[Ffmpeg-devel-irc] ffmpeg-devel.log.20160306

burek burek021 at gmail.com
Mon Mar 7 02:05:03 CET 2016


[00:06:37 CET] <jkqxz> To answer myself, it just means that num_ref_idx_lN_active_minus1 > 0.  Immediately after an IDR frame there is only one reference picture present, so you can't fill the RefPicListNs.  Maybe the message shouldn't be at error level.
[00:08:05 CET] <nevcairiel> streams that rely on the default ref frame behavior are rather odd
[00:08:19 CET] <nevcairiel> most do not in fact output such warnigs
[00:14:03 CET] <jkqxz> You expect that the frame immediately after an IDR frame in such a stream will contain ref_pic_list_modification which fills the other slots?  (With the same frame, presumably, since there's only one.)
[00:14:35 CET] <jkqxz> Or maybe just noone sets it in the PPS.
[00:20:03 CET] <jkqxz> Anyway, it seems the real answer to my problem is that the Intel driver does not support weighted_bipred_idc == 2, so if you set it your B frames will all be subtly wrong (presumably assumes 0).  Thanks for documenting that, Intel.
[01:15:55 CET] <nevcairiel> for encoding?
[01:16:00 CET] <nevcairiel> i never noticed problems decoding
[01:34:19 CET] <jkqxz> Yeah, encoding.  They certainly couldn't omit it for decoding.
[03:32:45 CET] <cone-701> ffmpeg 03Mark Harris 07master:238ddd6482d7: avformat/dump: Fix context/level for payload dump
[06:37:18 CET] <Timothy_Gu> ethe: https://gist.github.com/TimothyGu/23ec15872465cd295444
[08:08:44 CET] <Zeranoe> I built the latest (1.17) Intel MFX after some minor tweaking. FFmpeg compiled fine, but when I try to use the h264_qsv encoder I get "Error retrieving encoding parameters." Digging a little deeper and it's failing at MFXVideoENCODE_GetVideoParam in qsvenc.c line 604. ret is -3 or "MFX_ERR_UNSUPPORTED" which intel documents as "undeveloped feature".
[08:10:22 CET] <Zeranoe> I find it odd that 1.16 from lu-zero's mfx_dispatch
[08:10:22 CET] <Zeranoe>  works fine, but the newer version is "undeveloped". Does anyone have an idea what's going on here? Maybe nevcairiel?
[10:52:12 CET] <cone-161> ffmpeg 03Muhammad Faiz 07master:7dd60a0b54ee: avfilter/avf_showcqt: remove CoeffsType
[11:54:19 CET] <cone-161> ffmpeg 03Reimar Döffinger 07master:0a04c2885f02: aacenc: avoid double in quantize_bands.
[12:21:39 CET] <cone-161> ffmpeg 03Michael Niedermayer 07master:40aeff5f0287: avformat/avienc: Remove unneeded seekable tests
[12:36:25 CET] <cone-161> ffmpeg 03Reimar Döffinger 07master:5f5e6033cd98: bitstream.c: improve init_vlc error messages.
[13:24:47 CET] <cone-161> ffmpeg 03Umair Khan 07master:7d6c264e0227: avcodec/alsdec: treat quant_cof as a signed value
[13:55:30 CET] <cone-161> ffmpeg 03Kieran Kunhya 07master:247fe3e494ca: avcodec/cfhd: Fixes cfhd_odd.mov which has a resolution of 496x241
[13:57:02 CET] <Fyr> alright, let's wait a few years till somebody make it.
[13:58:00 CET] <Fyr> all the video editors can do that, only ffmpeg determines itself as something different.
[14:02:01 CET] <wm4> do what
[14:02:55 CET] <JEEB> Fyr: trying to imply that FFmpeg is a video editor is kind of derp, not to mention that filters like that are most probably better off be done on hardware that actually specializes in graphics. FFmpeg's job is to get video decoded or encoded, and the further you go from that the closer you are to just being better off making something on top of FFmpeg
[14:03:38 CET] <JEEB> wm4: he made that ticket :P
[14:03:53 CET] <wm4> whatever "3D cube effect" is, it sounds useless
[14:05:14 CET] <JEEB> but sure, since it's OSS if someone cares enough about it and produces a patch of good enough quality then libavfilter will have yet another funky filter
[14:05:52 CET] <wm4> it's so special and high effort that I can't imagine that
[14:05:52 CET] <JEEB> which I'm not even sure how you'd use because I'm pretty sure doing filtering on a "stitching point" of two clips is a PITA with libavfilter anyways
[14:06:05 CET] <wm4> even if you get opengl support in lavfi, would you hardcode the cube?
