libswscale/slice.c:178 and 184 [15:33:35 CET] <Daemon404> in english [15:33:36 CET] <endle> there is a cast from const pointer to non-const pointer [15:33:38 CET] <Daemon404> oh [15:34:25 CET] <Daemon404> i dont see any cats on those lines in HEAD [15:34:29 CET] <Daemon404> casts even [15:34:36 CET] <Daemon404> https://github.com/FFmpeg/FFmpeg/blob/master/libswscale/slice.c#L178 [15:35:06 CET] <endle> const uint8_t *src_[4] [15:35:27 CET] <Daemon404> implicit conversion then [15:35:33 CET] <endle> and s->plane[i].line is uint8_t * [15:35:49 CET] <endle> Yeah. That's why the compiler throws a warning [15:37:07 CET] <Daemon404> unrelated to your patch... [15:37:08 CET] <Daemon404> https://github.com/FFmpeg/FFmpeg/blob/master/libswscale/slice.c#L161 [15:37:13 CET] <Daemon404> why are these all stride[0] [15:37:29 CET] <Daemon404> that doesnt seem right at all, or just too 'clever' for me to get [15:40:22 CET] <durandal21707> Daemon404: looks like bug to me [15:40:22 CET] <Daemon404> the commit that added that seems to have never been reviewed on the ML either [15:40:27 CET] <wm4> later in the code: src_[i] + j * stride[i]; [15:40:32 CET] <Daemon404> yeah [15:40:35 CET] <Daemon404> it just seems wrong [15:40:39 CET] <wm4> so whatever it does [15:40:41 CET] <Daemon404> how could that ever work for subsampled chroma...? [15:41:16 CET] <wm4> also, what if there are less than 4 planes [15:41:19 CET] <Daemon404> https://github.com/FFmpeg/FFmpeg/commit/5bd62a1b3c3356b84818efca3fcaf52da27… [15:41:22 CET] <Daemon404> the commit in question [15:41:42 CET] <wm4> nice commit message [15:42:55 CET] <durandal21707> you can make your own conlusions what it does :) [15:43:18 CET] <Daemon404> michaelni, your knowledge is welcome here [15:43:25 CET] <Daemon404> you probably know swscale best, but it looks liek a bug. [15:50:07 CET] <endle> Thank you. Seems that libswscale is a little tricky for me to send my first patch [15:51:45 CET] <endle> Leaving for bed now. Bye [15:54:52 CET] <nevcairiel> Daemon404: slice-based scaling is practically unused and untested [15:55:24 CET] <Daemon404> i heard mplayer uses it [16:03:27 CET] <wm4> oh, that API [16:03:44 CET] <wm4> mplayer actually rarely uses the slice path (which it has spread all over its video chain) [16:03:53 CET] <wm4> because it doesn't work with modern things like h264 or threading [16:09:15 CET] <Daemon404> lol really? [16:09:39 CET] <Daemon404> or do you mean "work" like with sliced output [16:09:44 CET] <Daemon404> in draw_horiz_band [16:10:16 CET] <wm4> well in mplayer2 times I desperately tried to make it use the slice path (to test it), and couldn't trigger it [16:10:25 CET] <wm4> but yes, it works with draw_horiz_band [16:10:38 CET] <wm4> the idea is that decoding, filtering, and rendering works via slices [16:10:42 CET] <wm4> which should be more efficient [16:10:54 CET] <Daemon404> thats why utvideo uses slices [16:11:03 CET] <Daemon404> of course we used frame threading, eliminating its usecase [16:11:20 CET] <Daemon404> (i added sliced threads to ours once, and it was rejected due to being slower to decode linearly than frame threads) [16:11:46 CET] <Daemon404> (for NLE latency / seeking) [16:14:56 CET] <BBB> no response to my vp9/hwaccel/threading patch? [16:14:59 CET] <BBB> so sad [16:15:18 CET] <wm4> "it's awesome push it" [16:15:27 CET] <wm4> (there was a patch?) [16:16:19 CET] <BBB> ... [16:16:35 CET] <wm4> must have missed it [16:16:36 CET] <Daemon404> was it hidden in a giant thread [16:16:52 CET] <BBB> yes [16:17:01 CET] <Daemon404> thats why nobody commented [16:17:29 CET] <wm4> and I still can't find it [16:17:54 CET] <durandal21707> Daemon404: you added slice threads decoding to native utvideo? [16:18:08 CET] <Daemon404> at one point [16:18:14 CET] <Daemon404> the patch has since long been lost [16:18:18 CET] <Daemon404> it wasnt very hard to add [16:18:42 CET] <wm4> BBB: maybe resend it, and as a new thread [16:18:50 CET] <durandal21707> Daemon404: where it is? [16:19:08 CET] <Daemon404> i don't know, i don't keep track of patches i threw away 2 years ago [16:19:25 CET] <Daemon404> im not sure if it even made it to the ML; just irc [16:22:29 CET] <durandal21707> :( [16:31:51 CET] <j-b> atomnuker: yes. [16:33:20 CET] <JEEB> I might have the ut video slices patch... maybe. if my older laptop doesn't die when I touch it [16:34:54 CET] <cone-967> ffmpeg 03Tobias Rapp 07master:8b99c5e8daf2: avfilter/f_metadata: avoid float rounding problems [16:34:55 CET] <cone-967> ffmpeg 03Tobias Rapp 07master:202f97872891: avfilter/f_metadata: add support for file output [16:34:56 CET] <cone-967> ffmpeg 03Tobias Rapp 07master:6889deba68fc: avfilter/f_metadata: add starts_with string function [16:34:57 CET] <cone-967> ffmpeg 03Tobias Rapp 07master:730da5c20791: avfilter/f_metadata: whitespace clean-up [16:34:58 CET] <cone-967> ffmpeg 03Tobias Rapp 07master:80026a8ac35b: avfilter/f_metadata: rename "string" into "same_str" [16:34:59 CET] <cone-967> ffmpeg 03Paul B Mahol 07master:855d9d29ebf1: doc/filters: fix metadata example [16:39:28 CET] <durandal21707> anybody against loop filter, you would be able to set when to start looping and for how much? [16:40:32 CET] <wm4> seems like a kind of bad idea with the implied buffering, but it's not like anyone is going to be against it completely [16:43:36 CET] <durandal21707> default value would be some smallish number of frames to cache [16:44:05 CET] <wm4> doesn't it have to cache the loop duration? [16:45:02 CET] <durandal21707> no, frames received during looping are dropped [16:45:35 CET] <wm4> I mean the duration of the looped fragment [16:46:06 CET] <durandal21707> yes, that is cached [17:02:25 CET] <durandal21707> https://video.fosdem.org/2016/h2214/ :( [17:03:02 CET] <nevcairiel> why is fosdem so incompetent with video [17:03:47 CET] <Daemon404> nevcairiel, like i said before, theyre only giving us the FOSS UX [17:03:50 CET] <J_Darnley> Because video is not very free what with them patent things and all [17:04:11 CET] <Daemon404> they seem to have a single rpi transcoding everything [17:04:17 CET] <Daemon404> i have no other way to explain [17:04:18 CET] <JEEB> lol [17:04:24 CET] <J_Darnley> hue [17:04:26 CET] <JEEB> yeah, it does feel like that [17:04:36 CET] <J_Darnley> I'll loan them mine [17:04:41 CET] <J_Darnley> Twice the speed! [17:21:47 CET] <kierank> durandal21707: they've lost a lot of the video [17:21:49 CET] <kierank> that's why [17:22:00 CET] <JEEB> wow [17:22:32 CET] <Daemon404> kierank, ... how [17:22:41 CET] <wm4> maybe they deleted it out of embarrassment? [17:22:43 CET] <Daemon404> record to harddrive -> ??? -> profit [17:25:11 CET] <J_Darnley> Perhaps their wooden boxes burst into flames. [18:16:01 CET] <michaelni> durandal21707, release is planned ASAP, well i wanted to make it a week ago but too much random stuff kept me from workig on it [18:18:10 CET] <michaelni> and anyone who has an account on the server really can make a release i think, shouldnt need any special permissions [18:24:22 CET] <JEEB> mateo`: is your mediacodec stuff a decoder or hwaccel? (i think it supported both buffer and surface outputs) [18:26:08 CET] <mateo`> JEEB: it is both [18:26:19 CET] <JEEB> ok [18:27:02 CET] <JEEB> mediacodec is kind of interesting since we got libmpv to work with NDK this week [18:27:19 CET] <JEEB> so a nice opengl es video renderer [18:27:53 CET] <rcombs> the C API for it? [18:28:02 CET] <rcombs> mateo`: what did you do about the thread pool thing? [18:29:19 CET] <mateo`> rcombs: what are you refering to ? [18:29:50 CET] <mateo`> rcombs: i'm not using the C api in the end, i've done something kind of ~nasty [18:30:06 CET] <mateo`> i took the C api, and implemented it in FFmpeg with a jni backend [18:30:14 CET] <rcombs> yuck [18:30:22 CET] <mateo`> so it will be easiest to port to the C APi later on [18:30:31 CET] <mateo`> and it will work with all the android 4 devices [18:30:37 CET] <mateo`> yeah i know ... [18:30:57 CET] <rcombs> tl;dr android doesn't like that whole IPC model to work until the process has some thread pool running [18:31:15 CET] <rcombs> so unless you're using JNI (which you are, but later&) you need to do something to start it [18:31:24 CET] <rcombs> and apparently that thing is "call a private C++ API" [18:31:47 CET] <rcombs> someone should yell at the android guys about making that public and available to C [18:31:54 CET] <rcombs> or just doing it automatically [18:32:04 CET] <rcombs> if you try to use the IPC while it's not already running [18:33:47 CET] <mateo`> i did nothing about that ... i'm using all this stuff in the context of an application [18:33:56 CET] <mateo`> i was not even aware about this limitation [18:34:25 CET] <rcombs> yeah, the application already has it running; it comes from zygote [18:34:51 CET] <rcombs> but if you tried to execve (e.g. running ffmpeg.c) you'd need it [18:38:31 CET] <mateo`> What makes me decide to go with the jni stuff is that there is no API to access to codec name, and you still need it. In the first place I thought the MediaCodec implementations would have been improved with Android 5 but it's exactly the same thing when it output software buffers. [18:38:41 CET] <mateo`> OMX.SEC.avc.dec still lie about its stride [18:38:53 CET] <mateo`> and so forth... [18:39:56 CET] <mateo`> if you are interested I can put my work somewhere before sending the patch to the ml, so you can try it and flame me :) [18:46:46 CET] <JEEB> nice [18:47:13 CET] <JEEB> also lol at the JNI stuff not having access to codec name [18:55:20 CET] <mateo`> btw, awesome to hear that libmpv is running on the ndk :) [19:00:26 CET] <JEEB> started off as a joke and then before you knew it opengl-cb and openal worked :) [19:02:47 CET] <mateo`> rcombs: hello [19:02:53 CET] <mateo`> oops [20:52:31 CET] <furkan> in udp.c, could anybody explain the relationship between the buffer_size variable used to set SO_RECVBUF (UDP socket receive buffer), and the circular ("fifo"?) buffer? [20:53:09 CET] <Mavrik> One is an actual socket buffer you communicate to the OS. [20:53:21 CET] <Mavrik> It's where packets wait for the code in udp.c to read them an transfer them to the FIFO buffer. [20:53:34 CET] <furkan> Mavrik: ah i see [20:53:49 CET] <furkan> my kernel is dropping UDP packets so i'm trying to figure out how to increase the buffer size [20:54:19 CET] <furkan> netstat -s is reporting RcvBufErrors [20:54:30 CET] <Mavrik> that would be SO_RECVBUF then [20:55:09 CET] <furkan> it looks like i can set that with buffer_size [20:55:42 CET] <furkan> i think something like av_dict_set(&d, "buffer_size", x, 0); [20:55:54 CET] <furkan> trying to find out what kind of structure "x" is supposed to be [20:56:13 CET] <Mavrik> C string. [20:56:20 CET] <Mavrik> There's also av_dict_set_int. [20:56:37 CET] <furkan> great i'll give that a shot thanks [20:57:44 CET] <furkan> so is the circular buffer used for reordering RTP packets that are received out of order or something? [20:57:53 CET] <furkan> just trying to understand why the need for a second buffer [21:00:19 CET] <furkan> beautiful, increasing the buffer_size to 10M appears to have fixed my problem [21:03:54 CET] <nevcairiel> Daemon404: i dont suppose you know if clang/c2 is any smarter with stack alignment than gcc is? right now if I build a gcc build with -msse2 and call it from msvc it will just crash due to missing alignment, but building with -mstackrealign or -mincoming-stack-boundary=2 blows up in inline asm =p [21:04:18 CET] <kierank> furkan: the second buffer is a hack [21:04:21 CET] <nevcairiel> What I would really like would be an option to only apply the realignment to all publicly exported functions, but the compielr doesnt know which those are... [21:04:38 CET] <kierank> furkan: because otherwise ffmpeg receives udp packets in the same thread as everything else [21:04:46 CET] <J_Darnley> nevcairiel: there's a function attribute for that [21:04:51 CET] <kierank> it's not an application designed for streaming (like say vlc) [21:04:56 CET] <nevcairiel> J_Darnley: sure, but you want to flag every single public function? [21:05:13 CET] <furkan> kierank: ah interesting, i noticed that the circular buffer was only enabled if PTHREADS was enabled [21:05:19 CET] <J_Darnley> ah [21:05:21 CET] <nevcairiel> with -msse2 and auto vectorization it starts to use sse2 in all sorts of places [21:05:36 CET] <nevcairiel> kierank: that thread and buffer works quite nicely though [21:06:04 CET] <kierank> still a hack, best way to do it is with nonblocking io [21:06:07 CET] <J_Darnley> Wait. Don't we already mark public API with something? [21:06:10 CET] <kierank> and a separate thread [21:06:20 CET] <J_Darnley> I know I've seen around, at least in other projects [21:06:25 CET] <nevcairiel> J_Darnley: we dont [21:06:58 CET] <wm4> I thought every public API function is marked for exactly this reason (stack alignment) [21:07:17 CET] <wm4> attribute_align_arg [21:07:17 CET] <nevcairiel> the stack alignment attribute is only on a few of them which call into yasm code [21:07:25 CET] <nevcairiel> it doesnt account for auto-vectorization by the compiler [21:09:14 CET] <J_Darnley> Perhaps we need to disable it on Windows until it gets "fixed" [21:09:38 CET] <nevcairiel> it was an issue beofre the recent flag change already [21:09:45 CET] <nevcairiel> i personally just never build with -msse2 [21:10:01 CET] <nevcairiel> never been an issue with -mmmx -msse [21:10:25 CET] <J_Darnley> Well if it does it more often now then that just increases the need. [21:10:43 CET] <nevcairiel> it crashed before, still crashes now, nothing changed =P [21:11:06 CET] <nevcairiel> but i'm not sure everyone would be happy littering every single entry point with such an attribute [21:11:40 CET] <nevcairiel> its kinda sad the MS compiler never learned to produce 16 byte aligned stack, i'm not quite sure why not [21:11:44 CET] <J_Darnley> Propose new function attribute: AV_RETARDED_MSVC_STACK [21:12:00 CET] <nevcairiel> thats what attribute_align_arg already means :p [21:12:08 CET] <J_Darnley> The answeris probably "backwards compatability" [21:12:17 CET] <nevcairiel> well make it an option then! [21:12:35 CET] <J_Darnley> It probably is. [21:12:47 CET] <nevcairiel> i'm not aware of one [21:12:48 CET] <J_Darnley> But that only controls code it generates [21:12:55 CET] <nevcairiel> or I would just build with it and voila [21:13:25 CET] <J_Darnley> I guess we're going round in circles so I'll be quiet. [21:16:45 CET] <nevcairiel> in any case, the only really "complete" solution would be to add such an attribute to every public function that has more than 3 lines of code =p [21:19:00 CET] <wm4> or littering everything with dllexport attributes? [21:19:53 CET] <jkqxz> Flagging every public function with something is probably a good idea anyway for audit purposes (ignoring the effort of actually doing it). [21:20:21 CET] <wm4> public functions are already easily recognizable [21:20:26 CET] <wm4> they all have the av_ prefix [21:28:38 CET] <nevcairiel> anyway, hence wondering if maybe clang doesn't just blindly assume that the stack is aligned for its purposes [21:28:59 CET] <Daemon404> nevcairiel, i have no idea re: clang alignment [21:29:49 CET] <Daemon404> i think it's a pretty damn bad bug if so though [21:29:52 CET] <Daemon404> re: gcc [21:30:04 CET] <nevcairiel> its a couple year old issue on their bugzilla [21:30:08 CET] <nevcairiel> they decided not to care [21:30:11 CET] <nevcairiel> since its not a linux issue [21:30:16 CET] <nevcairiel> or OSS issue for that matter [21:30:22 CET] <nevcairiel> build everythjing with gcc, and it magically works [21:30:25 CET] <Daemon404> what about the mingw guys [21:30:37 CET] <nevcairiel> they are not very smart, remember? [21:30:38 CET] <nevcairiel> :) [21:30:45 CET] <Daemon404> as for clang you may have better luck [21:30:52 CET] <Daemon404> theyre specifically aimign for ABI compat with msvc [21:31:01 CET] <nevcairiel> indeed [21:31:07 CET] <nevcairiel> wonder what happens [21:31:15 CET] <nevcairiel> one of these days I should give clang/c2 a try [21:31:25 CET] <Daemon404> i had to hack configure to make it work [21:31:30 CET] <Daemon404> and change config.h [21:31:34 CET] <Daemon404> it misdetected some inline asm thing [21:31:45 CET] <Daemon404> inline asm builds if i flipped one bit in config.h [21:32:10 CET] Action: Daemon404 built with clang-cl and link.exe [21:32:10 CET] <nevcairiel> once its a bit more stable and "proven" we can work on getting it integrated and detected properly [21:32:17 CET] <Daemon404> right [21:32:27 CET] <Daemon404> basically i changed some of teh case statenents [21:32:36 CET] <Daemon404> the grep -q clang turned into 'if false' for me =p [21:32:36 CET] <nevcairiel> MS had a post somewhere which defines and whatnot clang sets in that mode [21:32:41 CET] <Daemon404> and cl changed to if true [21:32:44 CET] <Daemon404> maximum lazines. [22:09:20 CET] <Compn> Daemon404 : i dont think vfw/dshow exists in new windows versions. i have no idea what they are using now. i havent even seen new binary codecs besides dshow [22:09:32 CET] <JEEB> oh don't worry [22:09:35 CET] <JEEB> they still have that stuff around [22:09:42 CET] <Compn> wm4 : rewrite avidec when? [22:09:48 CET] <JEEB> actually the only thing you can rip out of a desktop windows is MF, funny enough [22:09:54 CET] <JEEB> which is the newest thing on the block [22:09:58 CET] <Compn> it was mediaframework yes [22:10:21 CET] <Compn> but i havent even seen anything use mf yet... [22:10:25 CET] <Compn> i havent really been looking [22:10:39 CET] <Compn> probably mf > vfw wrappers lol [22:10:57 CET] <JEEB> I've seen a couple of things utilizing it, but not too many [22:11:08 CET] <JEEB> firefox is probably the most popular MF thing around except for WMP [22:11:37 CET] <JEEB> and as I noted, you can actually install windows without MF, which makes it lulz if you make your app depend on it [22:16:10 CET] <wm4> apparently windows server comes without MF stuff? [22:16:46 CET] <JEEB> yeah [22:16:51 CET] <JEEB> Windows N ed too [22:16:57 CET] <Compn> why did windows shoot itself in the foot wrt video/multimedia processing ? [22:17:21 CET] <JEEB> I guess they actually made things somewhat separated and that's a side-effect [22:17:33 CET] <Compn> windows n is so difficult to reinstall a removed thing. i think i tried installing wmp on that for 3 hours and gave up [22:17:41 CET] <Compn> ah [22:17:45 CET] <JEEB> nowadays that's a single installer [22:17:45 CET] <Compn> more monopoly crap [22:18:05 CET] <wm4> what's windows n? [22:18:20 CET] <JEEB> "I don't want your media player" edition [22:18:22 CET] <Compn> i thought windows n was only in activity for 3 years after the lawsuit [22:18:28 CET] <Compn> then ms dropped it. [22:18:42 CET] <JEEB> 8.1 at least had it IIRC [22:18:44 CET] <JEEB> not sure if win10 [22:19:02 CET] <JEEB> https://support.microsoft.com/en-us/kb/3010081 [22:19:09 CET] <JEEB> exists since there's an installer for it [23:41:27 CET] <cone-576> ffmpeg 03Michael Niedermayer 07master:03c1129b2064: avfilter/vaf_spectrumsynth: assert that variables are initialized by switch() [23:41:28 CET] <cone-576> ffmpeg 03Michael Niedermayer 07master:62eb935bf932: avfilter/avf_avectorscope: assert that variables are initialized by switch() [00:00:00 CET] --- Fri Feb 12 2016

