[Ffmpeg-devel-irc] ffmpeg.log.20150331

burek burek021 at gmail.com
Wed Apr 1 02:05:01 CEST 2015


[00:01:36 CEST] <prelude2004c> sorry to keep bothering but i am still stuck
[00:01:37 CEST] <prelude2004c> :(
[00:01:45 CEST] <prelude2004c> nothing i do seems to get past this
[00:02:30 CEST] <IlIlIlIlIlIlI> prelude2004c: construct a MWE (no extraneous switches) and submit a bug?
[00:02:59 CEST] <prelude2004c> not sure what that means
[00:03:03 CEST] <prelude2004c> i am a novice at this
[00:03:21 CEST] <prelude2004c> i dont think this is a bug.. i think this is just me not knowing the proper config
[00:03:35 CEST] <IlIlIlIlIlIlI> http://en.wikipedia.org/wiki/Minimal_Working_Example
[00:03:46 CEST] <IlIlIlIlIlIlI> so try with a smaller example that _should_ work
[00:04:06 CEST] <prelude2004c> oh.. i could try that i guess...
[00:04:08 CEST] <prelude2004c> let me try with one
[00:04:27 CEST] <IlIlIlIlIlIlI> perhaps one everyone can run, without even that mcast input ...
[00:05:09 CEST] <prelude2004c> well the issue here is the problem..
[00:05:27 CEST] <prelude2004c> when i run ffmpeg -i multicast -c something - c somethign - c something
[00:05:33 CEST] <prelude2004c> it does't start the 3rd one
[00:05:39 CEST] <prelude2004c> if i just use 2 outputs its fine
[00:05:47 CEST] <prelude2004c> when i use 3 outputs it is not fine on the 3rd one
[00:06:09 CEST] <prelude2004c> some limit that can be adjusted ? seems like 2 multiple outupts by default only
[00:06:24 CEST] <prelude2004c> so without running any filters, i have no problem with 2 outupts
[00:06:27 CEST] <prelude2004c> the issue becomes on 3
[00:06:35 CEST] <IlIlIlIlIlIlI> what about with -i whatever?
[00:06:42 CEST] <prelude2004c> so i am trying to use filter_complex to resolve. but not sure if it will
[00:06:43 CEST] <IlIlIlIlIlIlI> -i wghatever.mp4, e.g.
[00:06:59 CEST] <prelude2004c> you mean a file instead of multicasdt ?
[00:07:02 CEST] <prelude2004c> one sec let me try that
[00:07:11 CEST] <IlIlIlIlIlIlI> I mean a MWE anyone can run to confirm!
[00:13:45 CEST] <prelude2004c> tried with basic just file
[00:13:50 CEST] <prelude2004c> :( .. only outputs 2
[00:14:59 CEST] <IlIlIlIlIlIlI> pastebin
[00:16:46 CEST] <prelude2004c> http://pastebin.com/1Cg9iuNy
[00:16:51 CEST] <prelude2004c> so, with 2 it works
[00:16:56 CEST] <prelude2004c> with 3 .. it gives me that error
[00:17:27 CEST] <prelude2004c> code is found here : http://pastebin.com/U1hFQc7i
[00:27:48 CEST] <IlIlIlIlIlIlI> prelude2004c: http://pastebin.com/Be5y3vSw
[00:28:07 CEST] <IlIlIlIlIlIlI> so if this doesn't work, strip down more switches, and if it still doesn't work
[00:29:12 CEST] <IlIlIlIlIlIlI> submit a bug report, I guess
[00:39:25 CEST] <prelude2004c> <IlIlIlIlIlIlI> - hey, thank you. Tried but same error " Error while opening encoder for output stream #2:0 - maybe incorrect parameters such as bit_rate, rate, width or height"
[00:42:44 CEST] <prelude2004c> works good with only 2 inputs
[00:42:46 CEST] <prelude2004c> i mean outputs
[00:42:54 CEST] <prelude2004c> but if i include the 3rd.. nada
[00:44:13 CEST] <IlIlIlIlIlIlI> prelude2004c: submit a bug report, or ask on #ffmpeg-devel in 18 hours
[00:45:19 CEST] <prelude2004c> cool, but why 18 hours :)
[00:47:35 CEST] <IlIlIlIlIlIlI> prelude2004c: it's getting late ;)
[00:50:40 CEST] <prelude2004c> last question.. anyone know when using the nvenc library for the GPU encoding.. how one would assign a stream to that GPU ?