[14:08:58 CET] <JEEB> anyways, just saying that it most probably won't get closed as a ticket, but really it is out of the main scope of FFmpeg (and I do agree there's plenty of such stuff already in libavfilter et al)
[14:09:55 CET] <JEEB> also talking of FFmpeg as a "video editor" is really funny in my opinion, even with just the intent of using it as a way of mocking the project
[14:11:42 CET] <wm4> if only it had EDL capabilities or so (instead of the crazy concat shit)
[14:12:24 CET] <JEEB> btw, as we're on the subject of libavfilter, does filter_complex let you do stuff like "if subtitle track exists in input, do something"?
[14:12:31 CET] <JEEB> IIRC no
[14:15:33 CET] <rcombs> JEEB: no but someone asked about a similar feature a while back
[14:15:51 CET] <rcombs> and it's probably useful enough to be worth implementing but also would be work
[14:15:56 CET] <JEEB> yeah
[14:16:14 CET] <JEEB> and I f.ex. haven't even gotten properly looking into libavfilter madness yet
[14:16:29 CET] <rcombs> like, I only recently added that to `-map` in ffmpeg_opt
[14:16:37 CET] <JEEB> yeah, I noticed that
[14:16:40 CET] <JEEB> quite useful
[14:16:45 CET] <JEEB> except when you want to do hardsubbing :)
[14:16:46 CET] <rcombs> well, maybe it could be easy if you did it in ffmpeg_filter.c in lazymode
[14:16:52 CET] <JEEB> hmm
[14:16:57 CET] <JEEB> hafta check that
[14:17:02 CET] <JEEB> because lazymoodo best moodo amirite?
[14:17:07 CET] <rcombs> do the same check as what I did for -map
[14:17:13 CET] <J_Darnley> u r right
[14:17:23 CET] <rcombs> and if the stream doesn't exist, make a fake one that never gets any packets
[14:17:29 CET] <JEEB> yeah
[14:17:42 CET] <rcombs> JEEB: oh also subtitle streams don't exist in lavfi yet
[14:17:51 CET] <JEEB> uh
[14:18:00 CET] <JEEB> then what is the input in lavfi for subtitle overlays?
[14:18:06 CET] <rcombs> a filename
[14:18:19 CET] <rcombs> something something ubitux something
[14:18:21 CET] <JEEB> you can't overlay a dvd/dvbsub track!?
[14:18:31 CET] <rcombs> that's the overlay filter
[14:18:45 CET] <rcombs> ffmpeg.c contains a hack to convert bitmap subtitles to video streams for lavfi
[14:18:46 CET] <JEEB> yeah, that's why I noted "subtitle overlays" since I remembered the overlay filter
[14:18:53 CET] <JEEB> hah :D
[14:18:58 CET] <JEEB> ffmpeg.c fug yeah
[14:19:00 CET] <rcombs> it's bad
[14:19:08 CET] <Daemon404> uh... the fuck
[14:19:12 CET] <Daemon404> does anyone speak dutch?
[14:19:22 CET] <rcombs> Daemon404: I think I know a guy, why?
[14:19:24 CET] <Daemon404> someone sent me an email in dutch, about ffmpeg
[14:19:31 CET] <Daemon404> and hls and some merge
[14:19:47 CET] <JEEB> "English motherfucker, do you speak it?"
[14:20:12 CET] <JEEB> probably caught your surname and thought you could habla dutch
[14:20:37 CET] <J_Darnley> That's unusual.
[14:20:46 CET] <Daemon404> i geuss my surname is dutch
[14:20:50 CET] <Daemon404> but still.
[14:20:56 CET] <J_Darnley> Most dutch speakers I've met speak good english
[14:21:03 CET] <Daemon404> https://www.dropbox.com/s/dugd1xlv5scqbzm/wat.html?dl=0
[14:21:03 CET] <rcombs> JEEB: so it'd probably be relatively easy to do for bitmap subs since the sub2video hack is already within ffmpeg.c
[14:21:07 CET] <ubitux> JEEB: the lavfi subtitles thing is a hack to extract eia subs, isn't it?