1 0

ffmpeg.log.20160211
by burek 12 Feb '16

12 Feb '16

[04:19:09 CET] <mike41> Does anyone know how to set the default or forced flags for matroksa subtitle streams? [04:31:13 CET] <llogan> mike41: maybe something like "-disposition:s:2 default" [04:31:20 CET] <llogan> 2 being the 3rd subtitle stream [04:34:38 CET] <mike41> I did [04:34:41 CET] <mike41> ffmpeg -i input.mkv -c copy -map 0 -disposition:s:1 default -disposition:s:2 default -disposition:s:3 default -disposition:s:4 default -disposition:s:5 default test.mkv [04:34:57 CET] <mike41> no change in the output file [04:39:51 CET] <llogan> i see thisquestion asked often but i don't know an answer. [04:40:17 CET] <mike41> the topic mentions this channel is logged, do we know where these logs go? [04:41:52 CET] <llogan> http://lists.ffmpeg.org/pipermail/ffmpeg-devel-irc/ [04:42:08 CET] <mike41> thanks, I'll grep this stuff and see what I can dig up [04:42:27 CET] <llogan> might be worth trawling the bug tracker too [04:42:50 CET] <llogan> IIRC, some of the similar questions involved mov instead of mkv [04:55:23 CET] <mike41> my input is: Stream #0:3(eng): Subtitle: ass [04:55:30 CET] <mike41> ffmpeg -i input.mkv -c copy -map 0 -disposition:s:3 forced -y test.mkv [04:55:35 CET] <mike41> this command line changes nothing [04:55:40 CET] <mike41> ffmpeg -i input.mkv -c copy -map 0 -disposition:s:s:1 forced -y test.mkv [04:55:47 CET] <mike41> this command line sets it as forced... [04:56:09 CET] <mike41> but :s:3 and :s:s:1 both refer to the same stream (in my mind) [05:05:58 CET] <relaxed> mike41: how many subs in the input? [05:07:03 CET] <mike41> 2 [05:07:43 CET] <mike41> 0:0 is video [05:07:50 CET] <mike41> 0:1 is ac3 [05:07:54 CET] <mike41> 0:2 is ac3 [05:08:05 CET] <mike41> 0:3 is subs (ass) [05:08:09 CET] <mike41> 0:4 is subs (ass) [05:08:54 CET] <relaxed> which one do you want to be default? [05:09:10 CET] <mike41> I want 0:3 to be forced [05:09:21 CET] <relaxed> -disposition:s:s:0 forced [05:09:31 CET] <relaxed> the count starts at zero [05:10:15 CET] <mike41> -disposition:s:s:0 forced sets 0:3 as forced [05:10:19 CET] <mike41> -disposition:s:s:1 forced sets 0:3 as forced [05:10:22 CET] <mike41> -disposition:s:s:2 forced sets 0:3 as forced [05:10:25 CET] <mike41> they all do the same thing [05:10:49 CET] <mike41> -disposition:s:3 forced does nothing [05:11:17 CET] <mike41> :/ [05:24:29 CET] <mike41> figured it out [05:24:36 CET] <mike41> to change 0:3 [05:24:42 CET] <mike41> use -disposition:3 [05:24:54 CET] <mike41> to change subtitle 0 (which is also 0:3) [05:24:59 CET] <mike41> use -disposition:s:0 [09:46:36 CET] <termos> codec_context->flags |= AV_CODEC_FLAG_GLOBAL_HEADER; is there an overview or table about which codecs+containers this needs to be set for? [09:59:46 CET] <Bluez_> termos: i havnt managed to find one, but you could look at the source to find where it prints the warning [10:00:12 CET] <Bluez_> hmm why doesnt it like a pts of 0 for video, i have to start at 1 [10:33:04 CET] <Ccdc_DuckZ> hello, can somebody help me with a segfault in my code please? http://pastebin.com/BvvDVgKA it crashes in avcodec_encode_video2(), and I'm not sure why [10:33:51 CET] <Ccdc_DuckZ> maybe I'm decoding the input data (whole png file) the wrong way? [11:24:09 CET] <k_sze[work]> You guys know MPEG-DASH? [11:24:24 CET] <k_sze[work]> I mean, does any of you know? [11:25:01 CET] <c_14> Know what specifically? [11:31:05 CET] <oldcode> is it possible to use my own malloc and free with ffmpeg as a shared object? [11:37:52 CET] <jkqxz> oldcode: Of course. Though note that malloc and free on their own are not sufficient - you will also need to replace realloc and probably posix_memalign (depends on what ffmpeg detected the system has). [11:39:07 CET] <oldcode> jkqxz, where can i find the API functions to set the memory management function pointers? i spent a few hours looking, no luck [11:41:37 CET] <jkqxz> Oh, right. There isn't any API to use like that: I was meaning ELF interposition. (With LD_PRELOAD, say.) [11:44:31 CET] <oldcode> ok, that sounds too fragile. i'll probably just leave ffmpeg to allocate memory in its own way then [11:51:11 CET] <jkqxz> There isn't anything special going on with memory allocation - ffmpeg (well, libavutil) just uses the C library. [11:54:00 CET] <oldcode> my app uses a very fast arena style allocator with leak detection - i prefer external libraries to use it instead of malloc - some libraries allow you to set your own allocator [11:57:45 CET] <jkqxz> Right, and ELF interposition by defining malloc/etc. in the executable would capture everything rather than just ffmpeg/lav*, which isn't necessarily what you want. [12:02:51 CET] <jkqxz> If you are building ffmpeg yourself then you can add a prefix to all allocation functions by adding "--malloc-prefix=foo". That would make it then use foomalloc, fooposix_memalign, etc. and you can then define those in your executable to only gets calls from ffmpeg. [12:05:02 CET] <jkqxz> *adding "--malloc-prefix=foo" to configure, that is. [12:06:48 CET] <oldcode> thats not a bad option, but its a cross platform app that only dynamically links to ffmpeg if it can find it. Users of my app can package their own ffmpeg dll or shared object if it isnt already installed on their machine [12:08:50 CET] <BtbN> good luck targetting various libav* versions that way [12:09:17 CET] <BtbN> or do you have an implementation for each ABI bump of the past few years? [12:11:05 CET] <oldcode> yes thats going to be fun - i'm going to have to create a search list of compatible ABIs / versions [12:11:47 CET] <BtbN> I'd say it's almost impossible to target all possible ffmpeg ABIs from a single binary, firefox is on the same quest currently [12:11:59 CET] <oldcode> if it finds an old version, it'll simply say no [12:12:36 CET] <oldcode> i'm happy to target only the most modern ABI, and i use only a common set of API calls which dont change a lot [12:13:02 CET] <oldcode> i'm dynamically requesting each function address [12:15:18 CET] <oldcode> and of the user want to be certain that their video will work, they can package their own ffmpeg [12:15:24 CET] <oldcode> of=if [14:04:25 CET] <momomo> anyone here alive? [14:05:12 CET] <andrey_utkin> momomo: sure [14:05:22 CET] <momomo> great :) [14:05:44 CET] <momomo> I am on a Java platform and would like to keep it as Java as possible if you know what I mean. [14:06:16 CET] <momomo> I have some streams that are in the mp2 format that I can connect to. These doesnt work in an html 5 video tag [14:06:53 CET] <momomo> I would like to connect to these on the server side, and apply a lightweight as possible convertion into appropiate format using as few resources as possible [14:07:10 CET] <momomo> and then serve this converted stream to the client [14:07:42 CET] <momomo> i have studied differnt tools and i think using ffmpeg seems to be the best option ... but I am not sure .. [14:07:51 CET] <gurabli> hi! i would like to ask for your help regarding ffmpeg with intel quick sync support on ubuntu server. Is there a guide I can follow? Thanks [14:07:55 CET] <andrey_utkin> you could start with invoking ffmpeg as system command execution [14:08:03 CET] <momomo> andrey_utkin, yes, that's what I was thinking [14:08:16 CET] <momomo> andrey_utkin, i have this for now: http://pastebin.com/xuAiDCVD [14:08:42 CET] <andrey_utkin> you could either consider conversion results as your "cache", or you could read conversion results from stdin (and to make ffmpeg put it to its stdout) [14:09:13 CET] <andrey_utkin> momomo: seems legit. What's the problem? [14:10:05 CET] <momomo> andrey_utkin, as of now ( current code ) writes to a file .. and then I am not sure how to pass it to the client .. ProcessBuilder can redirect to a file .. not sure how to get that to response output stream [14:11:11 CET] <momomo> but I just gotten started .. i think i am still far away from getting anything working .. the current example also relies on mp4 [14:18:08 CET] <momomo> andrey_utkin, is tehre a way to test this outside of java .. test a command and see the video ? is there a built in player in ffmpeg ? [14:18:15 CET] <momomo> or how would I go about this? [14:20:48 CET] <J_Darnley> There's ffplay [14:22:15 CET] <andrey_utkin> momomo: you can download the resulting file from your server and play in with any media player [14:22:44 CET] <andrey_utkin> the issue you are experiencing is not ffmpeg-related, you got to figure out how to let your users download streams [14:23:03 CET] <andrey_utkin> you could try to rely on functionality of some existing web servers, like apache/nginx/etc. [14:24:56 CET] <momomo> what would be the command to run ffplay to turn a mpeg2 into h264 and play it? [14:25:00 CET] <momomo> i tried ffplay -i http://85.229.9.206:9981/stream/channel/fdccb829c6e768d4643ae226c6e3b288 -y -vcodec libx264 [14:25:09 CET] <momomo> the url is a real one [14:25:38 CET] <momomo> andrey_utkin, yes, i need to continue exploring my options .. i am not sure what the correct path right now would be [14:25:49 CET] <DHE> sounds like you're doing a conversion for each user. is there any reason to not convert the video ahead of time and just serve one pre-converted copy instead? [14:26:12 CET] <momomo> DHE, that's my final intention .. but I'd like to get just one working first [14:26:17 CET] <momomo> :) [14:26:50 CET] <andrey_utkin> momomo: your _ffmpeg_ command is right. _ffplay_ doesn't convert, it plays ready files. [14:27:23 CET] <momomo> ah .. ok .. so I need to download some piece using ffmpef first ... lets try it .. and ffplay command then would be ? [14:28:23 CET] <momomo> does this command loook correct: [14:28:23 CET] <momomo> ffmpeg -i http://85.229.9.206:9981/stream/channel/fdccb829c6e768d4643ae226c6e3b288 -y -vcodec libx264 /momomo/Generated/Tv/a.mp4 [14:28:45 CET] <andrey_utkin> yes [14:29:12 CET] <momomo> it loads for 2 seconds and then I get : http://85.229.9.206:9981/stream/channel/fdccb829c6e768d4643ae226c6e3b288: End of file [14:29:27 CET] <momomo> I can't see the file a.mp4 under Generated [14:34:15 CET] <momomo> seems to be something wrong with the strams [14:34:25 CET] <momomo> i apoligzie .. i will try again soon [14:56:31 CET] <endle> Excuse me, I'm reading ffmpeg's code now, and faced a problem [14:57:17 CET] <endle> The coverage page(http://coverage.ffmpeg.org/) shows that libswscale got 54.7% coverage, but I can't find the test code for it [14:57:28 CET] <endle> tests$ git grep -i swscale [14:57:28 CET] <endle> ref/fate/source:libswscale/log2_tab.c [14:57:42 CET] <endle> So where is the test code for libswscale? [15:00:19 CET] <J_Darnley> in fate [15:00:58 CET] <J_Darnley> any time a change in pixel format or a change in size is needed swscale is ued. [15:01:01 CET] <J_Darnley> *used [15:01:46 CET] <endle> So the code in libswscale are tested indirectly? [15:02:40 CET] <J_Darnley> Not sure. You'd have to check the tests run [15:09:52 CET] <endle> A little confused [15:10:36 CET] <J_Darnley> Okay, I think there are some dedicated test programs but most of the testing is done by executing ffmpeg [15:11:10 CET] <J_Darnley> there are flac and h264 api tests for example [15:12:15 CET] <J_Darnley> oh checkasm too for some dsp functions [15:13:16 CET] <endle> Thanks. I was trying to send a patch fixing compile warnings in libswscale as my first patch to ffmpeg [15:13:44 CET] <J_Darnley> Oh a noble effort [15:13:58 CET] <J_Darnley> You can discuss this on #ffmpeg-devel [15:14:09 CET] <J_Darnley> You're more likely to find someone who knows then. [15:14:45 CET] <endle> Oops, I thought that this is for devel, and #ffmpeg-user is for daily use [15:15:17 CET] <J_Darnley> I don't think there is an #ffmpeg-user IRC channel [15:15:23 CET] <endle> Thanks [15:23:57 CET] <momomo> is "-s 360x480" a requiremetn ? [15:24:18 CET] <DHE> if you want to rescale the input, it's the easiest way [15:28:26 CET] <momomo> ook [15:30:46 CET] <momomo> If I don't want to record a stream, just send it along to the clients ... is there a way to make sure the file never gets bigger than say 50 MB ? [15:34:45 CET] <DHE> set the bitrate to a value that ensures it won't. you'll have to do some math on-the-fly if the video length varies [15:41:39 CET] <momomo> -maxrate 3000k -bufsize 6000k ? [15:42:31 CET] <DHE> the math says that a 50 MB file will be about 2.2 minutes long [15:42:39 CET] <DHE> (audio not included) [15:43:21 CET] <momomo> so what is the flags and values that I need to use? [15:47:01 CET] <YamakasY> what do you guys recommend for creating a mp4 out of an flv with best quality ? [15:47:28 CET] <J_Darnley> ffmpeg [15:48:35 CET] <YamakasY> yes but which parameters :) [15:48:50 CET] <J_Darnley> Depends on what's in the flv file [15:49:12 CET] <YamakasY> audio/video h.264 [15:49:17 CET] <J_Darnley> If they're suitable for mp4 then copy [15:49:21 CET] <J_Darnley> otherwise encoding [15:49:30 CET] <J_Darnley> h264 is not an audio format [15:49:31 CET] <YamakasY> yeah that's why [15:49:42 CET] <YamakasY> fps if 25 [15:49:45 CET] <YamakasY> of [15:50:08 CET] <J_Darnley> try just -c copy to begin with [15:50:54 CET] <YamakasY> yeah that was nog that good @ the past [15:51:00 CET] <YamakasY> *not [15:51:46 CET] <J_Darnley> "not good" is not an error message [15:51:57 CET] <J_Darnley> I will assume the the audio is the problem then [15:52:01 CET] <J_Darnley> -c:a aac [15:52:36 CET] <YamakasY> yap [15:52:48 CET] <YamakasY> -c is a shorttag for cop and a ? [15:52:51 CET] <YamakasY> async ? [15:53:15 CET] <J_Darnley> No [15:53:34 CET] <J_Darnley> The old syntax would be -vcodec and -acodec [15:53:47 CET] <J_Darnley> but apprarently that's too verbose today [15:56:14 CET] <at0m> hi, is there any (static?) builds of ffmpeg that support QSV? https://trac.ffmpeg.org/wiki/HWAccelIntro [15:56:24 CET] <at0m> looking to encode using hw accel [16:06:01 CET] <YamakasY> J_Darnley: is it or think people it is ;) [16:06:29 CET] <J_Darnley> Ask whoever changed it. [16:06:44 CET] <J_Darnley> I think the command line options just get worse and worse [16:08:29 CET] <Ccdc_DuckZ> I don't understand what cleanup I have to do here http://ffmpeg.org/doxygen/2.8/group__lavf__core.html#gadcb0fd3e507d9b58fe78… [16:09:23 CET] <Ccdc_DuckZ> it says I should call avcodec_close and avformat_free_context, but what should I pass to those? the example from muxer.c linked there passes AVStream::st to avcodec_close [16:09:46 CET] <Ccdc_DuckZ> so I guess that's one, but what about avformat_free_context? [16:10:13 CET] <J_Darnley> Does it not have a type in the function signature?> [16:10:55 CET] <Ccdc_DuckZ> it takes an AVFormatContext, so not an AVStream [16:11:55 CET] <Ccdc_DuckZ> and avcodec_close takes an AVCodec... so bonus question: how do I cleaup the AVStream pointer, since none of the suggested functions there accept one such pointer? [16:12:57 CET] <J_Darnley> I don't know. Just exit? [16:13:15 CET] <Ccdc_DuckZ> sounds like a leak tho [16:14:13 CET] <J_Darnley> Does any real OS not free memory when a program exists? [16:14:19 CET] <J_Darnley> *exits [16:14:59 CET] <Ccdc_DuckZ> idk, the problem is what happens if the program calls my function 1 million times without ever exiting? [16:15:38 CET] <Ccdc_DuckZ> rage quit? [16:16:09 CET] <Ccdc_DuckZ> (o° ° o5 ;;) [16:17:05 CET] <Mavrik> Ccdc_DuckZ, AVStreams are cleaned up in avformat_free_context. [16:18:00 CET] <Mavrik> https://www.ffmpeg.org/doxygen/2.5/libavformat_2utils_8c_source.html#l03574 [16:20:55 CET] <Ccdc_DuckZ> Mavrik: so whatever AVStream, it gets deleted when the associated context is killed? [16:21:54 CET] <Mavrik> Yes, since it gets attached to the format context when you create it. [16:22:14 CET] <Mavrik> Streams not attached to a format context don't really make much sense in ffmpeg API [16:22:50 CET] <Ccdc_DuckZ> ok, I see :) [16:43:34 CET] <YamakasY> can't I get a/v sync much much better ? I use -async 1 now [16:46:12 CET] <shincodex> Hi, How do i tell if libavformat or whatever in the stupid ass collection network stream dropped [16:46:15 CET] <shincodex> IS there a callback [16:46:42 CET] <shincodex> EG RTSP into a camera get stream av_decode everything is great and lovely... then some asshat kicks the router... or they reboot the camera [16:47:11 CET] <shincodex> IS there a callback of rtsp/avformat the library controlling connections of ffmpeg to tell me E_IDIOT_RECONNECT [16:47:19 CET] <shincodex> or errno [16:48:22 CET] <shincodex> OR... https://trac.ffmpeg.org/ticket/2694 [16:48:28 CET] <shincodex> Look at that would that work? [16:48:46 CET] <shincodex> Does that timeout value checked in the strucutre mean no communication from stream... (we might be dead or not) [17:05:15 CET] <Bluez_> hi guys [17:05:38 CET] <Bluez_> im trying to put raw aac frames into a quicktime container, the video works fine, but the audio only seems to play half of each frame [17:05:54 CET] <Bluez_> pts/dts is all correct, the audio is in the right place in the video [17:06:28 CET] <Bluez_> frame size is correct (confirmed on a xcode version) [17:07:13 CET] <shincodex> dont sleep anywhere. [17:31:24 CET] <Bluez_> argh this is driving me crazy [17:33:31 CET] <Ccdc_DuckZ> other question: I'm following the code this dude posted on SO http://stackoverflow.com/questions/31387067/segmentation-fault-with-avcodec… and I see he does AVFrame* sourceAvFrame = av_frame_alloc() [17:34:07 CET] <Ccdc_DuckZ> then the next time he uses that variable is: avpicture_alloc((AVPicture*)sourceAvFrame, blah) [17:34:56 CET] <Ccdc_DuckZ> so why the cast, and what do I pass to avpicture_free()? because he's not doing any cleanup at all [17:36:22 CET] <Ccdc_DuckZ> I'm cleaning my frames with av_freep(&frame->data[0]); av_frame_free(&frame); [17:39:46 CET] <momomo> guys, i can't get this to work, i know it's java.. but maybe you can spot something that I can't : http://pastebin.com/Zf2YEZ7a [18:01:31 CET] <Mavrik> Ccdc_DuckZ, that looks like ancient API that's mostly deprecated. [18:01:36 CET] <Mavrik> What are you trying to do? [18:05:16 CET] <Ccdc_DuckZ> Mavrik: I'm trying to encode jpg/png frames into an animated gif, but my code segfaults miserably in avcodec_encode_video2 :( [18:06:31 CET] <Mavrik> Anyway, avpicture_alloc will kill you on new reference counted code. [18:06:54 CET] <Mavrik> use av_frame_alloc(), and av_frame_get_buffer() to allocate it's buffer. [18:07:01 CET] <Mavrik> Make sure you set properties av_frame_get_buffer needs first. [18:07:16 CET] <Mavrik> Then you'll have valid AVFrame.data buffers. [18:07:22 CET] <Mavrik> And you'll be able to encode stuff. [18:07:32 CET] <Mavrik> Also, get a debugger so you'll know where avcodec_encode_video2 crashes. [18:09:15 CET] <Ccdc_DuckZ> I know, I built ffmpeg in release, so I won't see where it crashes anyways [18:11:21 CET] <Ccdc_DuckZ> btw I'm a bit confused now... first I call avcodec_decode_video2, which fills a frame structure, right? in that case packet=input and frame=output, right? [18:12:11 CET] <jmcdonald> Hey guys, I'm trying to install ffmpeg on CentOS Linux release 7.2.1511, but all guides that I'm following seem to be out of date and no longer valid. Does anyone have any tips, or should I just install from source? [18:12:12 CET] <Ccdc_DuckZ> then I call avcodec_encode_video2, passing the same frame I passed to decode and a different packet, because here packet=output and frame=input... or not? [18:12:49 CET] <Ccdc_DuckZ> jmcdonald: normally I ask those questions on the distro's channel [18:13:12 CET] <jmcdonald> Sorry, new to IRC. I'll switch to the distro channel, thanks :) [18:14:37 CET] <Ccdc_DuckZ> Mavrik: so that's to say, if my frame is (supposedly) already initialized, when do I have to call av_frame_get_buffer()? [18:15:09 CET] <Mavrik> Ccdc_DuckZ, ah, nop, if you decoded it then your AVFrame must be valid. [18:15:23 CET] <Mavrik> And you should really build ffmpeg with debug symbols for easier debugging. [18:15:31 CET] <Mavrik> Usually just seeing which field crashes is a huge help :P [18:16:22 CET] <Ccdc_DuckZ> can you have a quick look at my code, in case you spot any obvious mistake? 