[00:50:51 CEST] <prelude2004c> what if i wanted to put multiple channels on the same GPU
[00:50:58 CEST] <prelude2004c> not sure if there is a param for that
[01:12:53 CEST] <electronrotoscop> I'm trying to split a source file to two output files with different filters, and that's working fine, but if I set a -t flag for one it seems to be active for the second one too. Is that expected behavior?
[01:13:07 CEST] <electronrotoscop> I'm using something like ffmpeg.exe -ss 00:01:10 -i "infile" -filter_complex "[0:0]first filter filter, split [out1][in2];[in2]second filter[out2];[0:1][0:2][0:3][0:4][0:5][0:6]amerge=inputs=6[out3]" -map "[out1]" -t 00:46:11 -c:v mpeg2video -parameters -an "short_outfile.m2v" -map "[out2]" -c:v mpeg2video -parameters -an "long_outfile.m2v" -map "[out3]"  -t 00:46:11 -vn -c:a ac3 -ar 48k -b:a 448k "audio_outfile.ac3"
[01:44:32 CEST] <prelude2004c> hey .. can anyone help.. with ffmpeg. and libnvenc .. how do i set it all to just 1 GPU
[02:02:11 CEST] <tumagonx> when lto enabled, I expect ffmpeg become smaller but it's 1mb larger. Anyone experience the same? (gcc 4.7 here)
[02:59:49 CEST] <sideburns> Hello, I'm trying to use ffplay in a shell script on Fedora Linux, but can't see how to configure it correctly.  I know how to suppress any video, but I can't find out how to get rid of text output.  (This script may run when nobody's logged in.)  Suggestions?
[03:07:10 CEST] <sideburns> Just in case the first part of my question didn't go through, the program I'm working with is ffplay.
[03:09:11 CEST] <c_14> sideburns: -loglevel quiet ?
[03:15:21 CEST] <sideburns> That still shows the running stats; will that be a problem if it's run using at when there's nobody logged in?
[03:17:41 CEST] <c_14> running stats? It doesn't show anything for me when I try.
[03:18:28 CEST] <sideburns> Here's my command line, with the file itself left off.
[03:18:35 CEST] <sideburns> ffplay -nodisp -loglevel quiet
[03:19:13 CEST] <sideburns> The next line showed this: fd=   0 aq=    2KB vq=    0KB sq=    0B f=0/0
[03:20:19 CEST] <sideburns> If that just ends up in /dev/null, it's not important.  I'm just being cautious.
[03:21:30 CEST] <c_14> `ffplay -nodisp -loglevel quiet filename' doesn't display anything for me
[03:21:33 CEST] <c_14> What version are you running?
[03:22:31 CEST] <sideburns> 1.2.11 on Fedora 19.  (Can't upgrade because of hardware issues.)
[03:24:54 CEST] <c_14> aah, that's probably the issue. How about just appending 2>/dev/null ? (or 2&>/dev/null)
[03:27:00 CEST] <sideburns> No change.  Is there any difference between the two forms?  ( only tried 2>)
[03:29:02 CEST] <c_14> yes
[03:29:07 CEST] <c_14> try 2&>
[03:31:18 CEST] <sideburns> Both 2& and 2&> give me the sound of silence.
[03:37:45 CEST] <sideburns> Is there a different playback program that might work better?
[03:39:20 CEST] <c_14> Define work better, what are you trying to do?