[14:21:35 CET] <JEEB> rcombs: yeah - bitmap subs is what I was thinking of
[14:21:36 CET] <Daemon404> email is the html file i linked
[14:21:50 CET] <rcombs> JEEB: so yeah, you could totally lazymode that
[14:23:04 CET] <J_Darnley> He's asking whether you could address a ticket someone filed
[14:23:32 CET] <J_Darnley> He claims that the issue was created by a commit of yours
[14:23:50 CET] <J_Darnley> oh no, he says someone else claims that
[14:24:07 CET] <rcombs> Daemon404: turns out I indeed knew a guy
[14:24:11 CET] <rcombs> Daemon404: "looking at your last name, I believe I can just talk Dutch with you"
[14:24:16 CET] <Daemon404> ...
[14:24:25 CET] <rcombs> JEEB: nice call
[14:24:25 CET] <Daemon404> lol.
[14:24:26 CET] <JEEB> yeah, I guessed
[14:24:38 CET] <JEEB> lol
[14:24:43 CET] <Daemon404> nobody has spoke dutch since my great gandfather
[14:24:46 CET] <Daemon404> :D
[14:24:52 CET] <rcombs> "two weeks ago I opened a new ticket on the bug tracker. User 'kyl416' said the cause was something you commited. I don't know if you've seen the ticket already or knew the problem, but I would like it if you could take a look"
[14:25:07 CET] <Daemon404> all righty...
[14:25:17 CET] <rcombs> so basically "pls look at ticket"
[14:25:32 CET] <Daemon404> too bad i have bo protected hls streams to test
[14:25:38 CET] <Daemon404> no*
[14:26:13 CET] <Daemon404> unless the one in the ticket works
[14:27:14 CET] <rcombs> Daemon404: apparently google translate also does a fairly good job on that message
[14:27:22 CET] <Daemon404> interesting
[14:27:28 CET] <rcombs> so, <insert admonishment for not checking that first>
[14:27:46 CET] <Daemon404> bad part of doing merges: everything ever is your fault
[14:29:29 CET] <nevcairiel> of course
[14:29:50 CET] <nevcairiel> did he at least link the issue
[14:30:10 CET] <Daemon404> i would post on the trac issue
[14:30:11 CET] <Daemon404> but
[14:30:11 CET] <Daemon404> carl.
[14:30:13 CET] <Daemon404> so no.
[14:30:41 CET] Action: Daemon404 surprised he hasnt demanded a revert yet
[14:31:33 CET] <nevcairiel> apparently the link in the ticket still works
[14:32:02 CET] <Daemon404> ill check it on monday
[14:32:07 CET] <Daemon404> sunday is my no coding day
[14:32:29 CET] <Daemon404> i.e. xbox day
[14:32:59 CET] <Daemon404> someone with a trac account want to post something akin to "derek will look at it on monday"?
[14:33:37 CET] <Fyr> wm4, ffmpeg has effects.
[14:34:01 CET] <Fyr> 3D cube is an effect.
[14:34:10 CET] <nevcairiel> Daemon404: its not exactly a pressing issue, so just look at it when you have time without a comment =p
[14:34:34 CET] <Fyr> if there are many such effects are implemented, why is the cube rotation missing?
[14:35:29 CET] <Daemon404> nevcairiel, its pressing in the sense that i dont want to deal with carl demanding a revert
[14:35:33 CET] <Daemon404> thats all
[14:35:34 CET] <Daemon404> ;)
[14:35:48 CET] <wm4> I'll have to quote kierank here
[14:36:07 CET] <nevcairiel> last action on the ticket was 8 days ago
[14:36:11 CET] <nevcairiel> why would that suddenly show up
[14:36:43 CET] <Daemon404> :P
[14:37:30 CET] <nevcairiel> just wait until we push tep2!
[14:37:36 CET] <JEEB> :D
[14:37:41 CET] <Daemon404> lol
[14:37:56 CET] <kierank> tep2 isnt that extreme
[14:38:07 CET] <nevcairiel> high potential was subtle behavior changes
[14:38:12 CET] <nevcairiel> also, ffserver dies with it =p
[14:38:14 CET] <ubitux> kierank: it will probably break many things
[14:38:25 CET] <Fyr> wm4, ffmpeg developers are bad people and they know it.
