[Ffmpeg-devel-irc] ffmpeg.log.20150917

[burek]

[02:44:58 CEST] <maduro> can the -threads value passed during encoding have any negative effect on playback on resource-constrained devices? like if i encode on a 40-core beast with -threads 40 will that mess things up for playback on a smartphone?
[02:45:39 CEST] <maduro> particularly when trying to stream
[02:46:04 CEST] <c_14> no
[02:46:15 CEST] <maduro> thx
[02:47:32 CEST] <fred1807> how do I convert 1 single .png image into a 8 second .h264 movie file ?
[02:47:57 CEST] <c_14> ffmpeg -loop 1 -i png -c:v libx264 -t 8 out.h264
[02:48:19 CEST] <fred1807> thanks a lot
[02:57:08 CEST] <maduro> are there any video quality implications to using too many/too few threads/slices? or is it purely just affecting the time it takes to encode?
[03:04:32 CEST] <kepstin> assuming you're talking about x264?
[03:05:18 CEST] <kepstin> if you're using the regular thread mode (not slice-based), there's a slight quality hit to more threads
[03:06:06 CEST] <kepstin> and sliced mode explicitly divides the picture into multiple parts, meaning that the encoder thread doing one slice can't reference data from a different slice
[03:07:52 CEST] <kepstin> iirc, in the regular thread mode, the quality hit happens because it starts encoding the next frame before the previous frame completes, which means that when it's encoding the next frame, it can't reference data from the parts of the previous frame that haven't been encoded yet
[03:08:19 CEST] <kepstin> this is all pretty minor, particularly when encoding large videos with reasonable numbers of threads
[03:09:06 CEST] <kepstin> the sliced threads mode is mostly useful only if you're doing really-low-latency encoding (it's the only thread mode supported with "zerolatency" tuning), or have some other requirement for using slices anyways.
[03:10:48 CEST] <maduro> thank you kepstin
[05:41:00 CEST] <BullHorn> hello
[05:41:16 CEST] <BullHorn> i just downloaded the x64 zip of ffmpeg for windows and
[05:41:28 CEST] <C0nundrum>  How do i make it so that if a video file exist, it appends to it ?
[05:41:30 CEST] <BullHorn> i added the \bin\ to my windows10's environment Path
[05:41:50 CEST] <BullHorn> when i try to convert a .flv to .mp4 i get a weird 'could not find codec parameters for stream 0"
[05:41:52 CEST] <BullHorn> any ideas?
[05:42:34 CEST] <BullHorn> the command im using is ffmpeg -i D:\input.flv -c copy -copyts D:\output.mp4
[06:42:30 CEST] <IanWizard> How's Ogg Opus for music, particularly compared to Vorbis?  I've looked a bit, but there doesn't seem to be any good comparison or discussion of the merrits of each.
[07:02:48 CEST] <CounterPillow> Vorbis is better in terms of seeking, as far as I know, Opus is more efficient at some bitrates.
[08:58:57 CEST] <akae_> Good morning. Can anybody point a howto about compiling ffmpeg with Intel QSV capabilities under Ubuntu ?
[09:41:14 CEST] <Canar> IanWizard: Opus sounds transparent to me at 100kbps, though I've spent little to no time double-blind-testing myself with it so I may grow more sensitive
[10:03:32 CEST] <xenos> Hi all, does anyone know why vbr mode with libfdk_aac returns the following warning: "Note, the VBR setting is unsupported and only works with some parameter combinations" ? mediainfo reports that the resulting file has a constant bit rate mode but I would have expected it to report a variable bit rate mode.
[10:13:10 CEST] <Mavrik> Well, that's probably the "other parameter combinations" where VBR doesn't work.
[10:16:14 CEST] <xenos> I followed the VBR mode in the wiki here: https://trac.ffmpeg.org/wiki/Encode/AAC
[10:17:59 CEST] <akae_> Good morning. Can anybody point a howto about compiling ffmpeg with Intel QSV capabilities under Ubuntu ?
[11:23:13 CEST] <DelphiWorld> hi guys
[11:23:21 CEST] <DelphiWorld> is it pocible to include a transcript in a m4a file?
[12:07:52 CEST] <xenos> Does anyone know why the h264 encoding settings are lost from a file when extracting a segment using ffmpeg?
[12:10:02 CEST] <iive> does it happen if the segment includes the start of the original file?
[12:12:10 CEST] <BtbN> h264 encoding settings?
[12:17:28 CEST] <xenos> iive: No, the encoding settings are present in the segment if the segment includes the start of the original.
[12:17:46 CEST] <xenos> BtbN: The encoding settings that are displayed using mediainfo on the MP4 file.
[12:23:14 CEST] <iive> BtbN: x264 stores parameters used for the encoding. but it seems to put it only once in the bitstream.
[12:38:37 CEST] <xenos> iive: Is there anyone of copying that data to the new segment?
[12:39:03 CEST] <iive> i don't know...
[12:47:05 CEST] <xenos> iive: Ok, thanks for your help.
[15:20:03 CEST] <jules> hey, I have a basic development question: I want to compare the (encoded) sizes of I and P (or B) frames in movies. I assume that the thing I want to look at is the AVPacket.size field? I did this for a movie and it turns out that the P and B frames are significantly larger that the I Frames. Am I missing something?