[03:40:05 CEST] <sideburns> I'm trying to write a shell script that will use at to set an alarm for a specific time, even if I'm not logged in.
[03:40:37 CEST] <sideburns> Right now, I'm using an alarm applet that's part of my session; if I'm not logged in, it doesn't run.
[03:40:59 CEST] <sideburns> That'd be OK, but I've got a hardware issue, and my box sometimes reboots itself.
[03:41:28 CEST] <sideburns> If it does that late at night, I'm not logged in when it's time to wake up, and the alarm doesn't play.
[03:42:36 CEST] <sideburns> I want to use a program that doesn't care if there's a gui or terminal so that the alarm goes off even if my box has rebooted.
[03:43:35 CEST] <c_14> Sounds like a job for cron or at (not sure if at survives reboots though)
[03:44:29 CEST] <sideburns> AFAIK, at survives reboots, which is why I picked it.  I'm not using cron, because it's only needed now and then; not every day.
[03:46:51 CEST] <sideburns> You can specify month-day-year on the command line for at, so I presume it survives rebooting.
[03:47:31 CEST] <c_14> I'm pretty sure ffplay doesn't require a controlling terminal, but you can always use ffmpeg directly with `ffmpeg -i file -f alsa default' (also works with -f pulse), or use a media player like mpv (which iirc also works without a controlling terminal)
[03:50:35 CEST] <sideburns> Which do you recommend?  I have ffmpeg and can easily install mpv.
[03:52:06 CEST] <c_14> Depends. If you only want to play one file I'd stick with ffmpeg just because I know all the options for that. If you want to play multiple files or something like a playlist, use mpv
[03:52:22 CEST] <sideburns> Just tried ffmpeg, and got a whole bunch of text output.
[03:52:34 CEST] <c_14> -loglevel quiet -hide_banner
[03:53:23 CEST] <c_14> Though actually, it looks like -loglevel quiet already hides the banner
[03:54:33 CEST] <sideburns> That got me silence.
[03:54:52 CEST] <sideburns> OK!  Leaving off the hide_banner worked.  Thanx!
[07:30:20 CEST] <prajj> hello good people
[07:30:46 CEST] <prajj> i'm currently facing this issue, made a stack overflow post about it. check it out maybe?
[07:30:46 CEST] <prajj> http://stackoverflow.com/questions/29341161/encode-h264-video-using-ffmpeg-library-memory-issues
[09:37:01 CEST] <haasn> Is there a way to get information about the video source's (per-block) motion prediction vectors out of libavcodec?
[09:44:32 CEST] <haasn> ah, https://ffmpeg.org/pipermail/ffmpeg-devel/2014-August/160938.html is helpful
[09:56:42 CEST] <haasn> Hmm. I guess with the way it's setup, it's basically giving me a discrete list of vectors?
[09:57:07 CEST] <haasn> What I really want is something resembling a continuous surface (the resolution is irrelevant, but it should map 1:1 to the image)
[11:50:03 CEST] <Ders> I'm trying to use -pix_fmt rgba for the output of my file. I noticed that apparently not every codec supports this. WHich one should I use? I tried some I heard about but they don't accept it as well
[11:58:11 CEST] <relaxed> Ders: you could use -v:c rawvideo
[11:58:19 CEST] <relaxed> er, -c:v
[11:58:32 CEST] <Ders> Let me try that one
[11:58:39 CEST] <relaxed> the output will be huge
[11:59:37 CEST] <Ders> Yeah. It gives errors constantly: buffer underflow...
[11:59:49 CEST] <Ders> I do need it to be as small as possible though
[12:03:11 CEST] <relaxed> would bgra work?
[12:03:45 CEST] <relaxed> why do you need rgba?