[14:38:29 CET] <Daemon404> nevcairiel, i cant see killing ffserver being allowed
[14:38:33 CET] <Daemon404> i expect full scale riots 
[14:38:37 CET] <kierank> tep1 was the patchset that turned this project from a neckbeard project into a 21st century project
[14:39:14 CET] <JEEB> Daemon404: "muh company depends on ffserver"
[14:39:22 CET] <nevcairiel> Daemon404: the new api makes it literally impossible
[14:39:23 CET] <JEEB> "muh router streaming depends on ffserver"
[14:39:37 CET] <nevcairiel> we could keep supporting it for the deprecation period if you use the deprecated st->codec
[14:39:47 CET] <wm4> I still don't get what exactly ffserver does
[14:39:55 CET] <wm4> I mean I know some of the mechanism it uses
[14:40:06 CET] <JEEB> it's a crappity crappy streaming server
[14:40:26 CET] <Daemon404> lol
[14:40:27 CET] <JEEB> of which there are multiple by now
[14:40:39 CET] <JEEB> ffserver just happens to come with FFmpeg
[14:40:45 CET] <JEEB> which is why many poor souls try to run to it
[14:40:49 CET] <JEEB> and then hit a brick wall
[14:40:56 CET] <JEEB> because it's bad and undocumented, among other things
[14:40:58 CET] <Daemon404> we broke someone usign m3u8 as local playlists...
[14:41:01 CET] <Daemon404> i forgot people did that
[14:41:05 CET] <Daemon404> 'broke'
[14:41:41 CET] <wm4> to be fair testing hls from filesystem is nice
[14:41:44 CET] <JEEB> yeah, I think I still have old libavformat because I often grab live playlists and segments
[14:41:48 CET] <wm4> but it probably shouldn't work by default
[14:41:50 CET] <JEEB> for debugging/testing
[14:42:03 CET] <Daemon404> its not hls, wm4 
[14:42:08 CET] <Daemon404> it's a mplaylist of mp3s
[14:42:11 CET] <wm4> oh lol
[14:42:14 CET] <Daemon404> aka where m3u started
[14:42:16 CET] <Daemon404> iric?
[14:42:20 CET] <wm4> tell him to get a real player then
[14:43:33 CET] <nevcairiel> where are these whitelists defined anyway
[14:43:40 CET] <Daemon404> there is a default one
[14:43:44 CET] <Daemon404> for each protocol
[14:43:51 CET] <Daemon404> other than that, it is a user
[14:43:53 CET] <nevcairiel> but hls is a demuxer, not a protocol
[14:44:06 CET] <Daemon404> pretty sure its a protocol
[14:44:09 CET] Action: Daemon404 checks
[14:44:09 CET] <wm4> yeah, but http and file are protocols
[14:44:21 CET] <wm4> no hls is a demuxer
[14:44:29 CET] <Daemon404> well it's in .default_whitelist
[14:44:29 CET] <nevcairiel> we have a hls protocol, but its old and deprecated
[14:44:38 CET] <wm4> eh
[14:44:45 CET] <Daemon404> or something
[14:45:10 CET] <nevcairiel> so should crypto have a default whitelist of allowing it to be used on everything?
[14:45:34 CET] <wm4> crypto is basically just a filter, isn't it
[14:45:39 CET] <nevcairiel> exactly
[14:45:42 CET] <nevcairiel> no reason to restrict it
[14:45:43 CET] <nevcairiel> is there?
[14:45:46 CET] <Daemon404> https://github.com/FFmpeg/FFmpeg/blob/9c75148e6ebc88a0501e3d0242defb6dbdc3c23d/libavformat/file.c#L348
[14:46:07 CET] <Daemon404> should that not work?
