[Ffmpeg-devel-irc] ffmpeg.log.20180707

[burek]

[00:04:22 CEST] <karlll> It seems to be up again :)
[00:23:04 CEST] <ChocolateArmpits> building ffmpeg is pretty easy with msys2+mingw
[11:28:17 CEST] <analogical> It's it possible to use FFmpeg to decompress ALAC (Apple Lossless Audio Codec) files to WAV ??
[11:28:37 CEST] <analogical> ** Is it possible to use FFmpeg to decompress ALAC (Apple Lossless Audio Codec) files to WAV ??
[11:32:11 CEST] <JEEB> yes, there's been an ALAC decoder for ages
[11:32:11 CEST] <furq> sure
[11:32:15 CEST] <furq> -i foo.m4a bar.wav
[11:32:41 CEST] <analogical> that simple huh?
[11:33:22 CEST] <furq> or preferably -i foo.m4a -compression_level 12 bar.flac
[11:33:47 CEST] <analogical> no I don't want to create a flac :)
[11:34:22 CEST] <JEEB> &34
[11:35:42 CEST] <analogical> furq, can I use you example -i foo.m4a bar.wav to process many files in the same directory somehow?
[11:36:04 CEST] <analogical> *many files simultaniously
[11:39:35 CEST] <furq> not with ffmpeg by itself
[11:39:44 CEST] <furq> but yes with your shell
[12:34:35 CEST] <baschdel> Hello
[12:52:26 CEST] <baschdel> In Kdenlive there's a render option called 'MPEG4-ASP/MP3 (DivX compatible)' is there an equivalent for ffmpeg?
[12:53:29 CEST] <furq> mpeg4
[12:53:35 CEST] <furq> you really shouldn't be using that in this day and age though
[12:53:49 CEST] <furq> unless you need to support some ancient dvd player or something
[12:55:22 CEST] <baschdel> thank you
[12:56:03 CEST] <baschdel> And yes I'm recoding it for an 'ancient' TV (6 years old)
[12:57:51 CEST] <furq> you can also use libxvid if your ffmpeg was built with it
[12:57:54 CEST] <furq> that ought to be higher quality
[12:58:05 CEST] <furq> a tv that recent probably supports h264 though
[12:58:09 CEST] <furq> you should use that if you can
[15:13:21 CEST] <Yukkuri> hi, having a list of N `file 'path/to/file'` entries, how to select 0'th video stream and 1'st audio stream of each file when using -f concat -i index.list?
[15:13:55 CEST] <Yukkuri> -map seems to be affected by offset of each file in the list
[15:14:11 CEST] <Yukkuri> so -map 0:v:0 -map 0:a:1 would work only for first entry
[15:15:20 CEST] <Yukkuri> maybe i can perform selection in this list file?
[15:16:54 CEST] <DHE> -map is an app option. concat is a library feature that produces a single "file" illusion to the app.
[15:17:55 CEST] <Yukkuri> so how to combine -f concat and stream selection?
[15:18:14 CEST] <DHE> I'm looking at the docs and I'm not really sure...
[15:20:53 CEST] <DHE> https://ffmpeg.org/ffmpeg-formats.html#concat  "All files must have the same streams (same codecs, same time base, etc.). " so I suspect this is not supported.
[15:21:28 CEST] <Yukkuri> hau. okay, thank you, guess i can perform two-step conversion
[15:21:39 CEST] <Yukkuri> first to normalize inputs, then concat them
[15:49:22 CEST] <Hello71> https://trac.ffmpeg.org/wiki/Concatenate#differentcodec
[15:49:44 CEST] <Hello71> why does the frame rate have to be the same
[15:49:52 CEST] <Hello71> or is that just better
[15:53:38 CEST] <ariyasu> it won't cat if settings are different
[15:54:38 CEST] <furq> you can change those things midway through a file but only in certain containers iirc
[15:54:50 CEST] <furq> and obviously only in certain players
[15:55:33 CEST] <furq> just not in ffmpeg
[16:19:41 CEST] <Hello71> ah, true, not all containers support changing fps
[16:19:51 CEST] <Hello71> pretty sure matroska does though
[16:20:41 CEST] <Hello71> "MKV does not store a frame rate in its container" huh.