'cause I'm leaving soon, reubilding will have to wait until tomorrow [18:20:42 CET] <Ccdc_DuckZ> Mavrik: http://pastebin.com/JnzdwBGU and http://pastebin.com/jzesvhpW and http://pastebin.com/mj14TAvD [18:21:53 CET] <Mavrik> Ccdc_DuckZ, hrmf, I don't see you converting pixel formats anywhere. [18:22:09 CET] <Mavrik> You set output pix_fmt to PAL8 and input is almost certanly either YUV (jpeg) or RGB (png) [18:22:26 CET] <Mavrik> So when the encoder tries to read the palette it probably crashes. [18:22:59 CET] <Ccdc_DuckZ> hmm... I thought the encoder would know that... [18:27:12 CET] <Ccdc_DuckZ> I can't find how to do that... I'll check tomorrow again, thanks for the hint btw ;) [18:28:20 CET] <King_DuckZ> Mavrik: I'll be online/afk from here btw [18:54:06 CET] <shincodex> what about me dieing [18:54:15 CET] <shincodex> what can i check in ffmpeg to see if my network stream died [19:00:10 CET] <anewinckname> I am trying to make a script that when ran will execute a ffmpeg command.I have the script mostly made,minus one issue.I need to output the encoded video in to a new folder(which is made at the start of the script).Im having trouble finding out how to output to a folder. Here is the script: http://pastebin.com/DDeCjmwH [19:28:11 CET] <pzich> anewinckname: that looks like it should work, what are you seeing? [19:29:03 CET] <anewinckname> pzich: It does the encoding,but it errors out saying that "done" is a directory and exits. The encoded video erased [19:31:37 CET] <DHE> it's reading a .mkv as both input file and subtitle file? [19:32:23 CET] <DHE> otherwise yeah looks okay to me [19:32:46 CET] <anewinckname> Yes. I need to hardsub the subtitles,the subtitles are softsubbed and the device i am using dosent support subtitles Running the command without variables in it, it encodes as expected [19:34:07 CET] <derekprestegard> hello, [19:35:25 CET] <derekprestegard> Im scaling a 1920x1080 video down to fit in an 720x512 frame. I can go to 720x480 no problem using -vf pad=720:480:0:60 but I also need to add 32 lines of pad to the top for the vbi area. if I chain this with another pad=720:512:0:0 I get [Parsed_pad_2 @ 0x7fdef1c13020] Input area 0:0:720:480 not within the padded area 0:0:512:480 or zero-sized [19:35:26 CET] <derekprestegard> [Parsed_pad_2 @ 0x7fdef1c13020] Failed to configure input pad on Parsed_pad_2 [19:35:26 CET] <derekprestegard> Error opening filters! [19:35:53 CET] <derekprestegard> sorry that was not clear& one second [19:49:52 CET] <situation> is this for disney? :D [19:52:32 CET] <Matador> solid [19:52:44 CET] <derekprestegard> sorry about that :D [19:53:50 CET] <Matador> its ok.. your an expert [20:12:03 CET] <_TheArtist_> Hello! Can ffserver accept rtmp input? [20:18:28 CET] <derekprestegard> thoughts on dithering options in zscale? is there a generally regarded best practice for 10 bit to 8 bit dithering? (prores to MPEG-2, DV, H264 etc) [20:18:51 CET] <derekprestegard> guessing error_diffusion gives nice results? [20:22:48 CET] <durandal_1707> use format after zscale to make sure its called and not scale [20:26:58 CET] <derekprestegard> durandal_1707: not sure what you mean - can you clarify? proposed command follows: [20:27:01 CET] <derekprestegard> ffmpeg -i Trimmed_Trailer_Source.mov -map 0:0:0:0:0:0 -vf zscale=w=720:h=360:filter=spline36:dither=error_diffusion,pad=720:480:0:60,pad=720:512:0:32,telecine=t:23,colormatrix=src=bt709:dst=bt601 -c:v mpeg2video -pix_fmt yuv422p -b:v 15000k -minrate 15000k -maxrate 15000k -bufsize 7500k -dc 10 -alternate_scan 1 -mpeg_quant 1 -flags +ilme+ildct -top 1 -aspect 4:3 -filter_complex "[0:1][0:2]amerge=inputs=2[aout]" -map "[aout]" -c:a mp2 [20:27:01 CET] <derekprestegard> -b:a 384000 -f vob ntsc_mpeg2_dithered.mpg [20:27:53 CET] <durandal_1707> why telecine? [20:28:04 CET] <derekprestegard> have to make a 59.94i otuput for broadcast [20:28:16 CET] <derekprestegard> also why Im making a 16x9 file letterboxed to 4x3 :| [20:29:53 CET] <durandal_1707> also zscale should change colormatrix not that awful filter [20:30:05 CET] <derekprestegard> heheh ok [20:30:22 CET] <durandal_1707> it operates in 8bit only [20:31:55 CET] <derekprestegard> so like zscale=w=720:h=360:filter=spline36:dither=error_diffusion:matrix=170m:matrixin=709 [20:32:03 CET] <furkan> jkqxz: so guess what, wireshark didn't reveal any packet loss, because everything is being received and in correct order, but it turns out that the kernel has been dropping packets (specifically UDP packets) [20:32:19 CET] <furkan> i think i need to increase some buffering somewhere, it's probably because of lack of flow control with UDP [20:32:40 CET] <derekprestegard> durandal_1707: noticed theres no 601 mentioned in zscale for matrix [20:33:11 CET] <durandal_1707> its another name, alias for sure [20:33:24 CET] <derekprestegard> 170m == 601 basically? [20:37:42 CET] <durandal21707> 601 is mentioned for other things primaries, etc [20:41:40 CET] <derekprestegard> got it [20:41:44 CET] <derekprestegard> fark this is slow!!! [20:42:05 CET] <derekprestegard> like under 2fps when I use zscale to dither and do all the colorspace conversions [20:42:22 CET] <derekprestegard> is that normal?? [20:42:26 CET] <durandal21707> perhaps you havent enabled asm when building it? [20:42:34 CET] <derekprestegard> I downloaded a zeranoe build [20:42:39 CET] <durandal21707> ah [20:43:29 CET] <derekprestegard> I wonder if its applying all the conversions when the video is still 1080 then scaling it down. That would be wasteful [20:46:04 CET] <durandal21707> even if it is, difference should be marginal [20:46:28 CET] <durandal21707> i got 12-14fps with 1080i30 input [20:46:38 CET] <durandal21707> with asm enabled [20:46:52 CET] <durandal21707> and this is without threads unfortunately [20:50:50 CET] <derekprestegard> thats weird [20:50:55 CET] <derekprestegard> how does one thread it? [20:51:47 CET] <durandal21707> currently you cant [20:51:55 CET] <derekprestegard> gotcha [20:54:14 CET] <derekprestegard> how can i tell if asm is disabled? [20:59:27 CET] <derekprestegard> yeah dithering only gets me about 10fps [20:59:34 CET] <derekprestegard> dithering + color matrix goes under 5 [20:59:57 CET] <derekprestegard> and dithering + color matrix + transfer + primaries goes under 2& this is all on an Ivy Bridge Xeon too! [21:01:04 CET] <JEEB> derekprestegard: I think durandal21707 picked the slowest defaults for dithering [21:01:57 CET] <durandal21707> JEEB: default is none :) [21:02:03 CET] <JEEB> oh [21:02:04 CET] <JEEB> lol [21:03:02 CET] <derekprestegard> so can I get better than no dithering and better than 10 fps for SD ?? :) [21:11:30 CET] <JEEB> you might want to also try out vapoursynth, where zimg/zscale "comes from" and is threaded [21:11:40 CET] <JEEB> it doesn't really handle audio though [21:31:48 CET] <Bluez_> hey guys, this is driving me crazy. Im encoded h264 and aac (raw frames) into a .mov (quicktime) container. the video works, the audio also works in vlc [21:32:14 CET] <Bluez_> but in quicktime i dont get any audio, and media ispector in quicktime only shows the avc1/h264 track [21:32:15 CET] <Bluez_> http://pastebin.com/raw/yeACQUHD [21:32:32 CET] <Bluez_> im pulling out my hair :/ [21:43:18 CET] <llogan> Bluez_: did you use ffmpeg to encode? [21:43:36 CET] <Bluez_> no its raw h264 and aac from a hardware device [21:43:49 CET] <Bluez_> theres no adts headers either only rtp, which ive removed [21:44:01 CET] <llogan> how is ffmpeg involved? muxing? [21:44:02 CET] <at0m> hi, is there any (even static?) builds of ffmpeg that support QSV? https://trac.ffmpeg.org/wiki/HWAccelIntro ? looking to use hw encoding. will try a compile myself, but it's not my forte... [21:44:06 CET] <Bluez_> yes just muxing [21:44:39 CET] <Bluez_> im using the apis in C [21:45:01 CET] <Bluez_> http://pastebin.com/raw/RAkWCTjM [21:45:28 CET] <llogan> not my area, but maybe you can get help at libav-user mailing list if not here [21:50:59 CET] <derekprestegard> I heard ffmbc has been merged into ffmpeg? [21:51:09 CET] <derekprestegard> I want to make imx50 but -target imx50 doesnt appear to work [21:51:15 CET] <llogan> where did you hear that? [21:51:24 CET] <derekprestegard> dunno somewhere on the web lol [21:51:31 CET] <derekprestegard> in any case, can ffmpeg make imx50 in mxf? [21:52:25 CET] <JEEB> probably yes, -target imx50 is most probably just a shorthand for a bunch of options [21:52:31 CET] <derekprestegard> yeah [21:52:37 CET] <derekprestegard> seems to throw an error for me [21:52:49 CET] <JEEB> also you can't really merge things as-is from ffmbc [21:52:57 CET] <derekprestegard> Unknown target: imx50 [21:53:17 CET] <JEEB> yes, which is why I kind of wanted to hint at you actually going to look at the actual parameters that target sets :P [21:53:30 CET] <derekprestegard> yeah lol I did, I think Ive got a valid video essence [21:53:36 CET] <derekprestegard> bunch of params [21:53:53 CET] <derekprestegard> -pix_fmt yuv422p -minrate 50000k -maxrate 50000k -b:v 50000k -intra -flags +ildct+ilme+low_delay -intra_vlc 1 -non_linear_quant 1 -ps 1 -qmin 1 -qmax 3 -top 1 -dc 10 -bufsize 2000000 -rc_init_occupancy 2000000 -rc_buf_aggressivity 0.25 -bsf:v imxdump -tag:v mx5p [21:54:32 CET] Action: llogan is happy to not ever deal with mxf [21:55:51 CET] <derekprestegard> my concern is the audio - imx would properly have a 4ch aes3/s302m pcm track [21:57:40 CET] <JEEB> I've seen everything from multiple channels in a single track to multiple tracks for a set of channels [21:57:43 CET] <JEEB> in MXF [21:57:53 CET] <JEEB> thankfully I only have to deal with the waste other people throw at me [21:57:57 CET] <JEEB> not create more of it [21:58:08 CET] <dowdle> Any familiar with WebEx's .wfr files? Is that a completely closed codec that one needs their software to work with? Or is it something perhaps ffmpeg understands (I should check the --formats flag) [21:58:50 CET] <dowdle> Make that -formats [21:58:54 CET] <derekprestegard> yeah, xdcam generally requires one channel per track and imx usually requires 4 or 8 channels per track [21:58:56 CET] <JEEB> easiest is to build recent FFmpeg and see if "ffprobe file.wfr" tells you anything useful [21:59:10 CET] <derekprestegard> its pretty straighforward once youve had some experience with these files :) [21:59:20 CET] <JEEB> derekprestegard: I also love it when people have out-of-band data in the video [21:59:28 CET] <llogan> dowdle: a sample file would be nice to have if it can't [21:59:30 CET] <derekprestegard> yeah lol!!! that part is totally looney tunes [21:59:58 CET] <dowdle> llogan: I need to find a sample. I was just asked about them today and am researching. [22:00:38 CET] <dowdle> llogan: They have their own player here - https://www.webex.com/play-webex-recording.html and that page mentions .arf and .wrf files. [22:01:12 CET] <llogan> ARF. The file for dogs. [22:05:54 CET] <derekprestegard> how do I make silent audio tracks? [22:06:39 CET] <llogan> use anullsrc filter [22:06:57 CET] <llogan> or aevalsrc if you prefer [22:11:00 CET] <derekprestegard> turns out theres a special mxf muxer for imx/d10. its -f imx_d10 [22:15:43 CET] <derekprestegard> if anyone needs to encode 16x9 PAL IMX 50, heres a working command that is valid in sony tools [22:16:06 CET] <derekprestegard> -c:v mpeg2video -pix_fmt yuv422p -minrate 50000k -maxrate 50000k -b:v 50000k -intra -flags +ildct+ilme+low_delay -intra_vlc 1 -non_linear_quant 1 -ps 1 -qmin 1 -qmax 3 -top 1 -dc 10 -bufsize 2000000 -rc_init_occupancy 2000000 -rc_buf_aggressivity 0.25 -bsf:v imxdump -tag:v mx5p -filter_complex "[0:1][0:2][0:3][0:4]amerge=inputs=4[aout]" -map "[aout]" -c:a pcm_s24le -r 25 -f mxf_d10 pal_imx50_16x9.mxf [22:18:28 CET] <dowdle> Ok, I got a sample .wrf file and no... ffprobe doesn't know what it is... and neither does the file command. "data" [22:23:59 CET] <llogan> dowdle: you can make a feature request on the bug tracker and provide a sample [22:33:15 CET] <dowdle> llogan: Ok. [22:57:23 CET] <dowdle> llogan: Without any kind of spec for .wrf, filing a request for feature ticket would be a waste of time... unless someone likes reverse engineering in their free time. [22:59:25 CET] <JEEB> there's some people who like that [22:59:43 CET] <JEEB> there's almost always someone with Mrs. IDA up on their laptop at the usual meetings [23:10:22 CET] <momomo> when I execute the command : ffmpeg -i http://85.229.9.206:9981/stream/channel/channel4 -y -vcodec libx264 /momomo/Generated/Tv/store.mp4 [23:10:31 CET] <momomo> the file store.mp4 gets created [23:11:00 CET] <momomo> if I try to play it before exiting (CTRL+C) then it won't play ... first when I exit does it do so [23:11:27 CET] <momomo> the thing is that I would like to send the stream to a webbrowser directly and keep the stream open ( not exit it ) [23:12:12 CET] <drv> the mp4 format is not streamable; it has a moov chunk that needs to be written after the file is finished before it can be played [23:12:42 CET] <momomo> what format is? hls262 == mp4 ? [23:12:49 CET] <momomo> 264 [23:13:18 CET] <drv> hls uses a bunch of short segments to accomplish streaming [23:14:18 CET] <momomo> the above command converts a stream to hls264 i believe ... [23:14:47 CET] <furq> it doesn't and you seem to have got hls and h264 confused [23:14:59 CET] <momomo> yes, maybe [23:15:05 CET] <momomo> i just found this: https://hetzel.net/2014-01-29/youtube-videos-will-process-faster-encode-str… [23:15:07 CET] <furq> hls is a bunch of .ts files [23:15:22 CET] <furq> also youtube doesn't use hls [23:15:22 CET] <momomo> yes, lets forget about that. i need web usable format [23:15:39 CET] <momomo> h.264 [23:15:53 CET] <furq> as far as i know you're limited to flv [23:15:59 CET] <furq> unless you use hls or dash, in which case ts [23:16:15 CET] <furq> and of course flv won't work with html5 video [23:16:32 CET] <momomo> what does work with html5 video ? [23:16:39 CET] <furq> mp4 and webm [23:16:47 CET] <furq> webm can't contain h264 [23:17:00 CET] <momomo> webm can be used for streaming ? [23:17:08 CET] <furq> maybe [23:17:10 CET] <furq> i've never tried [23:17:17 CET] <momomo> mp4/h.264 can not ? [23:17:51 CET] <pinPoint> Does anyone think the first comment at the bottom is spot on? What about the second comment? Does time(cpu time) not == overall time spend encoding? http://wideopenbokeh.com/AthenasFall/?p=303 [23:17:55 CET] <pinPoint> thanks ^^^^ [23:18:06 CET] <furq> momomo: your best bet is probably using HLS [23:18:09 CET] <furq> ffmpeg has a builtin hls muxer [23:18:24 CET] <furq> and iirc it will work with a single .ts file, it doesn't have to be chunked [23:18:28 CET] <furq> the playlist can just refer to offsets [23:18:57 CET] <furq> it won't work with plain html5 video, but you can use hls.js [23:19:40 CET] <furq> https://github.com/dailymotion/hls.js/ [23:22:01 CET] <momomo> furq, so I would use ffmpeg to convert original stream to hls somehow, and then use this gigantic library to play it [23:22:02 CET] <momomo> ? [23:22:49 CET] <furq> normally i would agree that 114KB of javascript is excessive, but this is for streaming video [23:23:08 CET] <furq> it's not like that's going to be a significant percentage of bandwidth [23:23:39 CET] <furq> https://www.ffmpeg.org/ffmpeg-formats.html#hls-1 [23:24:53 CET] <momomo> furq, true [23:25:01 CET] <momomo> i found this pretty big thread: http://stackoverflow.com/questions/21921790/best-approach-to-real-time-http… [23:25:51 CET] <momomo> it doesnt seem to work on iphone thoguh [23:26:24 CET] <furq> er [23:26:37 CET] <furq> hls was designed by apple [23:26:51 CET] <furq> it should definitely work on iOS [23:27:05 CET] <furq> i don't think DASH does, but DASH looks like a hot mess [23:27:22 CET] <momomo> hls.js is compatible with browsers supporting MSE with 'video/MP4' inputs. as of today, it is supported on: [23:27:22 CET] <momomo> Chrome for Android 34+ [23:27:22 CET] <momomo> Chrome for Desktop 34+ [23:27:22 CET] <momomo> Firefox for Desktop 38+ (with media.mediasource.whitelist=false in about:config) [23:27:22 CET] <momomo> IE11+ for Windows 8.1 [23:27:23 CET] <momomo> Safari for Mac 8+ (beta) [23:27:31 CET] <furq> right, iOS supports it natively [23:27:39 CET] <momomo> not in the browser [23:28:30 CET] <momomo> i have been in the web world long enough to understand that these mutha fuckers do these things on purpose ... there is nothing in the web world that just worjs [23:28:31 CET] <momomo> works [23:29:16 CET] <furq> no i'm pretty sure it's supported in the browser on mobile devices [23:29:32 CET] <furq> i have no way of testing though [23:30:20 CET] <furq> https://walterebert.com/playground/video/hls/ [23:30:35 CET] <furq> that claims it's supported in safari on osx and iOS [23:31:22 CET] <furq> obviously you'd still ideally want an RTMP fallback for people on old IEs [23:32:43 CET] <momomo> not really, don't care [23:32:47 CET] <momomo> about those [23:33:18 CET] <momomo> on that demo page, only the last one worked .. the flash one [23:33:27 CET] <furq> yeah that page isn't using hls.js [23:33:32 CET] <furq> that's using hls natively [23:33:45 CET] <furq> which should work on iOS safari and android browser [23:35:04 CET] <momomo> jeezus .. why would the not support it natively on the desktop .. always making life difficult [23:35:16 CET] <furq> dash is the anointed son of the desktop [23:35:23 CET] <furq> which of course doesn't work at all on iOS [23:35:35 CET] <furq> hls.js is actually pretty good though [23:37:23 CET] <momomo> i found this too: https://www.theoplayer.com/ costs money though [23:39:13 CET] <furq> that's probably just using hls.js anyway [23:40:07 CET] <TD-Linux> furq, well MSE is the anointed son of the desktop :) [23:41:22 CET] <momomo> I need to test this tomorrow ... and figure out the right command for executing ffmpeg into a hls [23:45:40 CET] <tp__> momomo: mkv works in chrome also [23:46:00 CET] <tp__> but thats probably the only browser mkv works in :P [23:46:22 CET] <kepstin> nah, works in gnome's "epiphany" browser. but nobody uses that :) [23:46:25 CET] <furq> does it actually officially work or is it just a byproduct of webm support which might break [23:46:49 CET] <kepstin> (epiphany will play anything that your gstreamer installation can decode in a video tag) [23:47:10 CET] <tp__> not officially, but it doesnt break [23:47:20 CET] <furq> by might i mean in the next chrome release [23:48:58 CET] <tp__> no idea, i cant see why they should break it [23:49:37 CET] <furq> famous last words [23:50:39 CET] <tp__> worked for quite a while now [00:00:00 CET] --- Fri Feb 12 2016