[14:46:17 CET] <nevcairiel> why would file have a whitelist of itself
[14:46:25 CET] <nevcairiel> how can a file protocol open another file
[14:46:57 CET] <Daemon404> i think it's for whitelisting stuff opened via that protocol
[14:47:03 CET] <Daemon404> i think
[14:47:04 CET] <Daemon404> i dunno
[14:47:06 CET] <Daemon404> ask michaelni 
[14:47:07 CET] <Daemon404> he wrote it
[14:47:25 CET] <wm4> I totally want a remote playlist poking around in my /dev
[14:50:03 CET] <nevcairiel> hls should just not allow mixing protocols, ie. a manifest from https should not allow opening a llocal file, but if the manifest starts as a file, no reason to block it from opening files
[14:50:06 CET] <nevcairiel> or so i would think
[14:50:40 CET] <nevcairiel> no clue if thats how this whitelist thing works
[14:54:24 CET] <wm4> nevcairiel: yeah, I think the security issue was mainly about a local playlist accessing remote URLs
[14:54:44 CET] <nevcairiel> btw the example playlist in that ticket looks like valid HLS
[14:54:46 CET] <nevcairiel> just downloaded
[14:57:34 CET] <wm4> yeah, that one might be HLS
[14:58:40 CET] <Daemon404> wm4, it says he got it from youtube-dl
[14:59:53 CET] <wm4> youtube-dl can present you HLS and other adaptive streaming crap in the same format, AFAIK
[15:00:08 CET] <wm4> can't be bothered to read what _exactly_ his youtube-dl script does
[15:00:33 CET] <Daemon404> pretty sure it could just concat them
[15:00:35 CET] <Daemon404> since theyre ts
[15:00:36 CET] <wm4> so, what youtube-dl does is exporting such things as list of segments, and you can piece them together yourself
[15:00:48 CET] <Daemon404> im pretty sure thats only one output method
[15:01:00 CET] <Daemon404> im pretty sure it can cobat
[15:01:41 CET] <wm4> yeah, youtube-dl by default concats them with ffmpeg or whatever
[15:02:00 CET] <Daemon404> o
[15:03:08 CET] <Daemon404> BBB, there was an inflate impl by mans btw
[15:03:24 CET] <Daemon404> wait that was reimar
[15:03:31 CET] <Daemon404> i shifted columns in my brain
[15:04:14 CET] <BBB> :)
[15:04:52 CET] <Daemon404> i dont want to NIH zlib
[15:04:58 CET] <Daemon404> its omni-pesent on everything but windows
[15:05:01 CET] <Daemon404> and it has asm for x86
[15:09:37 CET] <Fyr> KKKK
[15:13:28 CET] <Daemon404> ...lol
[15:14:13 CET] <wm4> lol?
[16:06:11 CET] Action: Daemon404 doesnt buy any of these arguments in favour of NIH zlib
[16:07:09 CET] <JEEB> "I'm lazy to build zlib for my cross-compile target
[16:10:04 CET] <nevcairiel> unfortunately its not only about NIH'ing zlib, but also littering the code with #ifs to support both, because it doesnt sound like he wants to emulate the zlib api =p
[16:10:37 CET] <JEEB> yup
[16:11:31 CET] <Daemon404> wtf?
[16:11:53 CET] <nevcairiel> i mean, whats the point of implementing such an API if it wouldnt get used anywhere
[16:12:37 CET] Action: Daemon404 cant wait for libavlibc
[16:13:56 CET] <j-b> [=
[16:13:59 CET] <j-b> ´
[16:14:00 CET] Last message repeated 1 time(s).
[16:14:12 CET] <j-b> oops
[16:14:21 CET] <j-b> my cat walked on my keyboard, sorry
[16:15:30 CET] <Daemon404> "cat"
[16:16:00 CET] <j-b> no, really :)
[16:18:41 CET] <Daemon404> ;)
[16:34:27 CET] <iive> j-b: now you are obligated to show us a picture of your cat :)
[16:35:39 CET] <durandal_1707> or mouse
[16:48:24 CET] <cone-161> ffmpeg 03Paul B Mahol 07master:5451be50a89d: avfilter/vf_waveform: use nonsubsampled yuv output format for lowpass filter
[16:48:25 CET] <cone-161> ffmpeg 03Paul B Mahol 07master:23d8b79a1833: avfilter/vf_waveform: add text to graticule
[16:54:15 CET] <Compn> ; ELSA QUick Codec / MJPEG Codec
[16:54:16 CET] <Compn> aha!
[16:54:20 CET] <Compn> its just mjpeg.
[17:30:05 CET] <cone-161> ffmpeg 03erankor 07master:dd34e89f313b: mov - support seek in encrypted mp4
[17:47:02 CET] <yashpatel> michaelni: For the alternate qualification task for 'FFv1 P frame support' project. what exactly is expected ? I went through the documentation of FFv1, am I supposed to experiment around with the parameters and check corresponding results  ? or I need to think for some alternate algorithmic approaches ?