[15:40:44 CEST] <zidanne> latest ffmpeg from master branch: this stream does not work I dont know why. It works with VLC. Can anyone check? http://trtcanlifm-lh.akamaihd.net/i/RADYO1_1@182345/master.m3u8
[15:42:16 CEST] <satiender> Hi anybody install sdl-devel package for ffmpeg developemnt
[15:44:54 CEST] <satiender> anybody can please
[15:44:56 CEST] <satiender> help me
[16:42:03 CEST] <ace040> hello
[16:42:54 CEST] <ace040> I'd need some help configuring a custom FFmpeg build
[16:43:40 CEST] <durandal_1707> ace040: yes?
[16:44:40 CEST] <ace040> I'd need a shared library that: do not deal with sound, is able to demux/decode any video file, is only able to write H.264 to MP4 and MSMPEGV3 to WMV
[16:44:50 CEST] <ace040> this is what I have so far:
[16:45:55 CEST] <ace040> http://pastebin.com/7yTqU2Gn
[16:47:20 CEST] <ace040> I'm not sure about some things, such as what is the name of the muxer for WMV? will it fail if I feed it a video with an audio stream (I only want FFmpeg to disregard it)?
[16:49:07 CEST] <ace040> also I'm doing the rendering myself and only need the decoded picture from FFmpeg, so may I safely --disable-hwaccels ?
[16:49:55 CEST] <durandal_1707> hwaccels do not do rendering
[16:51:17 CEST] <ace040> oh ok, tbh I assumed so because of their names
[16:56:13 CEST] <ace040> what I'd like to have is: any video file -> FFmpeg -> picture in system memory, and picture -> FFmpeg -> video file
[16:56:33 CEST] <ace040> would I require hwaccels for this?
[17:04:36 CEST] <fritsch> ace040: no, but they don't harm
[17:04:47 CEST] <fritsch> think of content your cpu is too slow to decode
[17:06:32 CEST] <ace040> mm right, our application is already quite gpu-intensive though :)
[17:07:46 CEST] <ace040> may I safely disable filters and parsers?
[17:11:02 CEST] <ace040> as the only 'filtering' I do is calling sws_scale for pixel format conversion
[18:12:09 CEST] <gordonbleuu> hi everyone, i'm trying to create a mpd-manifest and receive an error. i pasted the console-output here: http://pastebin.com/cCyAdCyy so i tried to specify the pixel format with "-pix_fmt yuv420p", but that didn't help. any ideas?
[20:09:30 CEST] <waressearcher2> is anyone here ?
[20:11:13 CEST] <retard> 299
[20:12:47 CEST] <waressearcher2> if I concatenate 3 files, video_1,video_2,video_3 and video_2 doesn't have audio track, in output video, after video_1 starts playing video_2 but the audio from video_3, is there a way to make it automatically add silence for a video without audio track ?
[20:17:30 CEST] <waressearcher2> or put it other way, is there a way to add silence to a video ? say I want to copy from video some part but I don't want to copy audio track but I also don't want to get video resulting video without audio, so is there a way to put silence audio in it ?
[20:20:23 CEST] <waressearcher2> there is a long way, like generating silence then cut out how much I need then join that silence audio to a video file, but I don't want to do all that stuff I need ffmpeg do all
[20:28:52 CEST] <relaxed> waressearcher2: ffmpeg -i input.mkv -f lavfi -i aevalsrc=0 -shortest ...
[20:29:40 CEST] <waressearcher2> right, I was about to use that
[20:55:02 CEST] <waressearcher2> but it could've been better if ffmpeg adds silence automaticaly if one of the concatenated videos doesn't have audio track at all
[23:19:41 CEST] <well0ne> good evening everyone
[23:20:09 CEST] <well0ne> i'm using a raspberry pi with openwrt and a asus wlan stick, which is recognised correctly and works
[23:20:25 CEST] <well0ne> but my wireless router is on channel 13 and i can't see / scan it
[23:20:54 CEST] <well0ne> i tried to set regulray domain, didnt work either, so i found the patch from jow (reghack) but its not prepared for raspberry pi use
[23:20:58 CEST] <well0ne> any advice for me?
[23:21:47 CEST] <jasom> well0ne: I think you are asking in the wrong channel?
[23:22:10 CEST] <well0ne> omg, i am :D i'm so sorry :d
[23:33:06 CEST] <waressearcher2> that guy has reall problems with channels
[23:48:56 CEST] <iH2O> what is the most straightforward way to convert f.mp4 into f.flv
[23:49:10 CEST] <iH2O> *to
[23:50:44 CEST] <chungy> ffmpeg -i f.mp4 f.flv
[23:51:31 CEST] <iH2O> oh, thank you
[23:51:39 CEST] <BtbN> Well, -c copy
[23:51:45 CEST] <iH2O> ?
[23:51:52 CEST] <iH2O> do I need -c ?
[23:51:56 CEST] <BtbN> otherwise it's the most straight forward way to ruin the quality.
[23:51:57 CEST] <portal> ffmpeg -i f.mp4 -c copy f.flv
[23:52:12 CEST] <iH2O> good, i want it lossless
[23:53:50 CEST] <waressearcher2> so it doesn't convert anything it just puts video and audio stream from one container format to another
[23:54:11 CEST] <iH2O> it converts the format only
[23:54:17 CEST] <iH2O> i gather
[00:00:00 CEST] --- Fri Sep 18 2015