1 0

ffmpeg-devel.log.20160210
by burek 11 Feb '16

11 Feb '16

[01:00:48 CET] <cone-086> ffmpeg 03Michael Niedermayer 07master:674cc26f257c: avfilter/vf_nnedi: Fix memleak [01:52:18 CET] <michaelni> who maintains the dca/dts decoder code ? [01:52:33 CET] <michaelni> there are a few issues coverity found in it ... [01:55:41 CET] <jamrial> foo86, supposedly [01:56:07 CET] <jamrial> he said he'd maintain it but afaik didn't send a patch to add himself to MAINTAINERS [03:05:10 CET] <Timothy_Gu> ugh why isn't there a pmulhud [03:11:44 CET] <michaelni> jamrial, i mailed him about a week ago about coverity but got no reply [07:26:58 CET] <jya> michaelni: you wrote in an earlier ticket, that the best way to test if we were running LibAv or FFmpeg was to test libavcodec microversion and if it was 100 [07:27:14 CET] <jya> but in commit ae5b2c525, "foo86" made the microversion 101 [07:27:17 CET] <jya> that rather sucks [07:27:49 CET] <jya> any reason to upgrade microversion rather than minor ? [07:50:51 CET] <Timothy_Gu> jya: you can test MINOR >= 100 [07:51:09 CET] <Timothy_Gu> *MICRO [07:52:51 CET] <jya> Timothy_Gu: sure... I opened ticket 5226. Problem for us is that we're riding the train and if a 2.9 version comes out with that, it will break our FFmpeg support for 12 weeks. [07:54:20 CET] <jya> so knowing that a new version will likely come for ubuntu 16.04, our release cycle may be causing issue. so much easier to have this changed in ffmpeg :) and also document that it won't happen again (can see someone resetting to 0 the way it currently is documented) [07:54:40 CET] <Timothy_Gu> jya: you mean you can't push a trivial patch like this to your release branch? [07:54:50 CET] <jya> that's right [07:55:11 CET] <jya> please don't close it as invalid without looking at the background [07:55:13 CET] <Timothy_Gu> :( [07:55:26 CET] <Timothy_Gu> jya: i closed it before you replied. fflogger is slow :) [07:57:32 CET] <Timothy_Gu> regardless, I don't think it is reasonable to make sure ffmpeg doesn't bump micro for 12 weeks [07:58:02 CET] <jya> Timothy_Gu: we should make sure it never goes under 100 though [07:58:26 CET] <jya> so that itself that worth a ticket so version.h is documented as such [08:01:27 CET] <jya> Timothy_Gu: I did misread the comments though :( [08:01:31 CET] <jya> my fault [08:01:55 CET] <Timothy_Gu> jya: i guess we can bump minor in the future, but there is no guarantee that a user won't use a version of ffmpeg that has micro that is larger than 100 [08:02:08 CET] <jya> indeed. [08:02:27 CET] <jya> so >= 100 I can deal with.. I just want to make sure it *never* reset to 0. [08:02:40 CET] <jya> Like is this something set in stone and known by every ffmpeg committer [08:02:43 CET] <Timothy_Gu> jya: no it will never be done [08:03:15 CET] <Timothy_Gu> it has been >=100 since 2011 and will be for the foreseeable future [08:03:26 CET] <jya> well, it has been 100 since 2011 :) [08:03:46 CET] <jya> so assuming it was always going to be 100 was a fairly reasonable assumption [08:04:07 CET] <Timothy_Gu> yes [08:04:10 CET] <Timothy_Gu> also check https://www.ffmpeg.org/download.html#releases [08:04:21 CET] <jya> IMHO, the changes were fairly significant, a minor bump would have been more usual [08:05:12 CET] <Timothy_Gu> technically, yes, for this change we should have probably bumped minor [08:06:18 CET] <Timothy_Gu> also note that libavresample's micro will never be above 100, since it is kept only for compatibility with Libav, and we don't maintain it per se [08:07:15 CET] <wbs> jya: it hasn't "always been 100 since 2011", this wasn't the first time it has been >100 - that happens all the time. see e.g. e9e623369 and 73e4565df [08:07:40 CET] <Timothy_Gu> and also the release page i linked [08:07:43 CET] <jya> Timothy_Gu: we only use libavcodec and libavformat [08:08:15 CET] <Timothy_Gu> ffmpeg-2.8.6: [08:08:15 CET] <Timothy_Gu> libavutil 54. 31.100 [08:08:15 CET] <Timothy_Gu> libavcodec 56. 60.100 [08:08:16 CET] <Timothy_Gu> libavformat 56. 40.101 [08:08:25 CET] <jya> we had some crashes happening because someone had installed two different version of libavformat, and libavcodec.56.so was loading the wrong libavformat [08:08:37 CET] <jya> made for some really weird traces ! [08:08:54 CET] <jya> sorry libavcodec and libavutil [08:09:11 CET] <jya> we use our own demuxer [08:14:38 CET] <Timothy_Gu> is it possible to emulate roundps with truncation in sse2? [09:59:22 CET] <cone-252> ffmpeg 03Paul B Mahol 07master:4ca8879d1989: avfilter: add metadata filters [10:06:16 CET] <ubitux> Stream #0:0(eng): Video: vp9 (Profile 0), yuv420p(tv, bt709/unknown/unknown), 3840x2160, SAR 1:1 DAR 16:9, 23.98 fps, 23.98 tbr, 1k tbn, 1k tbc (default) [10:06:17 CET] <ubitux> Stream #0:1(eng): Audio: opus, 48000 Hz, stereo, fltp (default) [10:06:25 CET] <ubitux> heh, youtube is getting better at delivering decent videos [10:07:08 CET] <ubitux> Timothy_Gu: no other comment from me, i like the idea [10:15:30 CET] <Timothy_Gu> ubitux: cool. will push later today [10:22:47 CET] <cone-252> ffmpeg 03Timothy Gu 07master:5f1aad68c4f5: tests: Add test for proper header guard [11:40:55 CET] <rcombs> these Firefox guys& [11:47:03 CET] <BtbN> Why don't they just support a defined version, and include it in their binary releases? oO [11:47:30 CET] <nevcairiel> who knows [11:47:37 CET] <nevcairiel> for some reason they dont want to ship any binary [11:47:45 CET] <nevcairiel> so they have one evil hack to support everything [11:47:49 CET] <BtbN> So no ffmpeg on windows, ever? [11:48:00 CET] <BtbN> Not even an lgpl one? [11:50:19 CET] <cone-252> ffmpeg 03Paul B Mahol 07master:408ea50ca6bc: avfilter/f_metadata: fix few logic errors [11:56:25 CET] <Daemon404> what does Outreachy provide if not money? [11:56:33 CET] <Daemon404> admin? [11:57:00 CET] <rcombs> nevcairiel: BtbN: patents [11:59:22 CET] <Daemon404> nevcairiel / BtbN - he's in here if you wanna argue ;) [11:59:32 CET] <nevcairiel> i just want to make fun of them [11:59:40 CET] <nevcairiel> i havent used firefox in years [11:59:47 CET] <nevcairiel> not starting any time soon either [12:00:18 CET] <Daemon404> i switched back from chrome ~2 years ago [12:00:25 CET] <Daemon404> when chrome iddnt support highdpi correctly [12:00:29 CET] <Daemon404> too lazy to switch back [12:00:44 CET] <Daemon404> plus chrome is ie6 [12:01:04 CET] <durandal_1707> Daemon404: it provides devs doing merges [12:01:15 CET] <Daemon404> ? [12:01:26 CET] <durandal_1707> Outreachy [12:01:37 CET] <Daemon404> how does it provide devs? they devs apply. [12:03:52 CET] <nevcairiel> Daemon404: the only thing they do is admin yes, which really kind of limits its usefulness [12:04:26 CET] <Daemon404> thats what i thought [12:04:32 CET] <Daemon404> nevcairiel, where is our i/o callback api? [12:04:47 CET] <nevcairiel> its called open_cb i think [12:04:51 CET] <nevcairiel> same place the new one goes [12:04:53 CET] <nevcairiel> avformatcontext [12:05:19 CET] <Daemon404> wait.... [12:05:29 CET] <Daemon404> so out formatcontext->opaque is used for multiple things then [12:05:40 CET] <Daemon404> while it is only documented for the avdevice garbage [12:12:38 CET] <Daemon404> looks like ours is much less in scope [12:12:47 CET] <Daemon404> is anyone going to rip my face off for deprecating it [12:12:55 CET] <Daemon404> michaelni, ^ opinions [12:12:58 CET] <Daemon404> since youre teh author [12:13:14 CET] <nevcairiel> how old is that thing anyway [12:13:51 CET] <Daemon404> Date: Mon May 11 17:45:13 2015 +0200 [12:14:23 CET] <Daemon404> it's only used for mov crefs, mlvdec, and img2dec [12:14:54 CET] <Daemon404> drefs* [12:17:59 CET] <durandal_1707> replace with what? [12:18:30 CET] <nevcairiel> two new callbacks for open and close [12:19:19 CET] <Daemon404> nevcairiel, we never had two callbacks [12:19:25 CET] <nevcairiel> i never said we did [12:19:34 CET] <michaelni> Daemon404, ask wm4, i think he was the primary user of the callback [12:19:40 CET] <Daemon404> wm4, ^ [12:19:48 CET] <wm4> didn't use it [12:19:53 CET] <Daemon404> lol. [12:19:59 CET] <wm4> (although planned to) [12:20:04 CET] <wm4> (still plan to) [12:20:10 CET] <nevcairiel> may'15 means it was never in a release? or did it make 2.8? [12:20:54 CET] <Daemon404> github search for av_format_set_open_cb mostly finds avformat.def files... [12:21:07 CET] <michaelni> its in 2.7 and 2.8 [12:21:13 CET] <nevcairiel> aw [12:21:26 CET] <michaelni> see git branch --contains 541d75f9a0b6e1b360345e289cb44e43a39643cd [12:21:34 CET] <Daemon404> TIL: --contains [12:21:53 CET] <nevcairiel> i knew it exists, but i'm usually too lazy to use it =p [12:21:58 CET] <Daemon404> i can't actually find any real project that uses it [12:22:04 CET] <Daemon404> maybe kodi does? [12:22:08 CET] <Daemon404> (probably not) [12:22:28 CET] <wm4> maybe none does [12:22:57 CET] <Daemon404> i think it's relatively safe to deprecate... [12:23:06 CET] <Daemon404> im not quite ready to break api with libav... [12:23:17 CET] <Daemon404> i'd rather need it compat atm [12:28:51 CET] <durandal_1707> Daemon404: haven't you stopped doing merges? [12:28:59 CET] <Daemon404> no? [12:30:27 CET] <durandal_1707> Daemon404: bunch of commits are not merged [12:30:33 CET] <Daemon404> thats nice [12:30:43 CET] <Daemon404> i took a short break. libav only pushed like 2 commits in the meantime. [12:30:52 CET] <Daemon404> can anyone explain what this AVIOInterruptCB stuff peppered everywhere? [12:30:57 CET] <Daemon404> is* [12:32:50 CET] <wm4> Daemon404: not freeze the process eternally if you really want to close the avio/avformat context [12:33:02 CET] <nevcairiel> its for controlling timeout, yea [12:33:10 CET] <nevcairiel> but libav has that too [12:33:18 CET] <Daemon404> does it? ok [12:33:34 CET] <nevcairiel> maybe not all the same calls in all protocol code, but the same mechanism anyway [12:33:46 CET] <Daemon404> just noticed because basically every single place that libav is usign open_cb has some whtielist opening func [12:33:51 CET] <Daemon404> which takes that callback [12:34:05 CET] <Daemon404> wrappers for wrappers for wrappers for wrappers <.. continue for 10 layers...> [12:35:32 CET] <durandal_1707> should I write lua filter that exposes all lavfi internals, so you can write your own filters in lua? [12:35:35 CET] <wm4> the callbacks needs to be passed down to every sub-context that is opened [12:35:40 CET] <wm4> and the whitelists too [12:35:55 CET] <wm4> durandal_1707: not before C does? [12:37:29 CET] <durandal_1707> do not need C [13:00:14 CET] <durandal_1707> i want to reply to mats thread, with something funny [13:00:59 CET] <ubitux> "yes" [13:05:33 CET] <wm4> man this hwaccel effort duplication between ffmpeg and libav is terrible [13:05:48 CET] <wm4> now elenril wrote he has his own partial vaapi implementation [13:07:32 CET] <Daemon404> ffmpeg doesnt exist man [13:08:27 CET] <wm4> he actually looked at the ffmpeg ones and voiced some unclear complaints [13:08:41 CET] <Daemon404> since i dont work on playback, i can only watch others suffering [13:08:51 CET] <wm4> this is for transcoding too [13:08:58 CET] <wm4> (the fabled full-hw transcode pipeline) [13:11:49 CET] <Daemon404> i dont know why id want to do that at scale [13:12:04 CET] <Daemon404> because i love pissin money away? [13:12:31 CET] <wm4> everyone knows that GPUs should do all real work, how can you doubt that [13:13:24 CET] <Daemon404> gpgpgpgpgpgpu etc [13:21:04 CET] <nevcairiel> full hw transcode is useless because hw encoders are terrible =p [13:22:13 CET] <Daemon404> that too [13:29:33 CET] <cone-252> ffmpeg 03Michael Niedermayer 07master:a82ff49bb829: ffmpeg_filter: Add YUV 444 to strict_std_compliance case for mjpeg [13:29:34 CET] <cone-252> ffmpeg 03Michael Niedermayer 07master:82d2aa2b32e8: ffmpeg_filter: Add missing pixel formats to strict_std_compliance case for ljpeg [13:29:35 CET] <cone-252> ffmpeg 03Michael Niedermayer 07master:d94b11a72138: ffmpeg_filter: Factor get_compliance_unofficial_pix_fmts() out [13:39:12 CET] <rcombs> nevcairiel: it's only useful on shitty consumer hardware [13:46:16 CET] <rcombs> wm4: link to this libav vaapi work? [13:46:21 CET] <rcombs> (I'm trying to keep track of all this) [13:48:26 CET] <wm4> rcombs: https://lists.libav.org/pipermail/libav-devel/2016-February/074667.html [13:48:33 CET] <wm4> the vaapi part isn't public [13:49:09 CET] <rcombs> ah [13:50:30 CET] <rcombs> wm4: is there a solution to the format conversion/auto-inserted scale filter thing in there [13:51:47 CET] <wm4> don't think so [13:51:55 CET] <wm4> ask elenril [13:52:12 CET] <rcombs> appears afk [14:10:55 CET] <Daemon404> michaelni, ping [14:11:05 CET] <Daemon404> i have a question: re avio+whitelisting [14:15:45 CET] <jkqxz> wm4: I saw that new libav patchset and was intending to go through it more carefully soon. [14:16:11 CET] <wm4> ok [14:18:32 CET] <jkqxz> I'm still missing how you do the enumerate-all-possible-render-targets for initialisation, and the copy in/out setup probably needs to be a bit more refined. [14:19:02 CET] <Daemon404> michaelni, hmm nevermind [15:01:57 CET] <kierank> phew that ticket is ancient [15:02:30 CET] <kierank> there really should be any more crashes with h264 and bad streams now. afaik I found them all and michaelni fixed them [15:02:50 CET] <BBB> h264 is fairly stable yes [15:02:56 CET] <BBB> werent most of the crashes in the mt handling? [15:06:47 CET] <kierank> yes iirc [15:11:16 CET] <kierank> But $bigsportingevent over the last few days was decoded with ffh264 and there were no crashes at all [15:11:32 CET] <kierank> so at least my revision from July is solid [15:15:07 CET] <rcombs> oh? ffh264 was used in the broadcast? [15:15:13 CET] <rcombs> in what capacity? [15:16:25 CET] <ubitux> well no crash with the same kind of input doesn't mean much [15:16:31 CET] <kierank> ubitux: the input was switched a lot [15:16:45 CET] <ubitux> yes but switch to what kind of input [15:16:56 CET] <ubitux> i mean, it's more interesting in a situation like youtube [15:17:00 CET] <kierank> between h264 and mpeg2 and between different h264 encoders [15:17:01 CET] <kierank> yes ofc [15:17:19 CET] <ubitux> where you have so many users whose purpose is to inject a maximum of insane shit [15:17:31 CET] <kierank> Sure but going black to 20 million people is also of concern [15:18:01 CET] <ubitux> i'd say you're pretty sure to long forever if you can handle one minute [15:18:14 CET] <kierank> rcombs: receivers in certain territories [15:18:26 CET] <rcombs> neat [15:19:11 CET] <BtbN> are you talking about the superbowl stream? [15:19:18 CET] <Daemon404> static int read_header(AVFormatContext *avctx) [15:19:20 CET] <Daemon404> this is a gem [15:19:39 CET] <BtbN> Tried to watch it in Kodi, and it did some weird stuff every few minutes. [15:20:18 CET] <nevcairiel> Daemon404: how so, that could be any function in any files, those static things are generally not named very well [15:21:18 CET] <ubitux> happens in 26 [15:21:20 CET] <Daemon404> nevcairiel, avctx is almost universally used for avcodecocntext [15:21:38 CET] <ubitux> ah, that [15:21:45 CET] <ubitux> issue present in mlv and tty [15:21:45 CET] <nevcairiel> eh, type safety will protect you =p [15:22:09 CET] <nevcairiel> using "s" for any kind of context isnt that much better [15:22:23 CET] <nevcairiel> or just plain "ctx" [15:22:42 CET] Action: Daemon404 thinks he finally finished the io stuff [15:22:43 CET] <Daemon404> fate time... [15:27:11 CET] <Daemon404> can we like, rate limit Mats? [15:29:02 CET] <nevcairiel> that would be nice, wouldnt it [15:36:26 CET] <Daemon404> man we have 4 sets of vsynth now? [15:36:32 CET] <Daemon404> no wonder FATE's been so slow for me [15:36:52 CET] <nevcairiel> 4 now? i thought 3 [15:37:07 CET] <nevcairiel> oh right we duplicated some because of lena [15:37:12 CET] <nevcairiel> but those barely take any time for me [15:38:20 CET] <Daemon404> theyre by far the longest part of fate for me [15:43:09 CET] <cone-252> ffmpeg 03Anton Khirnov 07master:9f61abc8111c: lavf: allow custom IO for all files [15:43:10 CET] <cone-252> ffmpeg 03Derek Buitenhuis 07master:bc9a5965c815: Merge commit '9f61abc8111c7c43f49ca012e957a108b9cc7610' [16:08:02 CET] <BBB> Daemon404: --disable-encoders [16:09:21 CET] <Daemon404> sometimes vsynth catches lavf issues [16:09:23 CET] <Daemon404> so i dont [16:12:02 CET] <BBB> that suggests we need more lavf tests [16:13:20 CET] <Daemon404> is anyone super familiar with hls.c? [16:13:35 CET] <Daemon404> help would be appreciated on merging 81306fd4bdeb5c17d4db771e4fec684773b5790f [16:13:42 CET] <Daemon404> our hls.c seems to be 10x more complex [16:22:36 CET] <JEEB> hmm [16:22:58 CET] <JEEB> I wonder what of all thngs can be controlled through stdin/signals with ffmpeg cli [16:23:07 CET] <Daemon404> ? [16:23:40 CET] <JEEB> f.ex. if the encode was started with -re but later I can remove that flag [16:23:44 CET] <JEEB> :D [16:23:53 CET] <JEEB> I'll probably need to use the API for this kind of stuff [16:24:22 CET] <Daemon404> you really can only go so far with script ffmpeg cli [16:24:30 CET] <JEEB> yeah [17:07:57 CET] <nevcairiel> someone should remove all of those projects that we've been carrying around for 10 GSoCs [17:10:09 CET] <wm4> that would probably start an edit war [17:10:39 CET] <nevcairiel> some are so old that the proposed mentors probably dont even remember anymore [17:11:46 CET] <Daemon404> nevcairiel, people would object because then the list iwll be tiny [17:11:50 CET] <Daemon404> and nobody has any ideas [17:12:03 CET] <nevcairiel> and no students work on those stupid old ideas anyway [17:12:20 CET] <nevcairiel> because they are either way too complex or not very interesting [17:12:52 CET] <Daemon404> https://trac.ffmpeg.org/wiki/SponsoringPrograms/GSoC/2016#MissingAACdecoder… [17:12:56 CET] <Daemon404> this one doesnt even exist anymore [17:13:01 CET] <Daemon404> i think all of those got done [17:13:03 CET] <nevcairiel> decoder [17:13:06 CET] <nevcairiel> and no [17:13:11 CET] <Daemon404> no? [17:13:12 CET] <nevcairiel> 960/120 is still not supported, neither is SSR [17:13:27 CET] <nevcairiel> no idea what BSAC is [17:13:29 CET] <Daemon404> im pretty sure i saw someone add the window [17:14:30 CET] <nevcairiel> http://git.videolan.org/?p=ffmpeg.git;a=blob;f=libavcodec/aacdec_template.c… [17:14:40 CET] <nevcairiel> 960 is only supported in eld streams, not in LC [17:15:16 CET] <Daemon404> ah ok [17:15:20 CET] <Daemon404> still doesnt seem like it could fill a gsoc [17:18:05 CET] <kierank> Shall I start a thread about it [17:18:35 CET] <kierank> yes I will [17:21:48 CET] <kierank> hmm why did sending to ffmpeg-users(a)ffmpeg.org not work [17:22:00 CET] <nevcairiel> isnt it ffmpeg-user [17:22:03 CET] <kierank> oh bleh [17:22:25 CET] <nevcairiel> outreachy actually started with a nearly empty list of tasks [17:22:29 CET] <kierank> yes [17:22:32 CET] <nevcairiel> or did someone copy-paste the old one over by now [17:22:32 CET] <kierank> because I did it like that [17:27:03 CET] <BBB> kierank: why dont you just delete the old tasks [17:27:08 CET] <kierank> because carl [17:27:12 CET] <BBB> screw him [17:27:20 CET] <kierank> I emailed him about it [17:27:22 CET] <BBB> for each newly added task, the only persont hat can add it is the mentor [17:27:27 CET] <BBB> no mentor, no task [17:27:34 CET] <BBB> I have a better idea then [17:27:38 CET] <BBB> lets keep the task list as is [17:27:41 CET] <BBB> abandon it for any future use [17:27:45 CET] <kierank> to be fair aac encoding was an old one that actually got sorted [17:27:48 CET] <BBB> and make a new one that is actually, you know, useful [17:27:54 CET] <Daemon404> ok so [17:27:58 CET] <BBB> I dont think aac encoding got sorted b/c it was on a task list [17:28:08 CET] <Daemon404> the obvious one hasnt been added, because it has no mentor and is too hard :P [17:28:12 CET] <Daemon404> (hevc asm( [17:28:14 CET] <nevcairiel> BBB: atomnuker came to us through GSoC [17:28:21 CET] <BBB> Im willing to write hevc asm [17:28:24 CET] <BBB> but someone needs to pay me [17:28:32 CET] <nevcairiel> 1c per instruction? [17:28:37 CET] <durandal_1707> how much? [17:28:40 CET] <nevcairiel> hm thats counter productive [17:28:45 CET] <nevcairiel> you would write slow verbose code [17:28:54 CET] <bencoh> :D [17:29:27 CET] <BBB> any company wiht a big pocket of cash and sincere interest in hevc asm is free to contact me [17:30:04 CET] <BBB> (Ive repeated this every few months over the past 1-2 years, and only one company ever contacted me, and they were hoping Id do it for free) [17:30:11 CET] <Daemon404> seems unlikely [17:30:11 CET] <nevcairiel> should ask netflix, i hear they stream in hevc these days [17:30:12 CET] <BBB> (from which I gather there is just no interest) [17:30:15 CET] <Daemon404> chicken/egg/patent pools [17:30:22 CET] <j-b> BBB: how much? [17:30:28 CET] <Daemon404> orite vlc [17:30:35 CET] <BBB> j-b: companies can contact me, but not cheap [17:30:46 CET] <kierank> nevcairiel: they use hw decoders [17:30:52 CET] <j-b> that's very very vague [17:30:52 CET] <BBB> it makes no sense to use ngo money for that [17:31:01 CET] <BBB> it is, I dont want you to pay for my manhattanitis [17:31:04 CET] <j-b> cheap could be 1000¬ or 100000¬ [17:31:05 CET] <BBB> its my fault, not yours [17:31:08 CET] <nevcairiel> kierank: proper hevc hw decoders are still somewhat rare in hardware though [17:31:19 CET] <fritsch> j-b: calculate in "time" [17:31:20 CET] <nevcairiel> at least consumer side [17:31:44 CET] <ubitux> kierank: cfhd, i think you should s/av_clip_uintp2_c/av_clip_uintp2/ [17:31:45 CET] <nevcairiel> although that will probably improve this year [17:31:57 CET] <kierank> ubitux: breaks arm because not a power of 2 [17:32:18 CET] <nevcairiel> isnt that function specifically designed for powers of 2 [17:32:19 CET] <nevcairiel> :D [17:32:30 CET] <ubitux> ... we still haven't a proper fallback? [17:32:48 CET] <ubitux> nevcairiel: the actual problem is the value being const or not [17:32:54 CET] <kierank> oh wasn't a power of 2 [17:32:57 CET] <kierank> yeah was const [17:32:58 CET] <kierank> sorry [17:33:23 CET] <nevcairiel> ubitux: sounds like a useless limitation, just throw out some half-baked optimization that only works in some cases [17:33:29 CET] <BBB> j-b: rough guess - review of all existing simd (sao, lf, mc) and writing missing (idct) for 8bit/10bit (thatll give you 12bit essentially for free, but nobody cares), sse2, ssse3, avx and avx2, 32bit and 64bit compat, $150k [17:33:33 CET] <jamrial> seriously, just check av_builtin_constant_p() in the arm version [17:33:49 CET] <jamrial> i doubt there's a compiler we support that has inline asm but not that [17:34:06 CET] <BBB> oh I forgot intra pred [17:34:08 