[12:10:40 CEST] <Ders> it's the output from opengl
[12:10:50 CEST] <Ders> I can change it bgra but I don't know how that would help
[12:11:49 CEST] <relaxed> try -c:v ffvhuff -pix_fmt bgra
[12:12:11 CEST] <relaxed> I didn't see any codecs that support rgba
[12:17:42 CEST] <Ders> thanks. I'll try that
[12:20:26 CEST] <Ders> lots and lots of bufferunderflow happening.
[12:29:17 CEST] <kvz> What's the difference between `-sn -dn` and `-map 0 -map -0:d -map 0:s`? I would like to use `-sn -dn`, because with `map`, I sometimes get 2 video streams in my output file and I only 1 "best" video & 1 "best" audio stream; ffmpeg's default when not using `map`.   However, when I use `-sn -dn`, I get the error "Codec AVOption flags () specified for output file #0 (./test-%05d.ts) has not been used for any stream. The most likely reason is either w
[12:29:18 CEST] <kvz> 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.". Just changing it to the `map` arguments converts the video succesfully (but potentially having 2 video streams if the input file had that)
[12:38:28 CEST] <seasc> Someone able to explain me the difference between lan-streaming and internet-streaming? As in, do i need to do some preparations before i can stream my desktop to a friend of mine? Or can i just use his ip as target?
[12:40:51 CEST] <spaam> higher delay on internet-streaming and you might need to stream in a lower bitrate. :D
[12:41:53 CEST] <seasc> I waited 2 minutes, with a 75mbps flatrate internet, that enough for full hd res?
[12:51:19 CEST] <relaxed> kvz: the former maps everthing but the data stream, -map 0:s wouldn't be needed
[12:51:39 CEST] <relaxed> unless you meant -map -0:s
[12:51:52 CEST] <kvz> yes, sorry
[12:51:59 CEST] <kvz> I'll write up a snipept
[12:52:07 CEST] <relaxed> what do you want in the output stream?
[12:53:32 CEST] <seasc> spaam, was that saying for me?
[12:54:58 CEST] <kvz> relaxed: Here's the script & output: https://gist.github.com/kvz/05b01ba5a13eb91de53e. I would like: 1 video stream. 1 audio stream. No matter what 'weird' input file I throw at it.
[12:55:05 CEST] <sfan5> seasc: when streaming to someone your upload bandwidth matters
[12:56:02 CEST] <kvz> I used map before, but ran into problems when an input file had 2 video streams. I learned from the ffmpeg docs that "By default, ffmpeg includes only one stream of each type (video, audio, subtitle) present in the input files and adds them to each output file. It picks the "best" of each based upon the following criteria: for video, it is the stream with the highest resolution, for audio, it is the stream with the most channels, for subtitles, it 
[12:56:02 CEST] <kvz> the first subtitle stream. In the case where several streams of the same type rate equally, the stream with the lowest index is chosen."
[12:56:06 CEST] <seasc> sfan5, thats about 2-3 mbs (at absolute minimium, if not more around 10mbs), telling to use a maxrate of 2560kb
[12:56:38 CEST] <relaxed> kvz: "ffmpeg -i input -map 0:v:0 -map 0:a:0 ..." would give you the first audio and video stream of the input.
[12:56:46 CEST] <seasc> sfan5, friend has the some connection, also do i keep using 'udp:' because httpp fails
[12:56:49 CEST] <kvz> That made me think I wanted that default behavior, and I could not rely on `map`. So I opted for `-sn -dn` but that threw errors
[12:57:40 CEST] <kvz> relaxed: Will try that, but it seems I would miss out on FFmpeg's alhorthm to select the "best" stream
[12:58:01 CEST] <sfan5> seasc: HTTP probably needs a http server to serve the stream
[12:58:07 CEST] <seasc> kvz, you could try -map 0:v -map 0:a to use all the audio/video streams available, otherwise you need to specify which stream it shall use.
[12:58:57 CEST] <seasc> kvz, err.. sry was already said. But saying... you cannot tell ffmpeg to use stream id 3, and be confused it uses stream 3 rather than id 4 which is better quality.
