[Ffmpeg-devel-irc] ffmpeg.log.20171125

[burek]

[00:00:16 CET] <kepstin> Scrappy: hmm. what container & codec?
[00:01:21 CET] <kepstin> in many cases the avg_frame_rate is actually probed or guessed by looking at a few frames, and in some cases i guess it might not be fully determined until decoding. I wonder how the ffprobe -show_streams output compares.
[00:12:00 CET] <AppleTor> Ola, regarding of my question above, is it possible to make ffmpeg retry downloading the failed chunk? instead of skipping it
[00:58:16 CET] <CCFL_Man> anyone here have a hacked tivo hd with mfs ftp?
[02:56:15 CET] <SortaCore> what are reference frames, and when should I need more of them?
[02:56:29 CET] <SortaCore> Video: h264 (h264_qsv) (High), 1 reference frame (avc1 / 0x31637661), nv12, 960x480, q=2-31, 1000 kb/s
[02:57:34 CET] <iive> reference frames are the frames that are used as basis for motion compensation
[02:58:47 CET] <SortaCore> when should there be > 1?
[03:28:06 CET] <DHE> more are usually helpful for image quality, especially when the video is prone to brief bursts of temporary change (eg: a flash of lightning)
[03:28:34 CET] <DHE> but diminishing returns sets in quickly, and the number of reference frames is constrained by the 'Level' capabilities of your player
[03:32:38 CET] <TheRock> can anyone confirm that ffmpeg builds with vs 2013
[05:12:13 CET] <factor> I get the error "ERROR: vaapi requested but not found." when trying to compile Kodi for android. Said to come here for ffmpeg erros.
[06:01:56 CET] <vigilante> hi can i remux using ffmpeg?
[06:02:06 CET] <vigilante> from Bluray dvd?/
[06:02:21 CET] <vigilante> and get a mkv output?
[06:28:03 CET] <TheRock> while buildin static ffmpe
[06:28:29 CET] <TheRock> do i need filters, bsfs, indevs, protocols,
[06:28:35 CET] <TheRock> or can i disable them all
[06:28:51 CET] <TheRock> *i just need to play a video* such as mp4, wmv, avi, etc.
[10:18:06 CET] <illegal> I am having a problem converting to webp, it gives the images transparent borders https://i.imgur.com/D3CKzHg.png | https://i.imgur.com/r1a9at6.png
[10:20:52 CET] <illegal> I am using this .bat to convert a whole folder https://ghostbin.com/paste/gvpje
[10:21:22 CET] <illegal> could just be whatever program is previewing the thumbnails
[10:53:54 CET] <illegal> Nevermind I believe it's the thumbnail previewer, sorry
[11:50:05 CET] <radicate> Hey guys, perhaps someone knows what's required in order to make the "rtmp_conn" option to be available? I'm getting this error when trying to use it: "Unrecognized option rtmp_conn"
[12:04:57 CET] <Yosh12345mpeg-gr> hi, i have one question :) does anyone have some kind of script which do precise video cutting which would re-encode  to first keyframe, and copy the rest of data ?
[12:06:02 CET] <Yosh12345mpeg-gr> or --- to dump precise keyframe offsets, I'll do script myself :)
[12:10:24 CET] <seleejane> hi
[14:17:48 CET] <pl0t> hi, mplayer works with alsa plugin, ffplay does not play (although pulseaudio plug appears, it is silent as seen in pavucontrol), youtube does work, soundcloud does not, what could be the problem?
[14:22:19 CET] <faLUCE>  Hello. I created this library, which wraps ffmpeg:  https://github.com/paolo-pr/laav . It includes a low-latency streaming system, maybe the first open source one. Give me a feedback
[14:23:16 CET] <DHE> in what way is it low latency?
[14:23:36 CET] <faLUCE> the correct term is "lowest possible"
[14:23:44 CET] <DHE> my question stands
[14:23:45 CET] <faLUCE> (I defined it as "optimum latency")
[14:23:58 CET] <faLUCE> well, all the modules are optimized in latency
[14:24:01 CET] <faLUCE> this is the meaning
[14:24:23 CET] <atomnuker> do you use draw_horiz_band()?
[14:26:08 CET] <JEEB> talking of low latency https://www.youtube.com/watch?v=A4L5xEXXlas
[14:26:08 CET] <atomnuker> if not and you want the best possible latency you should
[14:27:36 CET] <faLUCE> [14:24] <atomnuker> do you use draw_horiz_band()? <--- for what?
[14:28:37 CET] <atomnuker> to get lines of a frame before the whole frame's done rendering?
[14:28:50 CET] <faLUCE> atomnuker: I use gstreamer for rendering
[14:28:59 CET] <atomnuker> *decoding
[14:29:15 CET] <faLUCE> I use gstreamer for decoding too
[14:29:52 CET] <faLUCE> ffmpeg is on the server side
[15:30:40 CET] <durandal_1707> faLUCE: expect lawyers knocking on you door
[15:47:47 CET] <pl0t> ah, and ffmpeg output: https://pastebin.com/WwnQekiC https://pastebin.com/5mLiDAvq <- stucks at that
[15:50:30 CET] <redrabbit> https://images-na.ssl-images-amazon.com/images/I/71Y7Lx7S8dL._SL1100_.jpg < bwarf
[15:55:43 CET] <debianuser> pl0t: If your problem related to pulseaudio you may want to repeat your question in #pulseaudio channel.