CET] <fritsch> BBB: so ne year in sum [17:34:10 CET] <BBB> anyway [17:34:15 CET] <fritsch> s/ne/one/ [17:34:17 CET] <BBB> fritsch: I dont work hourly [17:34:27 CET] <fritsch> companies always want a timeline [17:34:30 CET] <BBB> fritsch: its like writing per instruction, counter incentivizing [17:34:55 CET] <fritsch> so you could tell them: pay me for a year (150K) and after the year you have results [17:34:57 CET] <kierank> well either we'd want hourly or we'd want per X [17:35:09 CET] <kierank> gotta put something tangible on invoice [17:35:18 CET] <BBB> I work per project, not per hour [17:35:20 CET] <jamrial> it would most likely need to review and probably rewrite parts of the actual decoder. mc is kinda messy as is [17:35:26 CET] <BBB> so the project would be what I just said [17:35:56 CET] <fritsch> yeah - if you decide to pause your work after 3 months for 10 years :-) [17:36:45 CET] <nevcairiel> well clearly there would be some idea of timeline on such a rpoject [17:36:45 CET] <atomnuker> do I need to bump libavcodec MINOR/MICRO when pushing my VC-2 encoder? [17:36:50 CET] <BBB> atomnuker: yes [17:36:57 CET] <jamrial> minor, yes [17:37:02 CET] <atomnuker> kthnx [17:37:08 CET] <ubitux> nevcairiel: i think rcombs had a patch [17:37:45 CET] <ubitux> 2015-11-05 12:13:32 rcombs https://gist.github.com/e5c55681a10f59ec192f duplicated implementation since it would produce circular header dependencies otherwise [17:40:31 CET] <BBB> nevcairiel: I typically add deadlines for the project as a whole. doing a deadline for particular subtasks would be fine also, but I dont provide specifications of how many hours Ill work on X, Y or Z [17:40:41 CET] <BBB> if its not done 6 months from now, the contract will be void [17:40:43 CET] <BBB> or so [17:41:44 CET] <BBB> anyway, Ill keep my eyes open for companies using ffhevc for whatever [17:42:04 CET] <Daemon404> there are plenty, but they all violate the gpl anway [17:42:06 CET] <Daemon404> anyway* [17:42:08 CET] <Daemon404> so no $ [17:42:58 CET] <BBB> no $ is a problem for contract work [17:43:26 CET] <kierank> cehoyos: hello, so we were proposing to start from a clean project list for gsoc/outreachy [17:43:33 CET] <BBB> hello carl [17:43:36 CET] <kierank> and asking the on the ML for new ideas [17:44:35 CET] <cehoyos> Hi! [17:44:50 CET] <cehoyos> I am all for new ideas, I don't think starting with a new page makes much sense. [17:45:09 CET] <cehoyos> If you believe that the current page contains bad ideas, please suggest to remove them. [17:45:54 CET] <cehoyos> Starting with a clean page has the additional disadvantage that we would have to think of all the little things that should be part of the page, like contact information, student help etc. [17:46:05 CET] <kierank> The problem is it's the same dated ideas [17:46:18 CET] <cehoyos> I really don't understand why old ideas are bad? [17:46:31 CET] <cehoyos> On the contrary, I think they get more important every year... [17:46:33 CET] <kierank> imagine a student visiting ffmpeg gsoc every year and seeing the same thing [17:46:39 CET] <kierank> just makes us look like a boring project [17:46:41 CET] <kierank> that doesn't do anything new [17:46:46 CET] <Daemon404> if nobody picked them for 5 years [17:46:51 CET] <Daemon404> chances are nobody will [17:47:11 CET] <cehoyos> As said; Please add new ideas, the more we have the better! [17:47:30 CET] <cehoyos> (And more new ideas will make us look less boring iiuc.) [17:47:51 CET] <kierank> it's better to have a few good ideas than a ton of mediocre ones [17:48:10 CET] <cehoyos> I wonder if Stephanie agrees... [17:48:22 CET] <kierank> "C coding skills, good x86 assembly coding skills, basic familiarity with git." too much to ask these days for example [17:48:52 CET] <Daemon404> not these days [17:48:54 CET] <durandal11707> C# and java [17:48:55 CET] <cehoyos> Please add simpler tasks! But in your email - iiuc - you asked for more difficult tasks, or did I misread? [17:48:55 CET] <Daemon404> it was always niche [17:49:01 CET] <BBB> python [17:49:07 CET] <kierank> no I asked for more interesting tasks [17:49:13 CET] <kierank> the tasks we have are the same boring old crap [17:49:21 CET] <kierank> mpeg-4 als [17:49:22 CET] <cehoyos> More interesting than vc-1, j2k and mvc? Good luck! [17:49:23 CET] <BBB> I do agree things like postproc optimizations are not very hot [17:49:28 CET] <BBB> mpeg4 als isnt hot either [17:49:30 CET] <kierank> How does this interest me as a student? [17:49:35 CET] <kierank> that's what you need to think about [17:49:39 CET] <jamrial> and do most of these even have mentors anymore? [17:49:39 CET] <BBB> we need tasks that teach students about stuff theyll use in their phd [17:49:44 CET] <BBB> for video coding, that is vp9/hevc [17:49:47 CET] <durandal11707> port vapoursynth mvtools? [17:49:51 CET] <BBB> for audio coding, that is likely opus or stuff [17:49:59 CET] <BBB> I dont know what filtering research is hot these days [17:50:05 CET] <cehoyos> Let's please stop this, it is not very constructive: Just add "hot" ideas and by itself, the boring ones will just disappear. [17:50:08 CET] <BBB> but vc1-interlaced is not hot [17:50:12 CET] <kierank> cehoyos: no the boring ones won't [17:50:15 CET] <kierank> that's the problem [17:50:16 CET] <BBB> not unless we remove them [17:50:20 CET] <kierank> exactly [17:50:28 CET] <BBB> stuff only disappears if it gets removed [17:50:37 CET] <BBB> bugs dont vanish in thin air [17:50:44 CET] <BBB> neither do uncomplated gsoc tasks in a todo list [17:50:50 CET] <BBB> uncompleted* [17:51:03 CET] <cehoyos> Please let's start with adding tasks, do not remove them because you don't want to work on them (if you were a student) [17:51:15 CET] <kierank> you've just contradicted yourself [17:51:32 CET] <cehoyos> Note that als has a mentor, some of you are not listed on the GSoC wiki (yet) [17:51:38 CET] <Daemon404> "has: [17:51:45 CET] <Daemon404> the mentors are copypasted year after year [17:51:48 CET] <durandal11707> they dont want to be mentors or students [17:51:51 CET] <kierank> copy and paste doesn't could carl [17:51:58 CET] <kierank> count* [17:52:07 CET] <Daemon404> e.g. ive been copypasted 3 years in a row as a mentor withotu even being consulted once [17:52:19 CET] <Daemon404> of course nobody ever picks that project [17:52:26 CET] <durandal11707> ok, who wants to be mentor? [17:52:29 CET] <cehoyos> Sorry, I just started a few minutes ago: I will send mails to every single one on the list. [17:52:40 CET] <kierank> cehoyos: don't put them on the list unless they agree [17:52:41 CET] <cehoyos> But I wanted to tell you that for one of the projects you mentioned, we have a mentor. [17:52:42 CET] <BBB> durandal_1707: depends on project [17:52:43 CET] <kierank> the copy and paste is bad [17:53:13 CET] <kierank> 4:49 PM <"BBB> we need tasks that teach students about stuff theyll use in their phd [17:53:15 CET] <kierank> or get them jobs [17:54:56 CET] <kierank> he-aac encoder is another task [17:55:09 CET] <Daemon404> lol phd [17:55:11 CET] <cehoyos> Will you mentor he-aac? [17:55:21 CET] <kierank> no but we can find people [17:55:30 CET] <kierank> cehoyos: it's not a tickbox contest you know [17:55:35 CET] <kierank> the mentor needs to be the right person [17:55:38 CET] <kierank> not the first one you find [17:56:26 CET] <ubitux> simple tasks are either done on a regular basis, or boring as hell [17:56:44 CET] <ubitux> and interesting ones require skill [17:56:49 CET] <ubitux> or involve large api redesign [17:57:42 CET] <BBB> kierank: yeah, very true, phd or jobs [17:58:07 CET] <BBB> hobby stuff is fine also, but I have issues with the concept that people would find vc1-interlaced interesting as a hobby [17:58:09 CET] <kierank> cehoyos: you can copy my task from outreachy [18:00:21 CET] <cehoyos> Thank you, done! [18:00:35 CET] <jamrial> BBB: i found a shitty audio only container that was used in one game a hobby once :p [18:00:41 CET] <cehoyos> ubitux: Will you mentor a subtitle task? [18:00:46 CET] <ubitux> no, sorry [18:00:49 CET] <BBB> jamrial: thats not suitable as a gsoc task :-p [18:02:58 CET] <kierank> j-b, wm4: what are your complaints that gsoc could solve? [18:03:07 CET] <durandal11707> why is truehd encoder still task? [18:03:14 CET] <kierank> durandal11707: because we gotta have tasks [18:03:15 CET] <cehoyos> Do we have one? [18:03:16 CET] <kierank> taskssssss [18:03:33 CET] <kierank> need tasks, it's like gold dubloons [18:03:40 CET] <BBB> lol [18:03:46 CET] <wm4> something that a student could do... [18:06:14 CET] <j-b> kierank: asm/hevc, outputting NV12 directly in software decoders, [18:06:39 CET] <wm4> what's that nv12 one for? [18:07:10 CET] <kierank> nv12 > planar basically [18:07:15 CET] <kierank> x264 is nv12 because it's faster [18:07:18 CET] <kierank> and hw uses nv12 [18:07:22 CET] <j-b> yep. [18:07:22 CET] <jamrial> a student wrote the hevc decoder. another can surely write a truehd encoder [18:07:26 CET] <wm4> but I suppose it also needs more code [18:07:47 CET] <j-b> not really :) [18:07:50 CET] <kierank> jamrial: not the point. The point is hevc was appealing at the time (coool, new next-gen codec) [18:07:55 CET] <kierank> truehd has been there for years [18:07:57 CET] <kierank> we don't have a spec [18:07:58 CET] <j-b> well, it's trivial to interleaving [18:10:34 CET] <jamrial> anyway, as j-b said, assembly. hevc/cfhd/dca/aac/mp3/opus, for x86 and/or arm/aarch64 [18:10:42 CET] <BBB> encoders and decoders are completely different beasts [18:11:01 CET] <nevcairiel> i'm not sure which is easier tbh [18:11:15 CET] <nevcairiel> with an encoder you can just start with something naive and super inefficient [18:11:18 CET] <BBB> I briefly looked at nv12, and the problem is youll have to rewrite much of the assembly for chroma-only [18:11:27 CET] <BBB> MC, for example [18:11:30 CET] <nevcairiel> with a decoder you have no choice but to support everything the spec asks for :D [18:11:31 CET] <BBB> its really no fun at all [18:11:49 CET] <j-b> BBB: so it's easier to reshuffle U, V planes? [18:12:03 CET] <jamrial> nevcairiel: "whoops, unimplemented" -> return -1 [18:12:03 CET] <BBB> its certainly less code, yes [18:12:05 CET] <jamrial> :p [18:12:52 CET] <BBB> I can see how sao/loopfilter would be simple to do in interleaved chroma, possibly even reusing the luma code (not sure about that, just hypothesizing) [18:13:06 CET] <BBB> but mc/intrapred and probably even inverse transform would need completely new code [18:13:09 CET] <BBB> that is very uncool [18:13:13 CET] <jamrial> our hevc decoder reports unimplemented features with some of the conformance suit samples, even [18:13:17 CET] <nevcairiel> in my particular case, I can't really do zero copy for software decoding without many complicated problems, so I accepted having one memcpy, and instead of doing a plain memcpy I wrote SIMD to re-shuffle the yuv420 -> NV12 in the process [18:13:26 CET] <nevcairiel> its not any slower than a plain memcpy [18:13:33 CET] <nevcairiel> so it practically comes for free [18:13:46 CET] <BBB> uv420p->nv12 chroma reshuffling tends to be super-cheap, yes [18:13:58 CET] <nevcairiel> now if your architecture allows zero-copy, that might be more of a problem [18:14:12 CET] <BBB> its minimal compared to the overhead of actual decoding, really [18:14:18 CET] <BBB> for modern codecs like h264, hevc and vp9 [18:14:42 CET] <BBB> I think for x264 thats different because of the amount of calls you do in the r/d loop, and uv interleaving halves that for chroma parts [18:14:52 CET] <BBB> code complexity is less of an issue for encoders, its even expected [18:14:59 CET] <BBB> but for decoders I dont know if it makes sense [18:15:04 CET] <nevcairiel> even for x264 its a tiny gain only, but then 2% are 2% [18:15:11 CET] <BBB> I used to be all for it, but after having tried it, I decided to give up fairly quickly [18:15:17 CET] <cone-252> ffmpeg 03Michael Niedermayer 07master:23261e600149: sws/output: fix ordered dither threshold for mono output [18:15:18 CET] <cone-252> ffmpeg 03Michael Niedermayer 07master:21b459e4bbfc: avformat/segment: Fix header_filename handling [18:15:26 CET] <BBB> I dont think youd get 2% for a decoder, sadly [18:17:22 CET] <BBB> j-b: still planning to move to canada? [18:18:43 CET] <j-b> at some point, sure. [18:19:29 CET] <kierank> meh [18:22:35 CET] <cone-252> ffmpeg 03Rostislav Pehlivanov 07master:4701be7198eb: options_table: update maximum bitrate limit [18:22:36 CET] <cone-252> ffmpeg 03Rostislav Pehlivanov 07master:ec9e87c922ad: avcodec: add a native SMPTE VC-2 HQ encoder [18:22:37 CET] <cone-252> ffmpeg 03Rostislav Pehlivanov 07master:135460383e93: avformat: add vc2 as an allowed rawenc Dirac extension [18:23:25 CET] <atomnuker> rejoice, a low delay, low overhead video encoder has been added [18:24:24 CET] <atomnuker> technically not made for anything else than broadcasting stuff, but works great as a screen recoding codec [18:27:27 CET] <wm4> that's awesome [18:28:09 CET] <wm4> is this libschroedinger thing still needed? [18:28:53 CET] <kierank> wm4: yes [18:28:56 CET] <kierank> if you care about inter [18:44:44 CET] <atomnuker> I still find the primitive quantization hilarious, it's just a left shift and a divide [18:50:31 CET] <atomnuker> shit, fate-source fails, on it [19:00:32 CET] <kierank> What is fate-source [19:01:27 CET] <jamrial> checks every source file to see if they have proper copyright headers [19:03:30 CET] <jamrial> it's also very slow. takes like a minute to finish on a core i5 [19:03:36 CET] <atomnuker> nope, I just used non-project-standard includion guards [19:03:42 CET] <atomnuker> which is what fate-source also checks [19:04:23 CET] <jamrial> on msys at least. it seems to spawn a lot of processes (sh, git, etc) [19:05:03 CET] <cone-252> ffmpeg 03Rostislav Pehlivanov 07master:5669aa2a8ac5: vc2enc: use project-standard inclusion guards [19:08:39 CET] <ubitux> having decoders output exclusively nv12 instead of yuv420p is going to have a negative impact any time filters are involved [19:08:48 CET] <ubitux> unless you actually update most of them to support nv12 [19:10:58 CET] <nevcairiel> jamrial: on my system its not really noticeable slow, but its on a ssd, so that may help [19:15:26 CET] <cone-252> ffmpeg 03Michael Niedermayer 07master:a73b23e3df15: avformat/hlsenc: Fix filename and options [19:17:18 CET] <rcombs> ubitux: maybe we should have better autonegotiation there [19:17:37 CET] <rcombs> like, decoder could output nv12 if the first filter in the chain supports it; otherwise yuv420p [19:17:57 CET] <ubitux> sure, but keeping 2 paths in the decoder might be annoying [19:18:42 CET] <ubitux> but after all, nv12 is annoying to deal with in filters [19:21:05 CET] <rcombs> inherently, or just because a lot of code already targets yuv420p [19:25:17 CET] <durandal11707> ubitux: what? first wilter will convert to what is supports, next filter uses that what first give [19:26:17 CET] <durandal11707> *filter :) [19:26:37 CET] <ubitux> yes but then you have a convert [19:27:01 CET] <ubitux> so you do a roundtrip to sws when there wasn't any before [19:32:07 CET] <durandal11707> atomnuker: copyright is ok? [19:33:38 CET] <cehoyos> Thanks to Kierans comment, we can now guess what is meant, I still don't understand why it isn't clearly explained... [19:33:43 CET] <durandal11707> dirac native is slooooooooooooooooooow [19:34:03 CET] <cehoyos> Bye! [19:34:17 CET] <atomnuker> durandal11707: decoding or encoding? [19:34:32 CET] <durandal11707> atomnuker: decoding, vc2!=dirac [19:35:07 CET] <durandal11707> encoding hd720 is slightly faster than realtime, decoding is nowhere realtime [19:35:16 CET] <atomnuker> yeah, decoding is still slow as piss (~80% of the time goes just to read golomb) [19:36:26 CET] <atomnuker> function is libavcodec/diracdec:coeff_unpack_golomb(), if anyone's interested [19:46:05 CET] <nevcairiel> high bitrate will do that for you [19:46:11 CET] <nevcairiel> most time spent reading the coeffs [19:46:48 CET] <atomnuker> yeah, but it's oddly asymmetric [19:47:06 CET] <atomnuker> writing is so fast [19:48:48 CET] <atomnuker> then again for writing there's a LUT [19:55:09 CET] <Timothy_Gu> durandal11707: any comments on blend asm? [19:55:55 CET] <durandal11707> it is bit exact? whats gain? [20:02:09 CET] <Timothy_Gu> it is [20:02:23 CET] <Timothy_Gu> checkasm reports about 10x boost IIRC [20:02:32 CET] <Timothy_Gu> (checkasm patch pending) [20:02:42 CET] <Timothy_Gu> durandal11707: ^^ [20:03:22 CET] <Timothy_Gu> nop: 28.4 [20:03:22 CET] <Timothy_Gu> screen_c: 3568.5 [20:03:23 CET] <Timothy_Gu> screen_sse2: 313.0 [20:05:29 CET] <cone-252> ffmpeg 03Paul B Mahol 07master:5486d7fa91f7: avfilter/dualinput: use pts provided by framesync [20:11:05 CET] <jrosser> someone still needs to factorise the vc2 golomb unpacking [20:11:29 CET] <jrosser> or alterntively i need to write a paper on how to do it [20:22:07 CET] <durandal11707> don is back and kicking! [20:26:36 CET] <cone-252> ffmpeg 03Timothy Gu 07master:c8b1612af03b: x86/vf_blend: Move multiplying to a macro [20:26:37 CET] <cone-252> ffmpeg 03Timothy Gu 07master:74f8d9aaef91: x86/vf_blend: Add SSE2 optimization for screen [20:28:07 CET] <furkan> atomnuker: what's this low-delay encoder? is it possible to reduce compression (for lower CPU usage) at the expense of bandwidth for streaming over a LAN? [20:28:32 CET] <RiCON> furkan: vc-2 [20:30:16 CET] <furkan> interesting [20:32:58 CET] <furkan> i asked this in the main channel, but i've spent hours trying to figure out why i get this behaviour when streaming high bitrate h264 (>6Mbps) over UDP: https://www.dropbox.com/s/8s6m1ssi1wowbky/IMAG0096.jpg?dl=0 [20:33:25 CET] <furkan> there's no packet loss, no congestion, and TCP works fine... if anybody has tips on how i can debug, i'd appreciate it [20:35:05 CET] <furkan> i'd like to make use of multicasting, with TCP that's obviously not an option [20:54:32 CET] <jrosser> furkan: maybe counterintuitively reducing the compression will increase the cpu usage for vc-2, as the total number of bits which need to be packed/unpacked will go up [21:03:02 CET] <furkan> interesting [21:48:05 CET] <atomnuker> furkan: it [21:48:26 CET] <atomnuker> furkan: it's low sense in that it can potentially only transmit parts of the image [21:48:41 CET] <atomnuker> instead of an entire image [21:48:57 CET] <atomnuker> so you can potentially start displaying the image as you recieve the first piece/slice [21:50:03 CET] <atomnuker> I say potentially since you can't do that yet (libavcodec only works on whole frames, you'd need to rip the encoder out of it) [21:50:25 CET] <Daemon404> avcodec has a way to do it line-by-line [21:50:27 CET] <Daemon404> but not slices [21:50:52 CET] <durandal11707> wtf avfilter_graph_parse2 two gotos [21:51:21 CET] <wm4> fail:end: [21:51:22 CET] <wm4> nice [21:51:23 CET] <atomnuker> Daemon404: nice to know [21:51:32 CET] <wm4> Daemon404: what do you mean? [21:51:34 CET] <Daemon404> draw_horiz_band or w/e [21:51:51 CET] <wm4> that can output arbitrary units [21:51:56 CET] <wm4> (I think) [21:51:58 CET] <Daemon404> orly [21:52:00 CET] <Daemon404> didnt know [21:52:31 CET] <wm4> well it can output multiple lines at once [21:52:41 CET] <wm4> which is why the function has "band" in the name [21:52:52 CET] <Daemon404> sounds like what atomnuker wants then [21:53:02 CET] <atomnuker> yep, exactly [21:54:07 CET] <atomnuker> wm4: know any examples? [21:54:53 CET] <wm4> not really... mplayer can use it lol [21:55:00 CET] <wm4> kierank also tried to play with it once? [21:55:18 CET] <wm4> just search for draw_horiz_band [21:55:23 CET] <Daemon404> one crappy lossless codec i write supports it [21:55:23 CET] <Daemon404> cant rememebr whuch [21:55:29 CET] <Daemon404> wrote* [21:56:55 CET] <TD-Linux> atomnuker, neat, is there any other comparable encoder in ffmpeg right now? [21:58:25 CET] <atomnuker> TD-Linux: I guess VC-3 (dnxhd?) is in the same niche, and JPEG2000 is basically a direct competitor [21:58:49 CET] <TD-Linux> I didn't know there was a dnxhd encoder in ffmpeg [21:59:44 CET] <atomnuker> there is, AFAIK quite a lot of the broadcasting world uses that encoder to encode stuff to store [22:01:00 CET] <jrosser> j2k is high computational complexity, comparitively [22:01:05 CET] <jrosser> all that bit plane stuff [22:02:27 CET] <gnafu> VC-3/DNxHD is basically a variant of MJPEG, right? [22:02:33 CET] <gnafu> (To simplify it a bit, of course.) [22:02:58 CET] Action: Daemon404 wonders what vc-4 was [22:04:33 CET] <atomnuker> Daemon404: I did too until a few days ago [22:04:56 CET] <atomnuker> some SMPTE invention which never got outside of a press release [22:05:05 CET] <Daemon404> olol [22:05:39 CET] <wm4> isn't ac-4 bitter reality? [22:10:09 CET] <JEEB> at least the spec was posted somewhere [22:10:27 CET] <JEEB> it's mostly a reality due to the fact that the AC-3 patents are going to run out [22:11:33 CET] <TD-Linux> "It's hard to believe that Dolby's lossy AC-3 audio codec is nearly thirty years old" [22:11:34 CET] <TD-Linux> no not really [22:12:27 CET] <atomnuker> they could resort to patent evergreening [22:13:09 CET] <atomnuker> I mean it's still used in pretty much all DVDs/BDs [22:13:25 CET] <wm4> patent everwhat? [22:14:26 CET] <atomnuker> "Evergreening refers to a variety of legal and business strategies by which technology producers with patents over products that are about to expire retain royalties from them, by either taking out new patents (for example over associated delivery systems, or new pharmaceutical mixtures), or by buying out or frustrating competitors, for longer periods of time than would normally be permissible under the [22:14:28 CET] <atomnuker> law" [22:15:12 CET] <atomnuker> so as far as I understand that, you can save your royalties by buying related but applicable patents from competitors (which for some reason you haven't bought yet) [22:15:32 CET] <wm4> how wonderful [22:16:49 CET] <TD-Linux> yeah I wonder what actually happened to ac-4. presumably it's not in TrueHD because it's decodeable as ac3? [22:17:56 CET] <TD-Linux> err I meant atmos [22:21:20 CET] <wm4> AV_PIX_FMT_GBRAP32FBE [22:21:22 CET] <wm4> yummy [22:21:51 CET] <nevcairiel> its floating point, do we really need LE and BE there as well [22:22:58 CET] <wm4> I just complained about this [22:24:20 CET] <Compn> whatever happened to NE ? :P [22:36:04 CET] <rcombs> atomnuker: the AC-3 patents have already lasted longer than they should've since they filed some late [22:36:21 CET] <rcombs> and then there's eac3 [23:13:40 CET] <BBB> michaelni: what is F"? [23:13:44 CET] <BBB> michaelni: are these float formats? [23:14:26 CET] <BBB> ah yes they are [00:00:00 CET] --- Thu Feb 11 2016