[12:59:18 CEST] <seasc> sfan5, so udp IS the way to go?
[13:00:03 CEST] <sfan5> seasc: yep, tcp is also an option but not very suited for live streaming
[13:01:09 CEST] <seasc> sfan5, thank you, one last question.. by http server, would it suffice to have just httpd running? Or would i change/install some mods?
[13:01:47 CEST] <kvz> seasc: relaxed: Docs tell me (https://ffmpeg.org/ffmpeg.html#Stream-selection) that FFmpeg should be able to select the best audio & video stream - it should do so by default even. That is also the behavior that I desire.
[13:02:43 CEST] <kvz> seasc: relaxed: This tells me I should use -sn -dn to ditch the other streams, but it errors out, compared to a `map` that would effictively do the same
[13:03:28 CEST] <seasc> kvz, this only accounts for streams with the same identifier, such as 'eng', '0x80' or alike. if the streams identifier (dont know proper term) is invalid/unset (und) this wont work.
[13:03:39 CEST] <sfan5> seasc: you'd probably tell your httpd to proxy example.com/stream to an instance of ffserver listening on localhost
[13:03:45 CEST] <seasc> kvz, using -sn tells ffmpeg to use No Sound at all... likewiese for -vn
[13:04:14 CEST] <sfan5> isn't -sn no subs and -an no audio?
[13:04:19 CEST] <seasc> sfan5, thank you, i'll keep trying with udp first then :)
[13:04:25 CEST] <seasc> err.. right
[13:04:30 CEST] <seasc> was ment to be -an ^^
[13:05:55 CEST] <kvz> Ok I didn't realize this only works for identified streams, that's helpful thanks seasc. So it seems the best way to achieve what I want is relaxed's approach with `-map 0:v:0 -map 0:a:0` - and hoping (or do scripting to figure out) that the first streams, are also the best.
[13:06:51 CEST] <seasc> kvz if you script on linux, i'd have something for you: http://ffmpeg.gusari.org/viewtopic.php?f=25&t=2020&sid=91406d224b87ec027a9e3fdbd42e92fb
[13:07:46 CEST] <sfan5> kvz: alternatively you could let your script read the output of ffprobe and find the best stream
[13:10:47 CEST] <kvz> seasc: vhs looks very nice! we're already deeply invested into ffmpeg however. e.g. we let customers tweak its paramters.
[13:11:05 CEST] <kvz> sfan5: seems like there's no pretty way out of this :)
[13:11:20 CEST] <kvz> seasc: sfan5: relaxed: thanks for the help
[13:11:44 CEST] <seasc> vhs is just a wrapper around ffmpeg, passing -A (as in Advanced) and you can still edit the generate commandline 'tempplate' before it gets executed
[13:12:00 CEST] <seasc> translating issue: deeply invested into ffmpeg, means?
[13:12:24 CEST] <kvz> seasc: ok i'll give it a more throrough look, thanks.
[13:13:59 CEST] <kvz> seasc: not a native english speaker myself so i might have gotten it wrong :) but i meant: we have a lot of code & usecases depending directly on ffmpeg. it would be hard to change the tooling underneath that and not break backwards compatibilty, especially since we open up the ffmpeg api to customers who so desire.
[13:14:51 CEST] <kvz> one thing still bothers me though. Why would the `-sn -dn` approach even fail in the first place https://gist.github.com/kvz/05b01ba5a13eb91de53e
[13:19:14 CEST] <seasc> kvz, is your input audio only?
[13:21:12 CEST] <seasc> kvz, nevermind..
[13:23:49 CEST] <seasc> ok just seeing *ts and *mpegts beeing used quite often, seems i need to 'write a container' for those two in vhs. ($HOME/.config/vhs/container)
[13:23:53 CEST] Action: seasc afk
[13:39:36 CEST] <makhzi> Hi, I'm struggling with "rc buffer underflow". Input stream is mpegTS, received from udp, output is also mpegTS. I need to transcode from h264 to mpeg2video. Here is my command line: http://pastebin.com/HwTscHT2 . Anyone knows how to tune parameters to get rid of this problem?