[16:26:42 CET] <faLUCE> durandal_1707: why?
[16:30:12 CET] <concatter98298> Hi
[16:30:53 CET] <concatter98298> please look into this code https://pastebin.com/AztN4S7w
[16:30:59 CET] <concatter98298> its batch
[16:31:21 CET] <concatter98298> basically loop all files in folders and add them to concat.txt
[16:31:37 CET] <concatter98298> file 'name' /n duration 0.3
[16:31:45 CET] <concatter98298> video ends up looking shit
[16:32:07 CET] <concatter98298> I want to make all my videos 0.3 seconds and connect them into one video
[16:37:37 CET] <faLUCE> what is the problem with my library?
[16:37:38 CET] <faLUCE> Ricordo che tempo fa qui ci fu una surreale discussione sulla Collezione Farnese, dove io tentavo inutilmente di far capire che questo patrimonio è superiore, per mole e importanza, a quello attualmente contenuto nel Louvre.
[16:37:39 CET] <faLUCE> A distanza di qualche mese, ecco che spunta questo video, dove il direttore del MANN, sottolinea appunto che si tratta del museo archeologico più grande al mondo.
[16:37:41 CET] <faLUCE> https://www.youtube.com/watch?v=vCkCU7xvekM
[16:37:42 CET] <faLUCE> Ma lasciamo pure le persone credere che il non plus ultra, in Italia, siano gli Uffizi.
[16:37:44 CET] <faLUCE> In ogni caso, a questo tizio (Giulierini), benché ignaro del tutto che l'arte/archeologia etrusca nella zona di Capua è importante e rilevante quanto quella umbra e toscana (ma lo scoprirà col tempo), l'Italia tutta dovrebbe fare un monumento.
[16:37:46 CET] <faLUCE> Quello che Giulierini è stato in grado di fare per riportare in vita il MANN ha dello stupefacente e commovente.
[16:37:47 CET] <faLUCE> Ecco: ci vorrebbero, in Italia, esattamente persone come Giulierini, e NON come acremone.
[16:37:49 CET] <faLUCE> Sentite cosa dice Giulierini di Firenze:
[16:37:50 CET] <faLUCE> https://www.youtube.com/watch?v=vCkCU7xvekM&t=10m20s
[16:37:51 CET] <faLUCE> Finalmente un toscano non toscanocentrico.
[16:37:53 CET] <faLUCE> Se acremone fosse toscano e se fosse in qualche ministero culturale, sarebbe la rovina dell'Italia.
[16:37:55 CET] <faLUCE> sorry
[16:37:56 CET] <faLUCE> (I pasted a long code in a wrong channel)
[16:40:30 CET] <pl0t> :D right after "what is the problem?" came a long paste, just like some navy seals stuff
[16:50:12 CET] <faLUCE> please can you help me understanding if all is ok about the license of my library?  https://github.com/paolo-pr/laav
[16:50:22 CET] <faLUCE> (it wraps ffmpeg)
[16:50:51 CET] <faLUCE> I used GPL 3.0
[18:42:11 CET] <leewdch> hello. just a question, if a codec (let's say opus) is compatible with two container types (.opus and .ogg), is it safe to simply rename a file 'test.opus' to 'test.ogg'? since I don't want to re-encode the file and just change the containers I can just rename to swap it, right?
[18:44:52 CET] <SortaCore> lol
[18:45:21 CET] <SortaCore> you may be able to export opus to a raw opus format, and then re-add it in ogg?
[18:45:41 CET] <SortaCore> but renaming doesn't change the container format
[18:45:58 CET] <SortaCore> most video players will read the type from the content, not from file ext
[18:46:50 CET] <leewdch> I see
[18:47:02 CET] <leewdch> then someone lied to me on the internet \o
[18:47:12 CET] <SortaCore> well, take H264
[18:47:49 CET] <SortaCore> you can have it in a MP4 or MKV
[18:48:18 CET] <SortaCore> but MP4 and MKV are particular container formats for several streams at once - audio, video, subtitles
[18:48:44 CET] <SortaCore> and those container formats are obviously used in the file's content, not in the file name
[18:49:14 CET] <SortaCore> so whether you have MP4 in a .mkv may upset some players (maybe VLC?) otherwise it'll just read the real format
[18:49:39 CET] <leewdch> so if a file can be wrapped in two different containers, can I just ffmpeg -i <input> -c copy <output.ogg> ?