1 0

ffmpeg.log.20160210
by burek 11 Feb '16

11 Feb '16

[00:00:25 CET] <relaxed> lee: great, so it worked? If so, I'm going to add a blurb to the ffmpeg vp8 wiki. [00:00:48 CET] <lee> yep, it worked great [00:04:00 CET] <J_Darnley> Ha. What a stupid format. [00:11:16 CET] <explodes_> Hello! - After an av_seek_frame, I'll still (sometimes) av_read_frame(..) a video frame at the previous position instead of only getting frames from the new positio [00:19:49 CET] <kadiro> hello, can i convert an encrypted (drm) wmv to any other format ? [00:23:21 CET] <kadiro> I mean, Is it possible [00:41:03 CET] <FilipeMaia> Does anyone know of a good cmake FindFFMPEG module? [01:10:47 CET] <mzbotr> Hi, I'm trying to recode some video streams to fit a display aspect ratio of 16:9 for DVD playback without stretchiness... when I'm doing this, I'm getting errors saying the data is not aligned. Is this worth troubling myself over? [01:10:55 CET] <mzbotr> [swscaler @ 0xe9fc80] Warning: data is not aligned! This can lead to a speedloss [01:12:28 CET] <J_Darnley> I should submit a message to silence that warning [01:12:48 CET] <J_Darnley> uha patch [01:12:56 CET] <mzbotr> ain't nothing? [01:15:27 CET] <J_Darnley> It just means assembly can't use movdqa but rather movdqu [01:15:56 CET] <J_Darnley> Unaligned moves aren't that slow (anymore) and metters less today with copious cpu poer. [01:16:00 CET] <J_Darnley> *power [01:29:24 CET] <dugz> Hello everybody! I have some hundreds of random video files, and I want to transcode them into a Roku-friendly format for streaming over minidlna. Any thoughts on what my target format should be? [01:29:51 CET] <dugz> I'm planning on using find and ffmpeg to do the conversion [01:30:02 CET] <ThomQ> Hi all. IŽm trying to render a transparent WebM video, from a series of transparent.pngŽs. The video IŽm getting does not contain any transparency, nor does it play well in certain players (while other transparent .webm videos I downloaded do work in those players) [01:30:16 CET] <kadiro> just googled about "Tunebite" can remove drm, wonder if i can do same with ffmpeg like -c:v copy -c:a copy ? [01:30:42 CET] <ThomQ> Im using ffmpeg -f image2 -i wow%3d.png -vf fps output.webm as a command. As a n00bie to FFMPEG I think I might have missed something. Anyone any idea? [01:32:08 CET] <ThomQ> I would like to add that Im not bound to WebM, the transparency is the most important part [01:33:24 CET] <furq> is it transparent video day today [01:33:56 CET] <furq> 21:21:23 ( relaxed) lee: try, -pix_fmt yuva420p -metadata:s:v:0 alpha_mode="1" [01:33:58 CET] <furq> ThomQ: ^ [01:34:53 CET] <ThomQ> Is transparent video all the hype? ;) [01:34:58 CET] <ThomQ> Thanks, trying it now. [01:35:05 CET] <furq> this is the second time someone's asked this exact question today [01:35:10 CET] <furq> and also the second time i've seen it asked ever [01:38:13 CET] <furq> dugz: from a brief google, the only worthwhile video codec the roku supports is h264 [01:38:27 CET] <furq> so x264+aac i guess [01:38:38 CET] <furq> in mkv [01:38:41 CET] <dugz> furq, I saw that too, but mine seems to choke on matroska files [01:38:51 CET] <ThomQ> ffmpeg -f image2 -i wow%3d.png -vf fps output.webm -pix_fmt yuva420p -metadata:s:v:0 alpha_mode="1" doesnt seem to work for me [01:39:00 CET] <furq> ThomQ: where are you playing the video [01:39:13 CET] <ThomQ> its not transparent and just black in one player [01:39:15 CET] <furq> afaik it needs to be in a video tag with background: transparent; [01:39:28 CET] <furq> if you just load the video up in chrome it'll have a coloured background [01:39:31 CET] <ThomQ> Furq: Chrome among other things [01:39:56 CET] <furq> also afaik it'll only work in chrome, it doesn't work in firefox [01:40:02 CET] <furq> and i doubt it works in safari or IE [01:40:12 CET] <ThomQ> im loading it on a simple php page that does load another webm video correctly [01:40:23 CET] <ThomQ> the rendered one still has the black background though [01:40:51 CET] <ThomQ> yeah it only works in Chrome / Chromium and other forks [01:41:28 CET] <furq> oh [01:41:34 CET] <furq> you need to specify those options before the output filename [01:44:22 CET] <ThomQ> ffmpeg -f image2 -i wow%3d.png -vf fps -pix_fmt yuva420p -metadata:s:v:0 alpha_mode="1" output.webm gets the same results. It is playing in all players im using though [01:44:34 CET] <ThomQ> PNGs are confirmed transparent btw :D [01:45:09 CET] <furq> pastebin the ffmpeg command and full output [01:45:13 CET] <furq> maybe someone will know better than i do [01:47:27 CET] <ThomQ> okok [01:47:28 CET] <ThomQ> http://pastebin.com/HDTM3yKv [01:47:31 CET] <lee> ThomQ: the exact line that I used (which worked for me): ffmpeg -i zt%04d.png -c:v libvpx -pix_fmt yuva420p -metadata:s:v:0 alpha_mode="1" -b:v 1000k output.webm [01:48:00 CET] <furq> oh weird, it's using vp9 [01:48:06 CET] <furq> `-c:v libvpx` should fix it then [01:48:11 CET] <furq> i thought vp8 was still the default for webm [01:48:29 CET] <kadiro> hola hello allo [01:48:40 CET] <ThomQ> Thanks Lee! Testing it now [01:48:52 CET] <ThomQ> Furq, where do i put that? Also in front of the output name? [01:48:57 CET] <furq> yeah [01:49:07 CET] <ThomQ> Trying them both :) [01:49:13 CET] <furq> also you will want to specify a bitrate (-b:v) because the default is 200kbps, which is probably less than you want for 720p [01:49:18 CET] <lee> ThomQ: -b:v 1000k <-- target bitrate, adjust as necessary [01:49:26 CET] <furq> -b:v 0 -crf 30 is probably better [01:50:22 CET] <furq> or whatever the equivalent of x264 crf 21 is (the best crf) [01:51:21 CET] <relaxed> ThomQ: https://trac.ffmpeg.org/wiki/Encode/VP8 [01:51:47 CET] <ThomQ> Ahh k thanks guys. Yeah i was planning on fiddling with the bitrate and adding audio as soon as I had the base. Im too n00b for crf though [01:53:25 CET] <furq> crf is easier than specifying a bitrate [01:53:58 CET] <lee> furq: thanks, I'll try that next time (having read relaxed's explanation of it on the wiki page) [01:54:14 CET] <ThomQ> furq: Thanks, it worked! [01:54:18 CET] <furq> i think low 30s are roughly analogous to low 20s in x264 but don't take my word for it [01:54:43 CET] <ThomQ> i compared it to leeŽs command and fixed my typo [01:55:01 CET] <furq> yeah alpha isn't supported in vp9 yet for some reason [01:55:15 CET] <ThomQ> Trying CRF now, it did look a lot worse then what i rendered before [01:55:33 CET] <ThomQ> -b:v 0 -crf 30 instead of -b:v? [01:55:44 CET] <DeHackEd> that's vp or h264? [01:55:46 CET] <ThomQ> -c:v i mean [01:55:48 CET] <furq> vp8 [01:56:00 CET] <furq> -b:v 0 -crf 30 is correct [01:56:04 CET] <furq> lower crf values are higher quality [01:56:07 CET] <ThomQ> and thanks Relaxed, that page helps alot [01:56:41 CET] <furq> also you still need -c:v libvpx in case that wasn't clear [01:57:01 CET] <relaxed> I only added the alpha section to the wiki page, fyi [01:57:09 CET] <ThomQ> it wasnt, but it was already erroring [01:57:17 CET] <ThomQ> added it now :) [01:59:00 CET] <ThomQ> yeah it looks great furq, thanks [01:59:29 CET] <xbloodhound> one day my head is just gonna pop with all these video encoding stuff [02:00:07 CET] <lee> it's not just encoding, it's all the cool shit you can do with ffmpeg [02:00:21 CET] <xbloodhound> yea [02:00:49 CET] <lee> I could do with a matrix-style "woah! I know ffmpeg" moment though, to just know what options to use (or even what they all are, then work on when to apply them) [02:04:57 CET] <ThomQ> another question: if I add audio (-i wow.wav) it seems to be a little out of sync (audio is about half a second too early) [02:05:25 CET] <ThomQ> Im thinking it probably has something to do with framerate, that would be logical, right?| [02:05:46 CET] <ThomQ> the source is 29.97fps, but i havent declared that anywhere in the command line for FFMPEG [02:06:03 CET] <kadiro> hello, can i convert an encrypted (drm) wmv to any other format or copy just video/audio stream ? [02:07:49 CET] <furq> ThomQ: if it's consistently half a second out then it's nothing to do with framerate [02:08:01 CET] <ThomQ> its too short too really tell furq [02:08:19 CET] <furq> `-af adelay=500` or something [02:08:44 CET] <ThomQ> does -crf 30 set the frame rate btw? [02:08:47 CET] <furq> no [02:08:58 CET] <xbloodhound> hell no [02:09:02 CET] <furq> for the image demuxer you'd set -framerate 30 before the input file [02:09:08 CET] <kadiro> furq, can you help? [02:09:11 CET] <ThomQ> ahh k, so I added -r 30, which should theoraticly rule out delay due to difference in fps [02:09:22 CET] <ThomQ> Ahh k, thanks furq [02:10:14 CET] <ThomQ> Aaand it fixed it, itŽs perfect now [02:10:21 CET] <ThomQ> thank you guys so much, Furq you in particular [02:10:47 CET] <furq> it beats working [02:11:16 CET] <dugz> i'm trying to use winff with ffmpeg, but i can't seem to find a preset for 16:9 480p. am I just being retarded? [02:13:10 CET] <kadiro> personne ? [02:14:09 CET] <kadiro> can i resume that ffmpeg can't do that? [02:15:36 CET] <kadiro> furq, did you hear me? see me, me vois ?? [02:16:17 CET] <kadiro> !test [02:16:58 CET] <relaxed> kadiro: ffmpeg won't work with encrypted files [02:17:17 CET] <kadiro> relaxed, thank you man you saved my time [02:17:59 CET] <kadiro> relaxed, any open source can you suggess to do that? [02:19:04 CET] <relaxed> I don't know of any [02:19:14 CET] <kadiro> ok thank you [02:22:36 CET] <xbloodhound> can anyone tell me whats difference between -crf and -qp? [02:22:57 CET] <xbloodhound> or rather what used more for [02:25:11 CET] <furq> xbloodhound: http://slhck.info/articles/crf [03:18:03 CET] <pingpong1109> So I am trying to stream to a livestreaming service. This is the command I am doing and the output http://pasteben.benbaptist.com/?id=2270181225612 [03:18:35 CET] <pingpong1109> Also I would like to be able to stream the video to two destinations [03:20:04 CET] <furq> [flv @ 0x24e6c30] Video codec rawvideo not compatible with flv [03:20:12 CET] <furq> there's a hint for you [03:20:40 CET] <pingpong1109> I see that but can't find a solution. Unless I am missing something [03:21:09 CET] <furq> you're doing -vcodec copy with /dev/video0 [03:21:18 CET] <furq> you need to reencode it to a format that is compatible with flv [03:21:24 CET] <furq> such as -c:v libx264 [03:21:44 CET] <furq> (vcodec is deprecated, use c:v) [03:22:02 CET] <pingpong1109> Ah ok I think it worked. Let me verify [03:23:26 CET] <pingpong1109> All it is doing is buffering [03:25:00 CET] <pingpong1109> Yeah it crashed, I think a pi 2 isn't pwoerful enough to do this [03:26:08 CET] <furq> the pi 2 has a builtin h264 encoder but you can't use it from ffmpeg [03:26:20 CET] <furq> i believe gstreamer can use it [03:26:43 CET] <furq> you could maybe try -c:v libx264 -preset ultrafast [03:27:38 CET] <pingpong1109> .5 FPS ftw [03:28:27 CET] <pingpong1109> No that didnt work [03:28:55 CET] <furq> you should get more than .5 fps at 480p [03:29:19 CET] <pingpong1109> How do I define a resolution? [03:29:27 CET] <furq> it's already using 480p [03:29:47 CET] <furq> if you want something smaller then -vf scale=320:240 [03:29:52 CET] <pingpong1109> https://gyazo.com/a1692a360020a94090d4b5b456d16015 [03:30:35 CET] <relaxed> pingpong1109: -c:v flv1 might be faster [03:31:37 CET] <furq> i suspect something else is wrong [03:31:54 CET] <pingpong1109> http://pasteben.benbaptist.com/?id=3971927114762 [03:31:58 CET] <pingpong1109> Yeah it didn't work [03:32:15 CET] <pingpong1109> Guess this site doesnt like that [03:33:01 CET] <relaxed> don't cherry pick errors, pastebin the command and all output [03:33:26 CET] <furq> is that command actually fully loading the cpu [03:34:08 CET] <pingpong1109> http://pasteben.benbaptist.com/?id=6144378073513 [03:34:44 CET] <furq> "./stream.sh" isn't a command line which is useful to us [03:34:52 CET] <furq> Codec AVOption preset (Set the encoding preset (cf. x264 --fullhelp)) specified for output file #0 (rtmp://ingest.wdc01.beam.pro:1935/beam/) has not been used for any stream. The most likely reason is either wrong type (e.g. a video option with no video streams) or that it is a private option of some encoder which was not actually used for any stream. [03:35:06 CET] <pingpong1109> working on copying the actual command [03:35:23 CET] <pingpong1109> ffmpeg -i /dev/video0 -c:v libx264 -preset ultrafast -an -r 10 -f flv -metadata streamName=player1 rtmp://ingest.wdc01.beam.pro:1935/beam/ [03:35:49 CET] <relaxed> that's not the command you used [03:35:49 CET] <pingpong1109> ffmpeg is only using 5%ish [03:36:35 CET] <pingpong1109> ffmpeg -i /dev/video0 -c:v flv1 -preset ultrafast -an -r 10 -f flv -metadata streamName=player1 rtmp://ingest.wdc01.beam.pro:1935/beam/ [03:37:02 CET] <relaxed> ok, now omit -preset ultrafast (which is a libx264 option) [03:37:09 CET] <furq> if i had to guess i'd say there's a problem with your webcam or whatever you're using [03:38:22 CET] <pingpong1109> I am using an old webcam but it works in other computers, I can just use an old laptop since this isn't working [11:49:59 CET] <maziar> i have a problem in concept , i want to create a web site like youtube, and i configure nginx with rtmp and mp4 and pseudo-streaming as my web server, but for converting i don't know how should i do? should i use ffserver ? is ffserver good for me for serving 240, 780 and 1080 quality ? [12:23:13 CET] <xbloodhound> I am getting Hardware accelerated decoding with frame threading is not supported when I try it am I missing something or it just won't do it with what I have setup? [12:27:42 CET] <durandal_1707> xbloodhound: hw acceleartion doesn't make sense with multithreading [12:32:04 CET] <xbloodhound> guess I can't take load off CPU then [12:38:01 CET] <Filarius> hello, is here demuxer in FFMPEG what will allow me to make easy muxing in my application and send fo FFMPEG video and audio in single annon pipe ? [12:38:45 CET] <Filarius> like, packet = video-frame, audio-frame [12:38:59 CET] <jkqxz> xbloodhound: What are you actually trying to do? Hardware accelerated decoding (necessarily with no threading) may still be faster than fully-threaded software decoding using all of the CPU. [12:39:33 CET] <Filarius> I want to develop application what will send raw video and mp3(or raw) sound to ffmpeg for encoding [12:40:20 CET] <Filarius> but with annon pipe I can send only one - or video or audio, and I need to send both at realtime [12:41:06 CET] <Filarius> I'm curious if it is possible with single pipe, to not use 2 names pipes [12:41:12 CET] <Filarius> *named pipes [12:41:42 CET] <votz> Filarius: ffmpeg only has one stdin [12:42:18 CET] <jkqxz> Filarius: You will need to pack it into some container. Most of the supported ones are pipe-friendly. [12:43:07 CET] <Filarius> yep, stdin, thats what i'm talking about, just i'm not sure in words so I name it "annon pipe" [12:43:32 CET] <votz> Filarius: if you can't touch the filesyste, you could always use the network. ffmpeg -i http://localhost:8080 -i http://localhost:8081 [12:43:46 CET] <Filarius> i'm not sure if here friendly packing modules, need to search for one [12:44:11 CET] <Filarius> yeh, I had this idea too [12:44:37 CET] <Filarius> its will be my "back door" [12:44:49 CET] <Filarius> if i will not find way with pipe [12:45:57 CET] <durandal_1707> you could use multiple pipes [12:47:26 CET] <votz> durandal_1707: if I understand Filarius correctly, he doesn't want to use multiple pipes: I'm curious if it is possible with single pipe, to not use 2 names pipes [12:47:50 CET] <xbloodhound> jkqxz: I was converting video here is gist of what I am using "ffmpeg -hwaccel dxva2 -i INPUT -c:v h264 -c:a aac OUTPUT" I was hoping to be able to convert in same way some of my recording software using GUI to convert video files but I guess with what you told me its just not possible with ffmpeg [13:26:41 CET] <Ccdc_DuckZ> hello, I'm having a segfault in avcodec_encode_video2(), this is the code I have http://pastebin.com/TijqCnJy [13:27:19 CET] <Ccdc_DuckZ> I'm not sure I initialized all I had to, I wrote that following the sample code in the docs [13:27:35 CET] <Ccdc_DuckZ> or maybe I'm passing in the wrong thing? any clue? [14:41:33 CET] <FilipeMaia> Hi. Im trying to compress some frames to h264. Im using -qp 0 and -preset medium for lossless encoding. Whenever I encode less than 100 frames the compression is much worse than when I encode 100 or more. There is a discontiniuty at 100. Any idea why? [14:42:43 CET] <furq> FilipeMaia: does the same thing happen with other source videos [14:43:56 CET] <FilipeMaia> With the videos I have available, yes [14:44:50 CET] <FilipeMaia> It looks like internally something changes when I encode more than 99 frames [14:46:57 CET] <FilipeMaia> Ok Ill continue to dig, around. I might be messing up something else [15:21:16 CET] <J_Darnley> FilipeMaia: are you saying that 101 frames makes a smaller video than 100 frames? [15:21:40 CET] <J_Darnley> I might understand compression improving from the Nth frame depending on lookahead and frame threads [15:21:40 CET] <FilipeMaia> J_Darnley: yes, but theres probably a bug somewhere in my code [15:22:03 CET] <J_Darnley> Oh, you mean I can't checkwith ffmpeg? [15:22:26 CET] <J_Darnley> I will try anyway [15:22:37 CET] <FilipeMaia> Im having problems reproducing it myself now [15:22:49 CET] <FilipeMaia> Now I always get the larger sizes [15:23:58 CET] <FilipeMaia> Does anyone know if av_image_alloc zeros the allocated memory? [15:26:51 CET] <J_Darnley> Heh. I guess "read the source" is the answer [15:27:07 CET] <FilipeMaia> I guess so [15:27:16 CET] <FilipeMaia> Looks like it doesnt. That could explain it [15:28:22 CET] <Bluez_> hi guys [15:28:57 CET] <Bluez_> im using libav apis to put raw AAC into a container, but the sound is choppy, any ideas what it might be? Ive checked my timestamps and they seem ok [15:29:18 CET] <J_Darnley> No encoding? [15:29:45 CET] <J_Darnley> Does AAC need one of the bitstream filters? [15:32:10 CET] <Bluez_> oh maybe it does duh! [15:33:04 CET] <FilipeMaia> J_Darnley: my problem was uninitialized Cb and Cr [15:42:15 CET] <J_Darnley> Ah "random noise" is hard to compress [20:24:03 CET] <furkan> is there any way that i can debug why i get this behaviour when streaming high bit-rate h264 over a LAN? https://www.dropbox.com/s/8s6m1ssi1wowbky/IMAG0096.jpg?dl=0 [20:24:13 CET] <furkan> it only happens over UDP, not TCP [20:40:48 CET] <TD-Linux> furkan, that's just packet loss and ffmpeg's plain UDP has no way to fix it [20:43:45 CET] <furkan> TD-Linux: but there's no packet loss on the network, it's on a LAN [20:43:57 CET] <furkan> i've tried iperf [20:44:24 CET] <TD-Linux> furkan, how do you know there's no packet loss? [20:44:35 CET] <furkan> because i've tested with iperf [20:45:23 CET] <TD-Linux> and it's always 0% packet loss at the rates you are using with h264? [20:45:24 CET] <furkan> i can run other tests if necessary, but based on iperf the packets arrive in correct order [20:46:08 CET] <furkan> is there any way to tell if there are re-transmissions when using TCP? [20:46:34 CET] <explodes> Anyone familiar with the lib? I'm still having problems seeking to the position I specify. I want to av_seek_frame to AVSEEK_FLAG_BACKWARD and then read ahead until I'm at the frame I need to be: http://pastebin.com/TyX2kEBt [20:46:54 CET] <TD-Linux> I believe wireshark can tell you, but I've never done it personally [20:48:45 CET] <furkan> ok so i just ran another test [20:49:11 CET] <furkan> sent 28987 datagrams (40.6MB), 0% packet loss, 1 datagram received out of order [20:49:26 CET] <furkan> 34.1Mbit/sec [20:50:16 CET] <furkan> same results at 80Mbps [20:50:39 CET] <furkan> 57476 datagrams, 80.6MB total, 1 datagram received out of order [20:54:09 CET] <jkqxz> Are packets being thrown away on the sender because your instantaneous spikes of packets (per-frame?) are larger than the UDP buffering on your system? [20:56:07 CET] <furkan> jkqxz: i did find some discussion about that online so i tried increasing the kernel's UDP buffer size but that didn't help [20:56:26 CET] <furkan> i'm trying to find out if i can increase any of ffmpeg's buffers [20:58:49 CET] <explodes> After I read a video frame do I *have* to call "avcodec_decode_video2" every time to avoid a future avcodec_decode_video2 call to receive busted images? [21:01:02 CET] <kepstin> for modern video codecs, predicted frames rely on having the decoded versions of previous frames available [21:01:44 CET] <jkqxz> furkan: Similarly the receive buffer on the receive side, then? [21:02:22 CET] <furkan> jkqxz: actually yeah it's the receive side i'm looking at, the sending side is an IP camera [21:03:22 CET] <kepstin> explodes: so for the seeking case, you have to seek to a keyframe (intra-frame), then decode frames and discard the result until you're at the frame you want. [21:04:05 CET] <explodes> nice, that's a solid plan. Do I "decode frames and discard" by putting "av_read_frame" in a while loop? [21:04:09 CET] <explodes> or is it some other function? [21:04:25 CET] <jkqxz> And you are definitely getting all of the packets on the receive-side? (That is, something else can receive the UDP stream and decode it completely, so the camera send isn't the problem.) [21:06:56 CET] <furkan> jkqxz: could i figure that out with tcpdump? [21:07:25 CET] <kepstin> explodes: interestingly, the ffmpeg command-line tool does it by inserting a trim filter [21:08:59 CET] <explodes> kepstin: hmmm strange [21:09:04 CET] <explodes> or maybe not strange [21:09:59 CET] <jkqxz> What is the UDP format? If it's RTP, for example, they have sequence numbers in and you can easily tell whether some are missing. [21:10:02 CET] <explodes> what do you call it when video frames aren't "fully" rendered, glitching? [21:10:47 CET] <FastJack_> hi [21:12:13 CET] <kepstin> explodes: I guess so. there's certainly people who do that on purpose to create "glitch art" :) [21:12:49 CET] <DHE> "bad" frames? [21:12:51 CET] <explodes> this works flawlessly for advancing the video, except that the video frame are glitched until the next keyframe: http://pastebin.com/ZyJW5MJU [21:12:59 CET] <explodes> DHE let me record a video for yaaa [21:13:16 CET] <furkan> jkqxz: it's RTSP, but i could do RTP as well, if i can capture the data with tcpdump and inspect it that way it would be ideal [21:13:19 CET] <DHE> oh I know what it looks like. I've made bad seek indexes in the past for fun [21:13:22 CET] <explodes> oh wait nvm, I think you know what I'm talking about [21:13:25 CET] <explodes> hehe [21:13:28 CET] <furkan> or actually i suppose i could always import the dump into wireshark on my PC [21:13:37 CET] <kepstin> explodes: yeah, that's the normal behaviour if you start decoding at a frame that's not a keyframe. You get fun motion prediction artifacts :) [21:14:00 CET] <explodes> kepstin: well first I av_seek_frame(...., AVSEEK_FLAG_BACKWARD) then advance forward :( [21:14:20 CET] <explodes> it theoretically shouldn't be full of glitches like at all [21:14:37 CET] <jkqxz> furkan: I don't think tcpdump can show the right information without difficulty, but importing it into wireshark would definitely be able to show it. [21:15:40 CET] <furkan> jkqxz: great thanks for the tip, i'll collect some data next time and hopefully it sheds some light on what's going on [21:16:20 CET] <kepstin> explodes: it doesn't look like you're actually decoding all the frames? [21:16:58 CET] <kepstin> reading the packets isn't sufficient, you actually have to run them through the decoder before discarding them. [21:17:03 CET] <explodes> oh shit- do i have to call avcoded_decode_video2 and avcoded_decode_audio4 ? [21:17:08 CET] <explodes> gotcha [21:18:16 CET] <kepstin> you also might have to account for out-of-order frames (dts != pts) [21:18:34 CET] <kepstin> I think you should be checking the pts on the frames after decoding, since I think the decoder will reorder them? [21:19:35 CET] <explodes> ok [21:36:06 CET] <explodes> Here it is, completed. http://pastebin.com/ydk0cPAH I'm using packet.dts because the frame's PTS was coming back unset, i'd be using pkt_pts/pkt_dts anyways [21:37:32 CET] <explodes> Thanks for the help! I've been working on this for a while now [21:59:53 CET] <effractur> hi [22:00:11 CET] <effractur> i am trying to use the libdc1394 [22:00:17 CET] <effractur> to connect to a firewire camera [22:00:22 CET] <effractur> but i get the following error [22:00:30 CET] <effractur> Can't find matching camera format for uyvy422, 320x240@10000:1000fps [22:00:46 CET] <effractur> any idea how to get it working? [00:00:00 CET] --- Thu Feb 11 2016