[17:55:50 CET] <michaelni> yashpatel, the idea is to improve the algorithm or anything thats part of the algorithm that cannot be adjusted via the command line
[17:56:40 CET] <lkajshdf> hello, i'd like to extend the format that ffmpeg's hls muxer accepts, that is i want my segments to contains some timestamp. Where would the most appropriate place to do such a change? Reading the code it seems that the filename generation is happening in hls_start 
[18:00:37 CET] <yashpatel> michaelni: Can you please provide me a benchmark video for the purpose ?
[18:04:56 CET] <durandal_1707> yashpatel: just about any video could do it
[18:05:17 CET] <yashpatel> durandal_1707: Thank You.
[18:11:07 CET] <Compn> any video _with motion_
[18:11:08 CET] <Compn> :)
[18:11:09 CET] Action: Compn runs
[18:11:54 CET] <durandal_1707> yashpatel: you are improving encoding performance, do you know how to do it?
[18:12:40 CET] <yashpatel> durandal_1707: I'm an undergraduate researcher in Computer Vision and Machine learning. So trying to figure out things :P
[18:13:11 CET] <yashpatel> durandal_1707: looking for appropriate research papers, and trying to get idea of things I can change in current algorithm used.
[18:13:38 CET] <yashpatel> durandal_1707: Any suggestion / guidance ? Please ?  
[18:14:53 CET] <yashpatel> durandal_1707: I'm planning to make changes and compare compression ration outputs and stuff!
[18:15:22 CET] <durandal_1707> first find out what current ffv1 use
[18:27:28 CET] <cone-161> ffmpeg 03foo86 07master:db44b599805f: avcodec/dca: clear X96 channels if nothing was decoded
[18:29:34 CET] <JEEB> anyone want to poke this with a review https://github.com/VFR-maniac/ffmpeg/commit/172fe56b1a4a063ff2feb14d3394129eb155b40d ?
[18:32:45 CET] <lorddoskias> is there an easy way to add a timestamp to the filename generate by the hls muxer?
[19:07:48 CET] <wm4> lol
[19:19:25 CET] <Prelude2004c> hey guys.. i hope you are all doing well. I would like some advice. Using ffmpeg with vdpau  but vdpau has some issues with decoding h264. Does anyone know if a fix is in the works or some other way to do it?
[19:19:56 CET] <wm4> Prelude2004c: I know no of such problems, either your code is broken or your GPU (or drivers) are
[19:22:57 CET] <Prelude2004c> my code ? seems pretty simple.. -hwdecoder vdpau 
[19:23:08 CET] <Prelude2004c> same code decoding mpeg2video works just fine ... 
[19:23:27 CET] <Prelude2004c> i see the GPU decoder work correctly without any issues with mpeg2video
[19:23:32 CET] <Prelude2004c> when it comes to h264 .. its a no go
[19:24:52 CET] <MelchiorGaspar> what do you have for a GPU?
[19:50:33 CET] <MelchiorGaspar> <Prelude2004c> what type of graphics card do you have?
[19:54:40 CET] <carpediem_> Hello developers. I have been doing the fuzzing tests on ffmpeg and found some files which cause it to hang. Still searching for files which can cause a crash. Wanted to ask if you guys use some specific test files(or format) to check ffmpeg stability?
[19:58:28 CET] <kierank> carpediem_: a hang is fine
[19:58:41 CET] <kierank> that's also a bug
[19:58:49 CET] <kierank> the trick is to use small files at around 10kB
[19:58:52 CET] <kierank> obscure formats too 
[19:58:57 CET] <Compn> carpediem_ : any files at http://samples.ffmpeg.org
[20:02:17 CET] <carpediem_> small files, but a large number?
[20:02:29 CET] <carpediem_> I have been using afl-fuzz for testing of files
[20:03:37 CET] <kierank> zzuf is probably better to begin with
[20:03:45 CET] <kierank> but hangs are bugs so please report them
[20:03:49 CET] <kierank> and fix if possible
[20:04:25 CET] <carpediem_> where should I report the files which caused hangs?
[20:04:56 CET] <carpediem_> will try fixing the bugs
[20:08:06 CET] <kierank> carpediem_: https://ffmpeg.org/bugreports.html
[20:08:39 CET] <Compn> http://trac.ffmpeg.org , read those intructions ^^
[20:08:40 CET] <Compn> :)
[20:10:58 CET] <carpediem_> Thanks (for almost pointing it out)! 