[18:49:58 CET] <sfan5> yes
[18:50:47 CET] <sfan5> .opus and .ogg are just two file extensions for the same format btw
[18:50:57 CET] <sfan5> so you can in fact just rename it
[18:51:10 CET] <leewdch> oh, you see? that's what I was asking in the first place
[18:52:17 CET] <leewdch> sometime in tutorials you are asked to rename .img to .iso for whatever reasons, so I was wondering if you can do that too if a file can be wrapped in two different containers anyway
[18:52:32 CET] <leewdch> like opus
[18:54:06 CET] <sfan5> you can't generally do that, it happens to be possible in this case
[18:54:13 CET] <JEEB> 32
[19:08:56 CET] <SortaCore> there we go then
[19:09:11 CET] <SortaCore> so mp4 <-> mkv is bad, but ogg and opus are same container format anyway
[19:09:42 CET] <SortaCore> although I've not heard of a set ext for ogg video
[19:32:23 CET] <shmibs> ogv?
[19:32:30 CET] <shmibs> and oga
[21:25:38 CET] <nahro> hello
[21:26:24 CET] <nahro> when using -itsoffset for audio, the audio glitches for a brief moment at the start
[21:26:38 CET] <nahro> anybody else have this issue
[21:56:02 CET] <durandal_1707> nahro: pastebin full output
[22:00:07 CET] <durandal_1707> lol
[22:01:12 CET] <TheRock> you made him close the page
[22:48:18 CET] <TheRock> is there a way to build ffmpeg with software renderer only
[22:48:24 CET] <TheRock> in ./configure
[22:49:14 CET] <furq> renderer?
[22:49:55 CET] <JEEB> while there's some funky things in libavdevice, FFmpeg itself is not really a thing to utilize for rendering
[22:49:59 CET] <JEEB> (of video)
[22:50:13 CET] <TheRock> i mean processing
[22:50:18 CET] <TheRock> for encoding and decoding
[22:50:26 CET] <TheRock> i saw some nvidia specific parts
[22:50:30 CET] <TheRock> i don't need that
[22:50:59 CET] <JEEB> yea, since the header is included it can get enabled by default in a lot of places, but will not get utilized by default
[22:51:09 CET] <TheRock> I think it was nvdec
[22:51:20 CET] <TheRock> i caused compile errors
[22:51:22 CET] <TheRock> it*
[22:51:24 CET] <JEEB> you have to specifically ask for hwdec or hwenc
[22:51:26 CET] <JEEB> huh
[22:51:43 CET] <JEEB> I haven't heard of compile errors with it, but it's always possible :P
[22:51:56 CET] Action: JEEB has built for *nix and windows
[22:51:56 CET] <TheRock> I compile with VS2013 and SDK 7.1 to support XP
[22:52:09 CET] <TheRock> and VS2013 treats some stuff as errors
[22:52:12 CET] <TheRock> where as vs2017 wont
[22:52:24 CET] <JEEB> let's see our MSVC FATE instances
[22:52:24 CET] <TheRock> the vs2017 compile went well.
[22:52:45 CET] <JEEB> http://fate.ffmpeg.org/report.cgi?slot=x86_64-msvc12-windows-native&time=20171125193353
[22:52:51 CET] <JEEB> the MSVC 2013 thing seems to be passing
[22:53:03 CET] <JEEB> it fails a test, but that's after compilation
[22:54:06 CET] <TheRock> The current git clone fails in \libavcodec\nvdec_264.c
[22:54:07 CET] <TheRock> line 111
[22:54:22 CET] <TheRock> with VS2013
[22:54:32 CET] <TheRock> so i configured with --disable-nvdec
[22:54:34 CET] <JEEB> that thing I just linked is MSVC 2013
[22:54:36 CET] <TheRock> it compiles now
[22:54:46 CET] <TheRock> Yeah, i know
[22:54:46 CET] <JEEB> and it seems like 8794e0c037  is current master
[22:54:56 CET] <JEEB> so current master seems to pass in MSVC 2013
[22:55:14 CET] <JEEB> and yes, you can disable it like that if you have some specific issues
[22:55:30 CET] <JEEB> that said, I'm just saying that our testing suite finds no compilation errors and has nvdec enabled
[22:55:49 CET] <TheRock> I downloaded the current git repository
[22:56:00 CET] <TheRock> is that 8794e0c037?
[22:56:03 CET] <JEEB> git show
[22:56:07 CET] <JEEB> shows you the current hash
[22:56:18 CET] <JEEB> but yes, currently it is at that hash
[22:56:25 CET] <JEEB> like, right now. which is what the test run tested
[22:57:03 CET] <TheRock> hmmm
[22:57:07 CET] <JEEB> anyways, what I was trying to point out is that whatever's happening to you is specific to something
[22:57:17 CET] <JEEB> are you on an older version of FFmpeg, or older version of MSVC 2013?
[22:57:25 CET] <JEEB> since 2013 did get a good few updates
[22:57:47 CET] <TheRock> i haven't checked, but i use the current ffmpeg version
[22:58:00 CET] <TheRock> maybe i should update 2013
[23:02:49 CET] <TheRock> the error was weird anway
[23:03:12 CET] <TheRock> I wouldnt know how to fix it
[23:14:52 CET] <TheRock> looks like it finished
[23:15:04 CET] <TheRock> w/o othe errors
[23:15:08 CET] <TheRock> other*
[00:00:00 CET] --- Sun Nov 26 2017