1 0

ffmpeg-devel.log.20160209
by burek 10 Feb '16

10 Feb '16

[00:03:30 CET] <llogan> or make it part of an outreachy/gsoc qualification task [00:07:10 CET] <llogan> i have no idea which project ideas to add to the new outreachy wiki page [00:19:13 CET] <kierank> a bit hard for an outreachy/gsoc [00:19:17 CET] <kierank> qualification [00:19:22 CET] <kierank> but not big enough for a full task [00:20:49 CET] <jkqxz> nevcairiel: Are you sure the mr[345]_tandberg streams really work properly with dxva2 (no error output, and not falling back to software because of the profile)? As far as I can tell, h264 does call the alloc function more times than it should. [00:23:01 CET] <Timothy_Gu> kierank: get outreachy girls to do it [00:23:13 CET] <kierank> see what I just wrote [00:23:34 CET] <llogan> not limited to only women anymore. [00:23:47 CET] <Timothy_Gu> oh [00:24:23 CET] <Timothy_Gu> Additionally, they are open to residents and nationals of the United States of any gender who are Black/African American, Hispanic/Latin@, American Indian, Alaska Native, Native Hawaiian, or Pacific Islander. [00:24:42 CET] <Timothy_Gu> im neither residents or national of the U.S. or any of those "minorities" [00:24:53 CET] <Timothy_Gu> ugh [00:25:01 CET] <llogan> looks like i qualify as "ambigously somewhat brownish guy" [00:25:32 CET] <Timothy_Gu> alaska native haha [00:25:44 CET] <kierank> llogan: yeah but imo we're not outreaching [00:26:19 CET] <llogan> do you mean we aren't going to tell them we want to participate this round? [00:27:08 CET] <llogan> Timothy_Gu: "We are planning to expand the program to more participants from underrepresented backgrounds in the future." [00:27:22 CET] <llogan> perhaps you can join the Minority Club later [00:28:04 CET] <kierank> llogan: I mean if you (llogan) were to apply to outreachy we wouldn't be doing any outreach since you were here before [00:28:17 CET] <Timothy_Gu> llogan: im pretty sure the next group will be white rather than Asian [00:28:29 CET] <Timothy_Gu> technically i'm not being outreached either [00:28:49 CET] <llogan> kierank: oh, heh. not that i would contribute anything interesting. [00:30:29 CET] <kierank> llogan: can you start making the 2016 page [00:30:49 CET] <llogan> working on it now [00:31:12 CET] <llogan> listing you as admin and michael and me as backup [00:31:53 CET] <kierank> ah ok [00:31:58 CET] <kierank> I have one task idea [00:32:11 CET] <kierank> is there a place you are editing it? [00:33:37 CET] <kierank> Should I include the url in my email? [00:33:46 CET] <Timothy_Gu> fyi, id really like to have a new documentation engine [00:34:31 CET] <llogan> Timothy_Gu: https://trac.ffmpeg.org/wiki/SponsoringPrograms/ProjectIdeas#Automaticdocum… [00:34:40 CET] <llogan> feel free to make less bad description, etc [00:34:44 CET] <Timothy_Gu> ah [00:34:48 CET] <Timothy_Gu> lol [00:35:35 CET] <llogan> kierank: fflogger just barfed the url. you can include url although it is incomplete [00:35:39 CET] <kierank> yeah [00:36:45 CET] <kierank> I'm going to try and improve the opening marketing [00:36:49 CET] <Timothy_Gu> i thibk we should make sure what kind of approach we want BEFORE the program itself [00:37:23 CET] <Timothy_Gu> so that tr ibtern can start doing it as soon as possible w/o all the drama of discussing it [00:37:35 CET] <Timothy_Gu> *the intern [00:43:54 CET] <llogan> kierank: currently opw@ffmpeg email goes to michael, timnich, reynaldo, stefano, reimar, nicolas george, klaussfreire, and paul b mahol. i can change it if you like (see /etc/aliases on ffbox0) [00:44:04 CET] <kierank> can you add to me [00:44:17 CET] <kierank> no idea if I have ffbox0 access [00:45:03 CET] <llogan> should i remove the others and just add you me and michael for now? [00:45:43 CET] <kierank> yeah ok [00:46:53 CET] <llogan> ok. i used to know how to do this off the top of my head... i forgot how to do anything postfix since it's been so long. [00:49:32 CET] <michaelni> you edit the aliases file and run newaliases [00:51:00 CET] <michaelni> llogan, ^ [00:51:39 CET] <llogan> thanks. sounds familiar. looking at postfix man pages remined me why i moved everything to fastmail [00:53:55 CET] <llogan> kierank: do you prefer kunhya or obe email? [00:54:00 CET] <kierank> kunhya [00:56:36 CET] <llogan> i updated the alias and ran newaliases [00:58:21 CET] <kierank> I'll email with obe email because that's subscribed to outreachy [00:58:24 CET] <kierank> but it shouldn't matter [01:19:31 CET] <llogan> why is the trac database locked? [01:20:26 CET] <kierank> dunno [01:20:39 CET] <Timothy_Gu> "locked"? [01:21:43 CET] <kierank> sqlite does that sometimes [01:21:45 CET] <llogan> just trying to submit chages to the outreachy wiki page. "Trac detected an internal error: OperationalError: database is locked" [01:23:39 CET] <kierank> can't login [01:23:56 CET] <kierank> ah can [01:23:58 CET] <kierank> just a bit slow [01:24:02 CET] <Timothy_Gu> ubitux: is "tests: Add test for proper header guard" ok to commit? [01:24:19 CET] <llogan> kierank: might be getting slammed by spambots. [01:25:00 CET] <llogan> i wonder how gitlab is working out for videolan (code.videolan.org) [01:25:59 CET] <cone-484> ffmpeg 03Timothy Gu 07master:cb8646af24bd: configure: Enable GCC vectorization on e4.9 on x86 [01:28:08 CET] <llogan> unfortunate timing with the email [01:28:50 CET] Action: kierank restarts apache [01:29:11 CET] <kierank> seems happy now [01:29:17 CET] <llogan> thanks [01:34:53 CET] <cone-484> ffmpeg 03Thierry Foucu 07master:020b75806f6e: lavf/mov: Extend extracting XMP in mov files using UUID Box [02:10:34 CET] <llogan> 1500 tapes and only 4 had fungus issues [03:17:07 CET] <Compn> hmm [03:17:38 CET] <Compn> wonder if llogan was talking about vhs [03:50:54 CET] <jamrial> http://pastebin.com/Revwi2dr these warnings started showing up with gcc 6 after the tree vectorization patch [03:50:57 CET] <jamrial> doesn't happen with gcc 5 [03:51:05 CET] <jamrial> what are the chances they are bogus? [03:51:35 CET] <jamrial> lcov reports that the entire file is used in a fate run [04:01:41 CET] <Timothy_Gu> jamrial: my bet is on gcc bug [04:01:45 CET] <Timothy_Gu> :sigh: [04:03:02 CET] <jamrial> most likely, yeah [04:11:24 CET] <J_Darnley> Ha ha ha. I was browsing the "C string handling" page on wikipedia. We get a mention for allegedly reimplementing strlcat/strlcpy. [11:50:22 CET] <jkqxz> nevcairiel: The required number of surfaces for H.264 is num_ref_frames * threads in degenerate cases. [11:51:23 CET] <jkqxz> I made a test file to show the problem in the software decoder: <http://ixia.jkqxz.net/~mrt/evil.264>. [11:51:40 CET] <nevcairiel> even more reason not to use it =p [11:52:02 CET] <jkqxz> This has num_ref_frames=16. The stream is a black IDR frame followed by all-skip inter frames with frame_num decreasing. [11:52:44 CET] <jkqxz> Then this patch instruments malloc to see the large frames: <http://ixia.jkqxz.net/~mrt/big_mem_test.patch>. [11:53:41 CET] <jkqxz> It fails in the software decoder with threads == 8 because you allocate more than 100 frame buffers for the stream. [11:53:47 CET] <rcombs> so I've been thinking about format negotiation with hardware formats [11:54:28 CET] <rcombs> swscale can't handle hardware surfaces, and the hardware lib tends to support a specific subset of software formats [11:55:49 CET] <rcombs> so, I'm not sure if they should stay separate filters and have lavfi handle the combinations of both that may sometimes be required, or if the surface converter should be merged into vf_scale and have it handle the combinations internally [11:56:22 CET] <rcombs> (also, there's the question of picking an intermediate format in some cases, but I _think_ there's already code to do something like that elsewhere) [11:56:22 CET] <jkqxz> Since you don't ever actually touch the frames lazy allocation hides the problem in real memory use, but that's still bad. [12:01:25 CET] <jkqxz> I'm somewhat dubious about surface-frame autoconversion. While it makes filtering setup a bit easier to use, any minor misconfiguration might slow everything down immensely without it being clear why. [12:02:55 CET] <rcombs> that's true with autoinserted swscale as-is (though not to quite as large a degree) depending on your filters [12:03:40 CET] <jkqxz> (Though I do admit that autoinserted swscale which barfs on all your hardware surfaces is a right pain to tame in the current setup.) [12:10:19 CET] <wm4> jkqxz: in what case does it "accidentally" insert swscale? [12:13:56 CET] <rcombs> accidentally? [12:15:22 CET] <jkqxz> Output of hardware surfaces from a hwaccel decoder initially sets up the graph for the software output format. When you then output a hardware format, the reconfigure throws in a swscale instance which then barfs on you. [12:15:59 CET] <jkqxz> (See discussion and hack at the bottom of <http://ffmpeg.org/pipermail/ffmpeg-devel/2016-January/188356.html>.) [12:16:01 CET] <rcombs> it throws in a swscale all the time, doesn't it? [12:16:18 CET] <durandal11707> kierank: have you registered FFmpeg to outreachy? [12:16:29 CET] <kierank> durandal11707: I sent email [12:16:31 CET] <wm4> well that's stupid [12:16:38 CET] <wm4> maybe ffmpeg.c shouldn't do this then [12:16:58 CET] <rcombs> it's usually harmless, since vf_scale doesn't actually call swscale if params are all identical on input and output [12:17:42 CET] <wm4> it's all because libavfilter needs the params up-front and can't handle format changes [12:17:51 CET] <wm4> in mpv I have to go out of my way to emulate support for it [12:18:48 CET] <durandal11707> cant it check for hwaccel pix formats [12:19:12 CET] <rcombs> it's meant to handle `-s` and such [12:20:20 CET] <jkqxz> The other problem with autoconversion is really the in making negotiation harder. For "-vf vaapi_scale=... -c:v vaapi_h264 ...", you need a format filter to force it into to do the right thing, and autoconversion would open up much more scope for it to mess up. I'm not sure how to solve this one. [12:21:20 CET] <durandal11707> hwaccel filters use hwaccel formats [12:21:21 CET] <jkqxz> (This is where I had the "force_vaapi_out" parameter in a previous version before I realised that the format filter does exactly what was required there.) [12:22:04 CET] <rcombs> the format filter shouldn't be necessary there either if the swscale filter isn't present, right? [12:22:14 CET] <durandal11707> does vaapi_scale takes normal pixel format as input? [12:23:03 CET] <rcombs> though I think it'd be most convenient to have e.g. `-s <size>` work when doing hardware transcodes (i.e. have it do a vaapi scale) [12:23:12 CET] <jkqxz> vaapi_scale can take some (hardware-dependent, detected at runtime) set of normal pixel formats as input and output. [12:24:21 CET] <jkqxz> vaapi_h264 allows nv12 input for convenience, which then unhelpfully gets picked when there are better choices. [12:25:18 CET] <rcombs> is it listed first? [12:25:57 CET] <jkqxz> AV_PIX_FMT_VAAPI is always first in these lists. [12:26:00 CET] <rcombs> oh, heh, avcodec_default_get_format [12:26:08 CET] <rcombs> `while (*fmt != AV_PIX_FMT_NONE && is_hwaccel_pix_fmt(*fmt))` [13:57:58 CET] <jkqxz> nevcairiel: ^ Same problem, hwaccel is not relevant at all. (Though it does make it fail harder, because virtual memory doesn't save you. Maybe threading+hwaccel shouldn't be allowed... :) [13:59:11 CET] <wm4> lol [14:01:01 CET] <nevcairiel> threading+hwaccel isnt allowed =p [14:02:34 CET] <nevcairiel> luckily allowing it again isnt actually a problem for the majority of users as they dont use it anyway, and those that do can suffer their own fate [14:03:07 CET] <nevcairiel> trying to stop users from shooting themself in the foot is apparently not a good thing anymore [14:15:25 CET] <kierank> can we have some more outreachy tasks please [14:15:29 CET] <kierank> https://trac.ffmpeg.org/wiki/SponsoringPrograms/Outreachy/2016-05 [14:25:14 CET] <wm4> nevcairiel: it's all about vlc [14:25:24 CET] <wm4> and the inability of debian maintainers to troll vlc properly [14:25:44 CET] <wm4> vlc is going to have to solve this, there's no other way, or they'll have to live without vp9 hwaccel forever [14:25:59 CET] <wm4> apropos, why not push your dxva main10 patch? [14:26:40 CET] <Daemon404> wm4, packagers truly believe that you should fix everything at the lowest level possible [14:26:44 CET] <Daemon404> because it's the least work for them [14:26:49 CET] <Daemon404> and fixes the greatest # of packages [14:27:18 CET] <wm4> they also think unmaintained software is worth keeping around [14:27:43 CET] <wm4> so they don't really have much credibility [14:29:38 CET] <jkqxz> How about recommending that they solve it at a lower level: the kernel should serialise all threading. I bet that would solve loads of odd race conditions in other programs too! [14:54:07 CET] <nevcairiel> hwaccels+mt dont even multi-thread really [14:54:16 CET] <nevcairiel> it just switches which worker thread to serially run on =p [14:54:21 CET] <nevcairiel> so still consuming all the resources [15:00:10 CET] <cone-229> ffmpeg 03Michael Niedermayer 07master:43a69655699b: avcodec/dirac: Fix memleak of dsh on error [15:00:24 CET] <kierank> I have no response from outreachy... [15:04:58 CET] <BtbN> jkqxz, does your vaapi patch series come with ffmpeg_vaapi.c, so the ffmpeg cli tool can use vaapi acceleration? [15:07:10 CET] <jkqxz> BtbN: Yes. [15:14:11 CET] <durandal11707> ok to push metadata filters? [15:15:49 CET] <durandal11707> kierank: perhaps last day to register was 8th [15:15:55 CET] <kierank> I sent on 8th [15:35:48 CET] <Timothy_Gu> kierank: for your fuzzing testsuite project, how do you determine which samples are "likely to crash"? [15:36:44 CET] <kierank> Well my point is that for h264 commits you fuzz h264 and not opus [15:37:02 CET] <Timothy_Gu> oh [15:37:10 CET] <kierank> At the moment the current strategy doesn't focus on making sure stuff stays fixed [15:37:27 CET] <Timothy_Gu> ok [15:37:55 CET] <Timothy_Gu> anybody wants to mentor the documentation project? [15:41:04 CET] <Timothy_Gu> also anyone against making fatebeta default? [15:53:24 CET] <BBB> nope [17:48:25 CET] <durandal11707> anybody against (a)loop filters? [17:48:49 CET] <nevcairiel> you act as if we know what those are supposed to do [17:49:57 CET] <nevcairiel> looping in filters is going to be extremely limited however, since you cant issue seek commands, so you have to cache all the looping area, consuming memory to no end =p [17:50:41 CET] <durandal11707> it would cache limited number of frames and at eof loop them [18:02:08 CET] <BBB> kierank: re 5224, Im not sure that can be fixed [18:26:37 CET] <Timothy_Gu> i'm having a hard time trying to simd-ify "2 * a * b / 255," where a and b are bytes unpacked into words [18:27:20 CET] <Timothy_Gu> i've already done a * b / 255 and that works fine [18:27:45 CET] <Timothy_Gu> but I can't pre-multiply a or b by 2 since that would overflow [18:28:11 CET] <Timothy_Gu> i've also tried to multiply the result by 2, but it seems to lose a bit of precision [18:30:38 CET] <jkqxz> BBB: 5224 fell out of hwaccel/vaapi multithread testing. When you need to use preallocated hardware surfaces, the limit really does kill you (mr3_tandberg_b in FATE does the same thing in a less-extreme way). [18:30:39 CET] <durandal11707> yea you need ints [18:30:46 CET] <jkqxz> In the software-only case it doesn't directly matter, but it's possible that a degenerate stream could do nasty things to memory-limited contexts. (Maybe you can make something which falls over on a 32-bit player iff it uses threads?) [18:31:23 CET] <nevcairiel> jkqxz: as long as it just barfs gracefully, its probably not something to worry about if its only on crafted streams [18:32:56 CET] <Timothy_Gu> durandal11707: ugh [18:32:58 CET] <nevcairiel> any crashes on out of memory would of course be bad [18:35:45 CET] <Timothy_Gu> is there a "move quarter" macro in x86inc? [18:36:10 CET] <nevcairiel> move quarter? [18:36:30 CET] <jkqxz> Well, even if it does barf gracefully, calling malloc() too much could exhaust something else which barfs nongracefully. So if no change is to be made it should probably be documented so that memory-limited people know that they should be careful with threads and untrusted video. [18:41:05 CET] <jamrial> Timothy_Gu: pmaddwd? [18:42:54 CET] <nevcairiel> i think he wanted to avoid 32-bit intermediates [19:10:16 CET] <Timothy_Gu> yeah, I don't want to unpack bytes to doublewords [19:10:51 CET] <Timothy_Gu> is it okay if i commit a non-bitexact version (with proper checks)? [19:11:23 CET] <Timothy_Gu> wait there isn't a bitexact flag in libavfilter [19:11:58 CET] <durandal11707> why you don't want doublewords? [19:13:44 CET] <Timothy_Gu> it's too big :p [19:14:06 CET] <Timothy_Gu> plus i need to find a way to move a quarter of an xmm to the memory [19:14:50 CET] <Timothy_Gu> or maybe not. i can probably unpack it to another reg [19:33:23 CET] <BBB> jkqxz: I think we all agree that you shouldnt allocate more than 1-thread worth of surfaces for MT/hw, since you never have actual concurrency with hw enabled [19:38:13 CET] <jkqxz> Timothy_Gu: Something like a * b, then the high half of multipling with 514 (pmulhw)? That's probably not quite right across the whole range, but you have a few more bits above to play with to multiple by something a bit bigger and shift. [19:51:33 CET] <jamrial> Timothy_Gu: a single dword? movd [19:52:31 CET] <jamrial> if it's not the first dword, use psrldq to shift it [20:19:56 CET] <kierank> 7:14 PM <kierank> Did my email about FFmpeg actually make it to the list? [20:19:56 CET] <kierank> 7:19 PM <marina> kierank: it did! thank you for arranging for FFmpeg's participation! [20:20:45 CET] <Timothy_Gu> jamrial: didn't know movd works on xmm too [20:22:00 CET] <jamrial> Timothy_Gu: http://www.felixcloutier.com/x86/MOVD:MOVQ.html [20:25:19 CET] <jamrial> you also have pextrd to avoid having to shift dwords, but that's sse4 only [21:08:24 CET] <durandal11707> kierank: what happened to adding SIMD to overlay? [21:09:00 CET] <kierank> Might be a priority soon [21:35:35 CET] <jkqxz> The first constant that will actually do it exactly is 131587 (a * b * 131587 >> 24 == 2 * a * b / 255), which is unfortunately outside 16 bits :( [21:42:29 CET] <cone-229> ffmpeg 03Michael Niedermayer 07master:331a33d74a0e: nut: Add PAL8 support [21:47:11 CET] <Timothy_Gu> jkqxz: oh well. i'll do 32 bits [21:54:18 CET] <Timothy_Gu> fyi the patches arejust sent are different from the one i was trying to do [00:00:00 CET] --- Wed Feb 10 2016