[20:18:49 CET] <Prelude2004c> sorry for the huge delay.. i am using an M4000 card
[20:32:32 CET] <cone-161> ffmpeg 03Paul B Mahol 07master:02f84215461a: avfilter/vf_histogram: add 12bit depth support
[20:32:33 CET] <cone-161> ffmpeg 03Paul B Mahol 07master:6e19120ea206: avfilter/vf_vectorscope: add 12bit depth support
[21:19:38 CET] <cone-161> ffmpeg 03Boris Nagels 07master:1109ed7973c7: avformat/rtpenc: Fix integer overflow in NTP_TO_RTP_FORMAT
[21:23:17 CET] <cone-161> ffmpeg 03Carl Eugen Hoyos 07master:719276b8609e: lavc/cfhd: Sanitize avpriv_report_missing_feature() usage.
[21:39:27 CET] <cone-161> ffmpeg 03Carl Eugen Hoyos 07master:59793785f4ba: lavc/cfhd: Remove one more \n from avpriv_report_missing_feature().
[21:40:47 CET] <cone-161> ffmpeg 03Carl Eugen Hoyos 07master:aebfbe5c1368: lavf/riffdec: Remove \n from avpriv_report_missing_feature().
[22:29:28 CET] <cone-161> ffmpeg 03Carl Eugen Hoyos 07master:e356487fdae8: lavf/img2dec: Use jpeg constants in jpeg_probe().
[22:32:23 CET] <MelchiorGaspar> <Prelude2004c> sorry for the huge delay.. i am using an M4000 card
[22:33:39 CET] <MelchiorGaspar> ^ Whats the maufacturer?  is it ATI, Nvidia, etc? because I've never heard of the m4000
[22:34:47 CET] <ethe> nvidia m4000 quadro maybe
[22:35:21 CET] <Prelude2004c> nvidia 
[22:35:26 CET] <Prelude2004c> yes quadro
[22:36:04 CET] <MelchiorGaspar> so this is a work station?
[22:36:46 CET] <MelchiorGaspar> quadro are best for CAD/CAM
[22:38:25 CET] <JEEB> please ask for support from nvidia then. if it works with other pieces of hardware it's either due to your drivers or hardware
[22:39:01 CET] <MelchiorGaspar> Geforce is the best of Nvidia I have a http://www.amazon.com/EVGA-GeForce-Display-Graphics-02G-P4-3757-KR/dp/B00J0ISHMQ
[22:40:18 CET] <Shiz> i don't think this is #recommend-a-gpu-devel
[22:41:46 CET] <JEEB> the m4000 is listed as vdpau level E in the latest nvidia drivers, so please goto nvidia if you are having issues that nobody else seems to be having. if it's actually noted as an issue in FFmpeg, it will get fixed but until then it seems more likely that either your driver or hardware is borked
[22:41:51 CET] <JEEB> so please
[22:41:52 CET] <JEEB> goto nvidia support
[22:46:17 CET] <wm4> most likely his code is just buggy
[22:46:36 CET] <JEEB> he seems to be a cli user
[22:46:43 CET] <JEEB> using whatever hack there is in ffmpeg cli
[22:47:06 CET] <JEEB> at least him noting that there his "code" was something that looked like a ffmpeg cli parameter
[22:50:13 CET] <dinux5> Hi everyone. I want to apply for Outreachy 2016.
[22:50:24 CET] <Compn> hello
[22:50:28 CET] <Compn> welcome dinux5
[22:51:12 CET] <dinux5> I have cloned the code base and need help on my further steps . :)
[22:51:36 CET] <Compn> sure , ask away
[22:52:10 CET] <dinux5> I have selected TrueHD encoder as the project where I would want to work(keeping in mind my set of skills.)
[22:52:27 CET] <dinux5> How do I proceed with completing the Qualification task?
[22:53:27 CET] <dinux5> Or is there something else i was supposed to do before I select any project :)
[22:55:43 CET] <nevcairiel> contact the mentor first to make sure everything is in order, he can then also inform you about how to start
[22:58:51 CET] <dinux5> Thanks ! I have already mailed my project's mentor. Is there anything suggested I should do till I get a reply. :)
[23:11:41 CET] <ethe> dinux5: getting familiar with the FFmpeg codebase (particularly libavcodec), if you're not already might be a good idea
[23:13:51 CET] <dinux5> ethe: Oh k. Thanks a lot :) .
[00:00:00 CET] --- Mon Mar  7 2016


More information about the Ffmpeg-devel-irc mailing list