[13:44:57 CEST] <makhzi> I had also a problem with properly generated mpegTS stream directly from ffmpeg so, i'm piping output stream to VLC and then send it to multicast.
[14:53:30 CEST] <Peace-> arm static ffmpeg link to downalod ?
[14:53:35 CEST] <Peace-> download*
[15:55:00 CEST] <seasc> streaming, mpegts, are there any codec recomendations?... as i understand its basicly just mpeg, but i've seen the use of other codecs for this 'container'.. google just brings a list of h264 vs mpeg videos, or mpeg2 vs mpeg4 comparisions, and alike.
[15:55:16 CEST] <seasc> likewise question for *.ts
[16:23:34 CEST] <Mavrik> seasc, mpegts can hold alot of different codes
[16:23:48 CEST] <Mavrik> you choose the best one your target devices can play ;)
[16:24:08 CEST] <Mavrik> mostly that's H.264/AAC
[16:24:42 CEST] <seasc> Ty Mavrik
[16:28:39 CEST] <seasc> Just installed x265 and run ffmpeg -encoders | grep 265, but nothing was found, also not for "hevc".
[16:30:12 CEST] <Mavrik> you'll have to recompile ffmpeg to link to that x264.
[16:30:15 CEST] <Mavrik> *x265
[16:30:25 CEST] <Mavrik> also, why the heck would you want to use that
[16:30:33 CEST] <seasc> experimenting? :)
[16:31:11 CEST] <seasc> Seeing what bitrates are suggested for 4k++, i assume a new codec makes sense, and i do hope 265 is as good as 264 was when it 'came out'
[16:32:08 CEST] <Mavrik> huh.
[16:32:12 CEST] <Mavrik> It's nowhere near.
[16:32:24 CEST] <Mavrik> It's about 10-30x slower for worse quality
[16:32:29 CEST] <seasc> oh
[16:32:57 CEST] <seasc> so you already tried?
[16:33:17 CEST] <Mavrik> x265 is still in active development, it's not even in "experimental" phase yet
[16:35:10 CEST] <Mavrik> it's very new ;)
[16:36:41 CEST] <__jack__> x265 is ready for many purpose
[16:37:02 CEST] <__jack__> def. not still experimental, can be used in production
[16:37:44 CEST] <flying_toaster> hello there
[16:38:04 CEST] <flying_toaster> I have a strange behaviour using ffmpeg...
[16:38:55 CEST] <flying_toaster> I am using ffmpeg to decode a h264 stream : ffmpeg -i stream.h264  -f rawvideo -pix_fmt yuv420p out.yuv
[16:39:18 CEST] <flying_toaster> my ffmpeg stream has exactly 60 frames; but ffmpeg produces 61 frames
[16:39:31 CEST] <flying_toaster> the first frame is duplicated
[16:40:32 CEST] <flying_toaster> using git bisect, it seems that the commit that introduces this new frame is exactly this one : c5092025
[16:41:00 CEST] <flying_toaster> commited by michaelni, I guess
[16:42:11 CEST] <flying_toaster> I do not know a lot about ffmpeg internals, so I can't tell if this is a bug or not
[16:44:27 CEST] <seasc> __jack__, me confused now, Mavrik said no, you say its ready for many things... what is it ready for?
[16:53:01 CEST] <flying_toaster> ok, it seems adding -vsync passthrough solves my problem
[16:55:59 CEST] <__jack__> seasc: you can encode file & stream, from 4k to dvdrip, with a great quality & size reduction
[16:57:11 CEST] <seasc> __jack__, as in great quality while size reduction?