1 0

ffmpeg.log.20160209
by burek 10 Feb '16

10 Feb '16

[00:41:43 CET] <derekprestegard> how do I speed up ( like 24 to 25 fps) without introducing frame duplication? [00:41:58 CET] <derekprestegard> just setting -r 25 introduces one dupe per second, which is awful [00:42:07 CET] <c_14> setpts [00:42:13 CET] <derekprestegard> ok [00:42:21 CET] <derekprestegard> saw some guides on that, was wondering if there was any other way [00:42:22 CET] <derekprestegard> thx [00:46:20 CET] <DHE> can't you set -r on the input to reprogram the input framerate? [00:54:03 CET] <derekprestegard> DHE: thats what I thought, but Im seeing duplicated frames this way - strangly enough [00:54:14 CET] <derekprestegard> still seeing them with setpts.. not sure whats happening [01:30:56 CET] <votz> Just to confirm my understanding of the source, ffmpeg seeks (i.e. with -ss <foo>) to dts timestamps, not pts timestamps, correct? [02:00:16 CET] <kepstin> votz: seeks are to pts [02:00:48 CET] <votz> kepstin: Great. Thank you. [02:01:03 CET] <kepstin> dts doesn't really make sense at all for seeking due to frame reordering... [02:01:46 CET] <votz> On further inspection, I realize my confusion: the dts and pts values of the keyframes used to seek to are identical. [02:01:52 CET] <kepstin> that's also the reason why seeking is normally to keyframes, since they have dts=pts [02:02:49 CET] <kepstin> you seek to the keyframe with pts less than or equal to desired seek point. Then (in accurate seek mode) decode to the desired exact point, throwing out the extra frames. [02:03:18 CET] <votz> Right. [02:03:59 CET] <votz> kepstin: Thank you for your help and input. [04:44:59 CET] <C0nundrum> WHat's the best way to capture an m3u8 stream ? I'm recording to a browser playable format what i use is ffmpeg -i "http://sourcemedia.com:1935/live-origin/67b88_aac/playlist.m3u8" -vcodec copy -acodec copy -bsf:a aac_adtstoasc -f mp4 "E:\capture 2-8-2016 3.25 PM.mp4" the problem that i have is 1.ffmpeg doesn't seem to try to reconnect and i end up with alot of 30sec-1min recordings 2. [04:45:00 CET] <C0nundrum> if it is iterrupted, the whole file is useless. 3.when playing in browser there are parts where the time keep moving but the image is still. I tried recording using flv to avoid the broken vidoes if interupted but i end up with video and no sound. [04:48:30 CET] <explodes> My AvFrame->pts is always something ridiculous like -9223372036854775808 and it doesn't change, I'm wondering if I need to use pkt_pts to get what I need [04:50:26 CET] <explodes> When I decode a frame, I may or may not seek. When I do, I want to use AVSEEK_FLAG_BACKWARD when I am seeking to a previous position, and 0 (flags) when seeking forward [04:52:44 CET] <explodes> ..let me see if pkt_pts is what I *really* want [06:21:53 CET] <explodes_> Ok.. never mind that. -- av_seek_frame is seeking but it is often off by a couple seconds: http://pastebin.com/pvJUd0Fw [06:23:00 CET] <explodes_> Is my av_rescale_q code just bad? [06:24:58 CET] <pzich> explodes_: no idea what it's like from the lib side, but I know old ffmpeg CLI used to have a fast and slow seek: https://trac.ffmpeg.org/wiki/Seeking [06:25:16 CET] <pzich> being off by a couple seconds might mean it's seeking to the nearest keyframe or something [06:25:40 CET] <explodes_> Yea, that is what the flags I pass in are specifying, but are keyframes really up to 4 seconds apart? [06:25:48 CET] <explodes_> It makes for kind of a crappy user experience [06:25:54 CET] <pzich> depends on how the file is encoded [06:26:15 CET] <pzich> if you wanted to you could put in one, but it sure would be a crappy user experience [06:27:50 CET] <explodes_> ios seeks just fine, android is seeking oddly- so it's probably my JNI code (which I pasted) [06:28:17 CET] <pzich> welp, hopefully someone here understands that and can see what might be wrong [06:29:34 CET] <explodes_> :P I agree with that [06:44:10 CET] <Bray90820> OFF TOPIC: If anyone knows the max resolution of S-Video [06:47:52 CET] <pzich> "Also, S-Video suffers from low color resolution. NTSC S-Video color resolution is typically 120 lines horizontal (approximately 160 pixels edge-to-edge)" https://en.wikipedia.org/wiki/S-Video#Signal [06:47:58 CET] <pzich> says citation needed, though [06:48:56 CET] <pzich> I'm seeing 1024x768 http://www.highdefforum.com/flat-panel-tvs/36931-s-video-cable-what-maximum… and http://arstechnica.com/civis/viewtopic.php?t=332338 [06:48:59 CET] <explodes_> I guess I can AVSEEK_FLAG_BACKWARD and decode frames until I'm "close enough" [06:49:22 CET] <pzich> explodes_: if you can, you'd probably want to step back to the previous keyframe and then slow seek forward [07:16:22 CET] <TD-Linux> Bray90820, S-Video is basically the same as NTSC, but with the chroma channel on a separate wire to prevent crosstalk [07:17:44 CET] <Bray90820> TD-Linux: Bot what would be that resolution [07:18:13 CET] <TD-Linux> so 480i just like NTSC, but with higher both chroma and luma vertical lines of resolution [07:18:31 CET] <Bray90820> Alright [07:18:42 CET] <TD-Linux> err horizontal [07:18:55 CET] <Bray90820> That's higher quality then Composite right? [07:18:59 CET] <TD-Linux> yes it is [07:19:14 CET] <Bray90820> Ok [07:19:17 CET] <Bray90820> Thanks [07:39:49 CET] <explodes_> This code uses av_read_frame to read frames until packet-> pts is within a threshold, it seems to be ok except for the fact that video frames are scrambled until the next keyframe is decoded normally: http://pastebin.com/2kLv7yU3 [07:42:20 CET] <explodes_> fixed memory leak: http://pastebin.com/hsjfZU5P [07:44:19 CET] <explodes_> Any way to NOT get scrambled frames here? [07:45:06 CET] <Valex> Do any of you guys know if stereoscopic 3D is still being supported? [07:48:45 CET] <Valex> Anyone? [07:49:29 CET] <pzich> ...in video containers in general? [07:49:54 CET] <Valex> Yes. [07:50:26 CET] <pzich> I haven't heard anything about them taking it away. I'd assume it'll stick around as an option at the very least. [07:50:52 CET] <Valex> But is it still being developed? [07:51:12 CET] <Valex> Or has the feature been left to rot? [07:52:57 CET] <pzich> is there a feature you need added? my understanding is it basically takes two video files and says "hey, here's a left channel and a right channel of a stereoscopic video: go" [07:55:06 CET] <Valex> That's basically what stereoscopic 3D is. But that type of method generally works on a projector or a monitor that renders the same image twice. What I'm wondering is if there's a working stable software solution that doesn't require the use of PowerDVD. [07:56:15 CET] <Valex> In other words, I'm looking for something more than a quick and dirty hack. [07:56:19 CET] <pzich> I don't really...I mean, it sorta all depends on what you're playing it back on and what files it understands. [07:57:47 CET] <pzich> I think the stereoscopic 3D implementation in the MKV container format is more than a "quick an dirty hack", if whatever is playing it understands that format. [07:59:40 CET] <Valex> But what about media that is already in 3D? For example, if I were to play Tron:Legacy in 3D, wouldn't MKV effectivly render the frames 4 times simultaneously? [08:01:06 CET] <pzich> 4 times? [08:02:34 CET] <Valex> Twice for the 3D encoding in the disc and twice for 3D in MKV. [08:03:22 CET] <Valex> Unless, I'm wrong about how encoding works on 3D Disc. [08:03:50 CET] <pzich> I have no idea how it works, it probably depends on the disc, are you talking about standard blu-ray or something else? [08:04:09 CET] <Valex> 3D Blu-ray. [08:04:12 CET] <pzich> most of them have some flavor of container format and a variety of video/audio/subtitle streams [08:05:04 CET] <pzich> Valex: apparently it might be using this? http://www.tomshardware.com/reviews/blu-ray-3d-3d-vision-3d-home-theater,26… [08:05:52 CET] <pzich> I guess that way instead of having two full streams (and therefore twice the bandwidth and disc space) you have a base and dependant view [08:06:48 CET] <pzich> if I had to guess, 3D blu-ray format is about as standard as you can get [08:06:52 CET] <pzich> maybe check out http://www.sonycreativesoftware.com/dostudio/understanding_3d_authoring_blu… ? [08:12:47 CET] <Valex> I guess that makes sense. But authoring is done with having a enterprise-backed standard in mind usually involving proprietary hardware and software (i.e. not open-source solutions). Generally, what I was trying to look for was some library or application that can read and render 3D Blu-ray content. [08:12:58 CET] <Valex> on a PC. [08:20:57 CET] <pzich> Valex: I'm sure there are ones, I've just never used one. [08:21:07 CET] <pzich> not sure if ffmpeg can be a part of that [09:15:24 CET] <dorp> When it comes to transcoding duration time, x264 with preset=ultrafast and qp=0, is supposed to be much more efficient than x265? [10:56:53 CET] <Bluez_> hi guys [11:00:28 CET] <Bluez_> if im muxing raw h264 packets, do i still need to set gop_size and bit rate in the AVCodecContext ? [11:19:05 CET] <database1982> hi there [11:28:53 CET] <waressearcher2> database1982: hallo, wie geht's es dir ? [11:34:27 CET] <database1982> hi waressearcher2 [11:34:51 CET] <database1982> you speak english ? [11:35:45 CET] <waressearcher2> manchmal [11:37:56 CET] <database1982> waressearcher2, es geht mir gut (google translated) ^^ [11:39:07 CET] <database1982> Ich erlebe ein Problem mit mp4 Decodierung mit libav [11:40:53 CET] <database1982> av_read_frame returns AVERROR_EOF when it shouldn't on some local files [11:50:07 CET] <andrey_utkin> database1982: ensure to use latest ffmpeg first, if you run into weird behaviour [11:56:45 CET] <database1982> andrey_utkin: I am using 2.8.6 [11:57:28 CET] <database1982> andrey_utkin, I will check ffplay if the same behavior happens [11:58:16 CET] <database1982> andrey_utkin, maybe an option i am missing to set on format or codec [12:10:04 CET] <IamTrying> http://i.imgur.com/TCR27zN.png - i am trying this FFmpeg ( https://trac.ffmpeg.org/wiki/DirectShow ) for BUG report in Windows. But when i type ffmpeg -f dshow -list_options true -i video="Integrated Camera" its failing what is i am missing here please? [14:11:18 CET] <lee> hi. I am trying to stitch together a sequence of PNGs into a VP8/webm video, with transparency, but the resulting video has a black background. I'm running "ffmpeg -i zt%04d.png -c:v libvpx output.webm", full output at http://pastebin.com/8vFtwSpZ . am I missing something obvious? [14:12:17 CET] <lee> the source PNGs are available at https://www.dropbox.com/s/zmtn759rq7dmxjd/explode-animation.7z?dl=1 [14:14:46 CET] <furq> what are you using to play the video [14:15:03 CET] <furq> some decoders don't support vp8 alpha [14:15:17 CET] <lee> chrome [14:15:26 CET] <lee> but testing in VLC [14:15:35 CET] <furq> does it work in chrome [14:15:39 CET] <lee> no [14:18:36 CET] <furq> are you playing it back directly in chrome or in a video tag with background: transparent; [14:19:05 CET] <furq> the former gives a black background for me [14:19:15 CET] <lee> I am playing it back in a video tag [14:21:59 CET] <___dewey> Hi all, is there anyone with experience with AVSubtitle::end_display_time [14:22:25 CET] <___dewey> its always the same value in my case, that shouldnt be the case I think [14:23:17 CET] <___dewey> printf("End display time: %i\n", sub.end_display_time) = End display time: 10000 [15:58:35 CET] <Bluez_> is there any helper code to decode SPS in libav? [16:45:05 CET] <cortexman> i'm trying to get my video looking good on youtube but they keep compressing it to utter crap [16:46:04 CET] <cortexman> ffmpeg -framerate 60 -i %d.jpg -c:v libx264 -tune stillimage -preset ultrafast vid.mp4 [16:46:08 CET] <cortexman> i'm thinking about adding -crf 0 [16:56:21 CET] <cortexman> now i am doing ffmpeg -framerate 60 -i %d.jpg -c:v libx264 -tune stillimage -preset -qp 0 -crf 0ultrafast vid.mp4 [16:56:25 CET] <cortexman> now i am doing ffmpeg -framerate 60 -i %d.jpg -c:v libx264 -tune stillimage -preset -qp 0 -crf 0 ultrafast vid.mp4 [16:56:36 CET] <cortexman> it looks like the video is going to be larger than the input images........ [16:56:49 CET] <c_14> don't set both -qp and -crf [16:56:56 CET] <c_14> And it will (unless the source is lossless) [16:57:18 CET] <cortexman> what would lead the video the be larger than then inputs [16:57:22 CET] <cortexman> i need to restrict this thing to 128GB [16:57:50 CET] <cortexman> right now it's encoding at about 3gbps [16:58:01 CET] <cortexman> (bitrate) [16:59:16 CET] <c_14> It's lossless [16:59:29 CET] <c_14> lossless encodings are larger than lossy encodings [16:59:29 CET] <cortexman> the raw images are about 128GB [16:59:37 CET] <cortexman> it looks like the video will be larger than that [17:00:02 CET] <cortexman> i could be wrong, it hasn't finished [17:14:13 CET] <LeonG> YouTube's VP9 60fps encodings look less blocky than the H.264 encodings. Check if your browser is playing back the VP9 version [17:18:02 CET] <cortexman> how can you tell [17:18:19 CET] <furq> right click > stats for nerds [17:19:47 CET] <cortexman> Mime Type: video/mp4; codecs="avc1.640020" [17:20:23 CET] <cortexman> {"ns":"yt","el":"detailpage","cpn":"rIdN_vTjaRP6qSjV","docid":"lLivPYzvRPY","ver":2,"referrer":"https://www.youtube.com/my_videos?o=U","cmt":"24.055","plid":"AAUrWKyLS__zsuFW","ei":"bxG6VvaiDYifuALIuLTYBQ","fmt":"299","fs":"0","rt":"63.464","of":"M7zcFnZlImSWMAymFtyqrw","adformat":null,"content_v":null,"euri":"","subscribed":null,"lact":2,"live":null,"cl":"114145475","mos":0,"osid":"MTkxMTQ3ZGI:AOeUNAbul4CX764JMGzEwFv3Uw7OpSpV7g [17:20:23 CET] <cortexman> ","state":"4","vm":"CAEQAA","volume":100,"c":"WEB","cver":"html5","cplayer":"UNIPLAYER","cbr":"Chrome","cbrver":"48.0.2564.103","cos":"X11","hl":"en_US","cr":"US","len":"8114.678","fexp":"3300132,3310848,3312305,3312381,9407610,9416126,9417058,9420452,9422542,9422596,9423661,9423662,9425838,9425967,9426055,9426729,9427037,9427858,9428293","afmt":"140","vct":"24.055","vd":"8114.678","vpl":"20.200-24.055,","vbu":"20.016-253.237,", [17:20:24 CET] <cortexman> "vpa":true,"vsk":false,"ven":false,"vpr":1,"vrs":4,"vns":2,"vec":null,"vvol":1,"lct":"24.055","lsk":false,"lmf":true,"lbw":"7570669.268","lhd":"0.073","ltd":"2.433","laa":"itag=140,seg=25,range=3980815-4033604,time=250.0-253.3","lva":"itag=299,seg=49,range=67323708-68693831,time=248.3-253.3","lar":"itag=140,seg=26,range=4139750-4298317,time=260.0-270.0","lvr":"itag=299,seg=51,range=70069422-71443018,time=258.4-263.5","lvh":"r19- [17:20:29 CET] <cortexman> --sn-qxo7sn7s","lab":"20.016-253.237,","lvb":"15.200-253.333,","ismb":18940000,"debug_videoId":"lLivPYzvRPY","gpu":"GeForce_GTX_960/PCIe/SSE2","cgr":true,"debug_playbackQuality":"hd720","debug_date":"Tue Feb 09 2016 09:19:59 GMT-0700 (MST)"} [17:20:33 CET] <cortexman> oops [17:21:11 CET] <cortexman> at any rate, this video looks like complete crap as compared to what i uploaded [17:21:27 CET] <waressearcher2> nicht schlecht [17:21:36 CET] <furq> there's not much you can do about it [17:21:42 CET] <furq> other than use a different site [17:21:59 CET] <cortexman> i am uploading a lossless version to see how much they f it up [17:22:29 CET] <furq> does youtube not support mjpeg uploads [17:23:55 CET] <cortexman> seems like they do [17:24:23 CET] <drv> it's a video hosting site, not a "sequence of unrelated images" hosting site [17:24:24 CET] <cortexman> i think i would need to stick it in an avi container [17:24:30 CET] <cortexman> the images aren't unrelated. [17:24:31 CET] <cortexman> go away [17:24:51 CET] <cortexman> -vcodec mjpeg out.avi [17:24:52 CET] <drv> unrelated in terms of pixels, not thematically or whatever [17:25:32 CET] <drv> it's no surprise that it can't encode a bunch of different still images at the bitrate intended for normal videos that have highly correlated frames [17:26:42 CET] <furq> afaik you can just mux a jpeg sequence into mkv if you want a lossless video [17:26:57 CET] <furq> but yeah it's not going to look noticeably better [17:27:08 CET] <cortexman> the video has major compression artifacts [17:27:17 CET] <furq> yes it does [17:28:34 CET] <furq> nothing you do is going to get around the fact that youtube is encoding it at far too low a bitrate [17:29:03 CET] <cortexman> ah, 5mpbs-ish i think [17:29:19 CET] <cortexman> i could drastically reduce the framerate [17:30:48 CET] <furq> "youtube-dl -F yourvideourl" will show the bitrates they're using [17:31:55 CET] <cortexman> wow, they compressed the hell out of it [17:32:12 CET] <cortexman> that was a 77GB file [17:32:20 CET] <cortexman> no, this one was 36GB [17:32:41 CET] <furq> they transcode everything to roughly the same bitrate [17:34:05 CET] <furq> dropping to 30fps will probably help a bit [17:34:13 CET] <furq> i doubt it'll be good enough though [17:34:31 CET] <cortexman> trying ffmpeg -framerate 30 -i %d.jpg -vcodec mjpeg vid.avi [17:34:40 CET] <furq> don't use -vcodec mjpeg [17:34:50 CET] <furq> use -c copy [17:35:39 CET] <cortexman> nice [17:36:21 CET] <jkqxz> What do you want the visible framerate to be at the end of this process? You could copy all of the input frames ten times - that won't significantly affect the size of your upload but would affect their bitrate calculations. [17:36:39 CET] <cortexman> 15 - 60fps [17:37:42 CET] <jkqxz> Then I think you are probably just doomed because of the bitrate constraints they will apply on their side. But you could still try a 60fps video with each frame in four times. [17:46:00 CET] <cortexman> jkqxz, can you provide some command line arguments? [17:46:27 CET] <cortexman> furq, i am uploading a 127GB copy'd file now, we will see how it looks in 1-2 dayS [17:48:36 CET] <cortexman> jkqxz, you can make the video up to 11 hours. there are about 480k images [17:50:12 CET] <jkqxz> Something like "ffmpeg -framerate 15 -i %d.jpg -vf fps=fps=60 -c:v libx264 ...", I think. (Though do make some of it and single-frame-step through the output to make sure it has acted sensibly with the frames which are the same.) [17:51:37 CET] <cortexman> i have to use H264 instead of copying, so that the file size stays ~ the same? [17:52:35 CET] <jkqxz> Yeah. If you copy then it will have to be four times the size. [17:54:53 CET] <cortexman> takes 13 hours to upload the max file size [17:55:09 CET] <cortexman> on biz class comcast [18:10:06 CET] <cortexman> the 77GB video i uploaded yesterday has only returned a 360p version thus far [18:10:12 CET] <cortexman> i suppose it's still processing... [18:10:16 CET] <cortexman> been 16 hours [18:10:19 CET] <cortexman> since it finished uploading [18:11:39 CET] <jkqxz> Is there some reason to believe that the result will be different to a shorter test with only a few hundred frames? [18:12:43 CET] <cortexman> seems like a good point [18:17:20 CET] <bencoh> how long was your 77G video? [18:32:51 CET] <podman> If you guys had to pick just one format for video delivery over the web using HTML5 to implement and support today, would it be HLS or DASH [18:34:42 CET] <podman> I was going with DASH, but i think iOS support might be a big enough issue to stick with HLS for now [18:34:53 CET] <Mavrik> HLS of course. [18:35:01 CET] <Mavrik> With flashplayer fallback for everything else. [18:36:06 CET] <podman> Mavrik: yeah, sounds like that's probably the way to go. HLS.js for everything that supports MSE, native on iOS, flashls for everything else [18:37:05 CET] <Mavrik> on the other hand [18:37:11 CET] <Mavrik> you san support both with a simple nginx plugin [18:38:25 CET] <podman> Mavrik: not really feasible, right now [18:38:41 CET] <bencoh> you'd still need to remux (on-the-fly or offline) [18:38:46 CET] <Mavrik> *shrug* [18:38:54 CET] <Mavrik> You'll be losing users. [18:39:19 CET] <podman> Our users don't really care about HLS vs Dash [18:39:27 CET] <bencoh> is MSE/HLS.js fully supported on Android? [18:40:55 CET] <podman> bencoh: looks pretty good on Chrome on Marshmallow at least [18:41:37 CET] <cortexman> bencoh, 2 hrs 15 mins [18:41:38 CET] <DHE> chrome for android played back HLS for me without any special software. That includes no javascript (direct m3u8 link). That's on Kitkat [18:47:56 CET] <podman> so what's the best HLS segmenter these days then? [18:54:34 CET] <DHE> umm... ffmpeg does pretty well. I would suggest changing the defaults though [18:55:56 CET] <DHE> eg: hls_time defaults to 2 seconds. I'd pick something bigger. 4 seconds would be my minimum... [19:13:53 CET] <bencoh> cortexman: why do you need 77G to upload a 2h15 video to youtube? [19:14:17 CET] <cortexman> i can give you 500,000 reasons [19:14:37 CET] <bencoh> it's not like their HD format is worth a 5mbps anyway [19:14:54 CET] <cortexman> maybe it is [19:14:55 CET] <furq> i assume he wants the best source possible for youtube to ruin [19:15:11 CET] <furq> it makes sense in theory but i don't think it's going to make any actual difference [19:15:33 CET] <bencoh> I don't either ... sadly [19:16:46 CET] <DHE> could be 4k video, or 60fps... or both? [19:16:59 CET] <furq> no it's 480,000 jpegs [19:17:03 CET] <DHE> (I honestly don't know what youtube does to ruin those, but the point stands) [19:17:26 CET] <furq> which from what i gather aren't related, so not inter-frame compressible [19:17:27 CET] <bencoh> DHE: they just bitstarve-encode it [19:17:35 CET] <bencoh> furq: hmm [19:17:46 CET] <cortexman> they are related, i.e., same model wearing different outfits at the same show across 30 frames [19:17:54 CET] <bencoh> does youtube allow 1080p@60fps ? [19:17:54 CET] <cortexman> same pose [19:18:06 CET] <bencoh> (as an enduser format) [19:18:21 CET] <cortexman> bencoh, i am trying to find that out. yesterday i angered the youtube gods by uploading a very large video, which it has so far only encoded as 360p [19:18:33 CET] <bencoh> :] [19:18:44 CET] <furq> i think it supports 2160p60 [19:18:52 CET] <furq> it'll probably be about 7mbps though [19:18:58 CET] <bencoh> ha.ha. :] [19:19:00 CET] <DHE> that's 4k... and yeah, higher bitrates will become available over time.. [19:19:08 CET] <DHE> also last I heard I thought youtube only allowed 64 GB as the max file size [19:19:16 CET] <cortexman> the uploading guide actually says higher framerates are acceptable [19:19:20 CET] <cortexman> ..higher than 60 [19:19:30 CET] <cortexman> and youtube has an example 8k video up [19:19:34 CET] <bencoh> acceptable as input, but what about output? [19:19:41 CET] <cortexman> not yet clear. [19:19:54 CET] <furq> https://www.youtube.com/watch?v=_fC53l3jGAc [19:19:58 CET] <furq> that's available in 2160p60 [19:20:05 CET] <furq> amazingly it seems to be about 30mbit [19:20:09 CET] <cortexman> 8k video: https://www.youtube.com/watch?v=sLprVF6d7Ug [19:20:09 CET] <bencoh> because seriously, considering their 1080p result, I wouldn't expect much of higher res/rate [19:20:20 CET] <furq> http://sprunge.us/VZCC [19:20:22 CET] <cortexman> i have a 4Ghz CPU and it can't play that 8k video [19:20:28 CET] <cortexman> also a nice video card [19:20:37 CET] <furq> 315 webm 3840x2160 DASH video 38636k , vp9, 60fps, video only, 1.24GiB [19:20:48 CET] <bencoh> hmm [19:21:09 CET] <furq> the x264 bitrates ;_; [19:21:21 CET] <furq> 10mbps for 1440p h264 or 20mbps for 1440p vp9 [19:21:29 CET] <furq> why do i feel like that's intentional [19:21:36 CET] <bencoh> :D [19:22:34 CET] <cortexman> i should try uploading in exactly the same format they encode to... [19:22:41 CET] <furq> that makes no difference [19:23:14 CET] <cortexman> webm 7680x4320 DASH video 59785k , 24fps, video only, 397.69MiB [19:23:15 CET] <bencoh> iirc dailymotion has (had?) a "original stream" player option [19:23:18 CET] <cortexman> is that 60 mpbs? [19:23:19 CET] <bencoh> at least for live streams [19:23:22 CET] <furq> yes it is [19:23:38 CET] <furq> youtube used to have an "original video" option but it's been gone since they introduced DASH [19:23:41 CET] <furq> which makes sense i guess [19:23:46 CET] <bencoh> indeed [19:23:51 CET] <cortexman> how do i get that 60mpbs rate [19:23:57 CET] <cortexman> specify 8K resolution? [19:24:02 CET] <furq> i would have thought so [19:24:09 CET] <bencoh> they probably dont want to mess with muxing variable GOP [19:24:11 CET] <furq> that 4K video i linked gets 35mbps [19:24:29 CET] <furq> 4K doesn't even show up in the options for me though [19:24:45 CET] <furq> i only get 1080p60, which is going to look like garbage [19:24:54 CET] <cortexman> i have the 4K option [19:25:13 CET] <furq> i wonder if it's detecting my screen size [19:25:25 CET] <cortexman> and 8K [19:25:28 CET] <cortexman> it does detect your video card [19:25:34 CET] <cortexman> i only have a 4K monitor [19:25:44 CET] <furq> this is only 1920*1200 [19:26:11 CET] <furq> maybe it just checks for >1920 width [19:26:19 CET] <cortexman> my computer can't play the 4K video [19:26:49 CET] <furq> but yeah even if you manage to get 35mbit at 2160p, most people aren't even going to see the option [19:26:58 CET] <furq> they'll be stuck on the 5mbit 1080p stream [20:39:44 CET] <t4nk328> hello [20:39:51 CET] <t4nk328> is anyone here to help [20:40:09 CET] <t4nk328> i need help to install ffmpeg [20:40:12 CET] <DeHackEd> that's not how it works. post your question, then someone will try to help. (patience required) [20:41:01 CET] <t4nk328> i am trying to installl fffmpeg [20:41:09 CET] <t4nk328> anyone able to help me [20:41:44 CET] <DeHackEd> well what problem are you having? [20:42:14 CET] <t4nk328> i am trying to install ffmpeg my linux hosting [20:42:22 CET] <t4nk328> anyone know how to do it [20:42:29 CET] <t4nk328> i login in ssh putty [20:44:18 CET] <J_Darnley> ./configure && make [20:44:19 CET] <J_Darnley> done [20:44:48 CET] <DeHackEd> depending on your distribution it might be as simple as "apt-get install ffmpeg" or some such [20:45:10 CET] <t4nk328> apt is it for linux or ubuntu? [20:45:20 CET] <DeHackEd> ubuntu, debian, and derivatives [20:45:42 CET] <t4nk328> i have linux server [20:46:00 CET] <t4nk328> i am not that experience [20:46:09 CET] <t4nk328> with this ffmpeg [20:46:20 CET] <DeHackEd> well, try it. worst case it just doesn't work [20:47:25 CET] <Mavrik> -_- [20:48:54 CET] <t4nk328> i have ffmpeg downloaded my host [20:49:01 CET] <t4nk328> but don't know how its work [20:49:31 CET] <J_Darnley> What *exactly* are you trying to do? Compile ffmpeg or run ffmpeg? [20:49:59 CET] <t4nk328> i have video script i want to make work ffmpeg [20:50:05 CET] <t4nk328> converting video [20:50:25 CET] <t4nk328> i type ffmpeg sais -bash: ffmpeg: command not found [20:56:10 CET] <J_Darnley> If you really have no idea of anything then grab a static linux build from here http://ffmpeg.org/download.html and run that. [20:57:25 CET] <t4nk328> how do i download that linux [20:57:30 CET] <t4nk328> what command needed? [21:00:10 CET] <J_Darnley> wget [21:03:34 CET] <Betablocker> + download URL [21:03:57 CET] <t4nk328> after download [21:04:06 CET] <t4nk328> i did download ssh puty [21:04:32 CET] <t4nk328> tar. gzip [21:04:34 CET] <t4nk328> i download [21:04:37 CET] <t4nk328> untar? [21:04:41 CET] <t4nk328> unzip? [21:04:49 CET] <t4nk328> what next command i have to use? [21:05:15 CET] <J_Darnley> tar [21:05:39 CET] <Betablocker> + downloaded filename [21:05:58 CET] <Betablocker> :) [21:06:19 CET] <DeHackEd> this is "Linux 102" material I think... [21:06:29 CET] <Betablocker> + 10 bucks donation to ffmpeg project [21:08:09 CET] <t4nk328> if i can install and work fine 10 dontaion [21:08:14 CET] <t4nk328> otherwise no [21:08:20 CET] <Betablocker> :) [21:08:24 CET] <Betablocker> just kidding [21:08:41 CET] <t4nk328> i have video script trying to get usr/bind/ffmpeg work [21:09:04 CET] <t4nk328> .usr/bin/ffmpeg [21:09:10 CET] <Betablocker> best thing about ffmpeg it is very good documented [21:09:23 CET] <J_Darnley> pfft. no it isn't [21:09:54 CET] <Betablocker> you want to compile ffmpeg or what are you working on ? [21:10:08 CET] <durandal11707> J_Darnley: what's not documented? [21:10:09 CET] <DeHackEd> it does rather assume you know about video encoding to begin with, like what a b-frame is. but assuming you do, the documentation is fine [21:11:01 CET] <J_Darnley> t4nk328 did you getanything out of the tar yet? [21:11:52 CET] <J_Darnley> durandal11707: I don't know what's *not* documented [21:12:22 CET] <furq> i don't think you need to know what a b-frame is to copy a command line off stackoverflow [21:13:32 CET] <Betablocker> hihi [21:13:41 CET] <t4nk328> j_darnley its just downloaded [21:14:23 CET] <J_Darnley> then let me give a slightly better hint: tar -xf [21:15:07 CET] <t4nk328> tar.gz i downloaded [21:15:35 CET] <t4nk328> saved 10 mb [21:15:51 CET] <t4nk328> x-gzip [21:29:10 CET] <lee> anybody around who can decode the apparently dark art of creating a VP8 video that preserves the alpha from the source PNGs? [21:31:04 CET] <relaxed> lee: did you try -pix_fmt yuva420p ? [21:31:37 CET] <lee> pretty sure I did, let me try again [21:33:29 CET] <lee> yep, no dice with: ffmpeg -i zt%04d.png -c:v libvpx -pix_fmt yuva420p output.webm [21:33:35 CET] <lee> the background continues to be solid black [21:36:48 CET] <BtbN> what do you expect? To see the desktop behind the player? [21:37:59 CET] <lee> I expect it to be transparent [21:38:23 CET] <lee> and show the background of whatever is on the page where the <video> tag is [21:38:33 CET] <furq> BtbN: http://simpl.info/videoalpha/ [21:38:47 CET] <furq> assuming you're using chrome [21:39:34 CET] <BtbN> I'm not, no transparency there for me. [21:39:55 CET] <kepstin> BtbN: yeah, only chrome/chromium supports it atm [21:39:57 CET] <kepstin> I think you have to explicitly use css to set the video tag background to be transparent, it defaults to black [21:40:35 CET] <lee> nope, that's not needed [21:42:00 CET] <kepstin> lee: can you pastebin your ffmpeg output so we can see what's actually going on? [21:42:40 CET] <lee> http://pastebin.com/8vFtwSpZ [21:43:53 CET] <lee> hrm, can I demonstrate with dropbox... [21:44:13 CET] <kepstin> lee: looks all correct, that should be a transparent video... [21:44:39 CET] <lee> nope, damn, [21:45:15 CET] <lee> kepstin: my video: https://www.dropbox.com/s/fb5eacclure1fq0/output.webm?dl=0 the dancer video from simpl.info/videoalpha: https://www.dropbox.com/s/kdctapgsasma5yl/dancer1.webm?dl=0 [21:46:32 CET] <lee> if you'd like to try it yourself the source PNGs are available at https://www.dropbox.com/s/zmtn759rq7dmxjd/explode-animation.7z?dl=1 [22:09:02 CET] <lee> I'm pretty sure I'm missing something obvious but I'll be buggered if I can find it [22:21:23 CET] <relaxed> lee: try, -pix_fmt yuva420p -metadata:s:v:0 alpha_mode="1" [22:53:16 CET] <lee> relaxed: THANK YOU! [22:53:51 CET] <lee> how did you figure that out? intimate knowledge of ffmpeg, or RTFMing something I couldn't find? [23:03:05 CET] <PiTutorials> Hello, how can I tell if I have the latest version of ffmpeg please. [23:03:27 CET] <J_Darnley> ffmpeg -versionwill tell you things [23:03:40 CET] <J_Darnley> such as the compile date [23:03:56 CET] <J_Darnley> and the git commit hash [23:04:10 CET] <J_Darnley> then you could go see what has happened since then [23:04:20 CET] <c_14> It won't tell you the compilation date anymore, just version/commit hash [23:04:33 CET] <J_Darnley> wtf? [23:04:34 CET] <PiTutorials> ok, so I get 'ffmpeg version N-77414-g484cc66' [23:04:37 CET] <J_Darnley> what did that heppen? [23:04:46 CET] <c_14> A while ago [23:05:18 CET] <PiTutorials> is the g484cc66 the comit hash? [23:05:27 CET] <J_Darnley> without the g, yes [23:05:49 CET] <J_Darnley> it looks atleast 300 commits old(er than mine) [23:05:52 CET] <PiTutorials> ok, thanks [23:06:54 CET] <c_14> J_Darnley: https://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=103e4c58633f210d… [23:07:06 CET] <J_Darnley> gah [23:07:30 CET] <J_Darnley> If you're going to remove something useful then replace it with something else useful. [23:09:13 CET] <jkqxz> git hash is plenty! [23:09:23 CET] <jkqxz> $ git log | grep -A 2 484cc66 | tail -1 [23:09:23 CET] <jkqxz> Date: Sun Dec 20 21:31:55 2015 +0100 [23:09:54 CET] <J_Darnley> Sure, if I have a .git dir around and git installed. [23:10:11 CET] <furq> it's fine i'm pretty sure everyone is an ffmpeg developer [23:10:37 CET] <J_Darnley> And then is abc1234 newer than def5678? [23:10:45 CET] <jkqxz> Well maybe not. I guess that isn't quite right because it tells you the date on the commit, not when it actually went into the repository. [23:12:07 CET] <J_Darnley> I want svn back with a nice linear history and a revision number. [23:12:27 CET] <furq> i'm going to assume that was a joke before someone else has a chance to assume it wasn't [23:12:52 CET] <J_Darnley> okay then [23:12:59 CET] <J_Darnley> I want *** back with a nice linear history and a revision number. [23:13:06 CET] <furq> that's better [23:13:07 CET] <J_Darnley> I censored the swear [23:13:47 CET] <furq> it's ok we all know you meant perforce [23:14:55 CET] <PiTutorials> Ok, that's the date I pulled and compiled, so makes sense. Is there an easy way to update, or do I need to pull and compile again. [23:15:19 CET] <J_Darnley> How can you update without compiling? [23:15:22 CET] <c_14> pulling and compiling is the easy way to update [23:15:34 CET] <J_Darnley> You can download a pre-compiled binary I guess. [23:16:19 CET] <PiTutorials> there is no built in update command then. [23:16:50 CET] <J_Darnley> Only bad programs have that. [23:17:37 CET] <J_Darnley> Such as Windows, every browser [23:17:56 CET] <PiTutorials> thanks, just wanted to check [23:21:21 CET] <relaxed> lee: that worked? I saw the metadata in the webm from the example website. [23:23:56 CET] <Nosomy> How to read xvid stats pass file? [23:24:06 CET] <Nosomy> sorry bad internet [23:50:59 CET] <FilipeMaia> How can I use libx264 from the API without getting console output like [libx264 @ 0x7fe66c000000] using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2 AVX? [23:51:58 CET] <J_Darnley> By changing the loglevel [23:52:21 CET] <J_Darnley> set it to warning [23:53:51 CET] <Nosomy> How to read xvid stats pass file? [23:53:59 CET] <Nosomy> [frametype] [qp] [?] [?] [?] [FrameSize] [?] [23:54:21 CET] <FilipeMaia> J_Darnley: thanks [23:54:54 CET] <xbloodhound> does anyone know if ffmpeg supports amd hardware acceleration by any chance? [23:55:15 CET] <J_Darnley> It supports a few APIs for that [23:55:20 CET] <J_Darnley> DXVA [23:55:23 CET] <J_Darnley> VAAPI [23:55:26 CET] <J_Darnley> VDPAU [23:55:39 CET] <J_Darnley> Something Apple-y [23:56:05 CET] <J_Darnley> Nosomy: read the xvid source [23:56:24 CET] <Nosomy> J_Darnley wheres? [23:56:48 CET] <J_Darnley> You're the one still building libxvid by the sound of it. [23:57:16 CET] <J_Darnley> Or if you mean "where in the source" then I have even less idea. [23:57:44 CET] <Nosomy> hmm libxvid? let me see [23:59:16 CET] <lee> relaxed: ahh, I don't know enough to do that, but thanks :D [00:00:00 CET] --- Wed Feb 10 2016

1 0