[16:57:56 CEST] <seasc> it could be understood like: great quality reduction and size reduction
[16:59:46 CEST] <__jack__> haha, no, great quality with small file size
[17:08:31 CEST] <seasc> Ty __jack__ i'll give the ffmpeg rebuild a shot then
[18:16:57 CEST] <baran> how to differ yuv420p from yuv420p(tv) in ffmpeg
[18:57:19 CEST] <elgatov> hi guys, im trying to strip some files from audio streams and just let english and spanish, is there a way to check audio stream by language?
[18:58:54 CEST] <c_14> -map m:lang:en -map m:lang:es should probably do it
[18:59:04 CEST] <c_14> You have to make sure the metadata tags are correct though.
[18:59:26 CEST] <c_14> https://ffmpeg.org/ffmpeg.html#Stream-specifiers-1
[19:03:28 CEST] <elgatov> c_14: thanks a lot
[21:15:23 CEST] <ChocolateArmpits> Trying to encode VP8 but the bitrate doesn't get lower than 56kbps, do I have to change some setting? minrate and maxrate doesn't influence this
[21:15:51 CEST] <c_14> You _want_ the bitrate lower than 56kbps?
[21:20:32 CEST] <ChocolateArmpits> Yes
[21:22:07 CEST] <ChocolateArmpits> My source is low res (96x128), is posterized to about 2 colors (white and black) and plays at 10fps. 56kbps is overkill
[21:26:28 CEST] <ChocolateArmpits> ok down to 30kbps
[22:52:25 CEST] <tlhiv_laptop> i am having a VERY difficult time streaming audio from my microphone to a format that HTML5 <audio> tags can understand ... i would appreciate any insight into this
[22:55:48 CEST] <seasc> tlhiv_laptop, try ogg
[22:56:03 CEST] <seasc> gn8
[23:12:43 CEST] <ac_slater_> when muxing video, do I have to explicitly set bitrate? What happens if I dont?
[23:14:06 CEST] <Mavrik> ac_slater_, that's a very nonspecific question :)
[23:16:12 CEST] <kepstin-laptop> ac_slater_: the answer is "different things depending what you have the video encoder set to"
[23:16:26 CEST] <ac_slater_> kepstin-laptop: ;) sorry.
[23:16:35 CEST] <ac_slater_> I'm only doing container muxing
[23:16:48 CEST] <ac_slater_> ie - I encode separately and I muxing into MPEGTS
[23:16:49 CEST] <kepstin-laptop> if you are using -c:v copy, then the bitrate is ignored.
[23:17:08 CEST] <ac_slater_> kepstin-laptop: ... I should have mentioned... I'm using libavformat, et al .. not the ffmpeg binary
[23:17:18 CEST] <iive> bitrate is option for the encoder
[23:17:29 CEST] <iive> because this is where the loss is controlled
[23:17:45 CEST] <ac_slater_> that makes sense. I guess I missed that
[23:17:48 CEST] <ac_slater_> thanks mate!
[23:18:29 CEST] <Mavrik> ac_slater_, the bitrate for the muxer controls some minor stuff
[23:18:36 CEST] <Mavrik> that depends on the container/muxer involved
[23:18:40 CEST] <Mavrik> mostly it's not relevant
[23:18:45 CEST] <ac_slater_> I'm only using mpegts
[23:18:53 CEST] <ac_slater_> I havent looked at the muxer code to see if it really uses it
[23:19:04 CEST] <Mavrik> it doesn't afaik
[23:19:37 CEST] <ac_slater_> awesome. I'm still confused on some other muxing things ... I guess I'll report back with questions if I have them
[23:25:16 CEST] <ac_slater_> Mavrik: actually I think I've asked you this before but I can't find it in the logs... If I write two streams to a container (mpegts), and neither have b-frames, should I have to do `av_interleaved_write()` ?>
[23:25:18 CEST] <ac_slater_> ? *
[00:00:00 CEST] --- Wed Apr  1 2015


More information about the Ffmpeg-devel-irc mailing list