[16:21:53 CEST] <JEEB> it just has PTS
[16:21:56 CEST] <furq> iirc with h264 it only works with annex b
[16:21:57 CEST] <JEEB> the presentation time stamp
[16:22:08 CEST] <furq> not sure though, i've only ever looked into this with m2v
[16:22:55 CEST] <zyme> Question, anyone know if there's an easy way to leave the video stream on a mkv as is, and add a new audio stream (2-channel) track to the video?
[16:23:06 CEST] <furq> zyme: -c:v copy
[16:23:14 CEST] <furq> and probably -map
[16:23:20 CEST] <Hello71> or even -c copy
[16:23:28 CEST] <Hello71> if the audio is already how you want it
[16:23:33 CEST] <furq> right
[16:27:05 CEST] <zyme> I find different programs that are non-specific in their methods are either very bad or very good at converting 7.1 or 5.1 to 2.0, when I had a receiver with native support and 7.1 speakers the voices would be pretty much only from the center front channel, and really kind of needs to be boosted when converted to understand what people are saying..
[16:28:01 CEST] <furq> if you want to add a mixdown track then -i foo.mkv -map 0:v map 0:a:0 -map 0:a:0 -c:a:0 copy -c:a:1 aac -ac:1 2 bar.mkv
[17:28:49 CEST] <styler2go> Hey everyone. I am trying to concat soem video files and convert the audi codec but if i am using -acodec mp3 i get the message: https://p.styler2go.de/6257852
[18:02:59 CEST] <wiak> can someone point me to the documentation on how i can link a external mkv file with internal tracks of another file? i want to make a play "video track 2, audio track 2 of a matroska file that has video track 1, audio track 2 as default"
[18:03:35 CEST] <wiak> styler2go do you get the same issue if you use flac instead of mp3?
[18:05:05 CEST] <wiak> styler2go seems the audio is corrupt
[18:06:39 CEST] <styler2go> wiak, it works when not using any codec (just concat withouth -acodec)
[19:29:32 CEST] <lvlscape_2> if im streamign over rtmp, and cant use h264, how to begin choosing fromt he 300+ remaining codecs for video?
[19:29:58 CEST] <furq> you don't
[19:30:10 CEST] <furq> rtmp only supports h264 and the shitty flash codecs
[19:30:36 CEST] <furq> and idk if most rtmp servers bother to support vp6 or whatever
[19:30:59 CEST] <furq> i'm pretty sure nginx-rtmp doesn't
[19:31:24 CEST] <furq> anything that republishes to hls won't support anything but h264
[19:39:27 CEST] <wiak> i tried i gave up lol
[22:15:25 CEST] <wyseguy> hi all
[22:15:51 CEST] <wyseguy> anyone to lend a hand on my ffmpeg command? I have messed with it quite a bit but cant get a better stream quality
[22:17:28 CEST] <DHE> that's not how IRC works. say what's wrong, include pastebin of commandline and output if applicable (commandline only in channel is fine) and wait for a reply from someone
[22:18:54 CEST] <wyseguy> DHE yes, curious if anyone was around. will post in a min
[22:24:34 CEST] <wyseguy> I ran this command and was getting a live stream that was super pixelated: fmpeg -i "udp://10.5.5.100:8554?fifo_size=10000" -f flv "rtmp://a.rtmp.youtube.com/live2/xxxx-xxxx-xxxx-xxxx  But then after getting some insight, I tried running this command and the live stream was much better but youtube was showing the stream was having some issues: ffmpeg -f mpegts -i "udp://10.5.5.100:8554?fifo_size=10000" -c:v copy -f flv "rtmp://a.rtmp.youtube.com
[22:24:34 CEST] <wyseguy> /live2/xxxx-xxxx-xxxx-xxxx Any ideas on how to improve this? My setup is a MacBook Air connecting to a gopro hero6 black over wifi and that stream is being pulled into camera suite, the MacBook Air is usb tethered to a verizon jetpack for internet. Terminal view: https://pastebin.com/xWeFT6VG
[23:27:02 CEST] <jleclanche> What's the correct way to merge a shorter wav with a longer audioless mp4? My mp4 is ~30 min, audio is ~15 min, and I just want the last 15 min to be silent.
[23:27:16 CEST] <jleclanche> I used `ffmpeg -i out.mp4 -i out2.wav -c:v copy -c:a aac -strict experimental final.mp4` which works, but past 15 min the video gets laggy and mpv complains a lot
[23:27:36 CEST] <jleclanche> [lavf] Too many packets in the demuxer packet queues:
[23:27:36 CEST] <jleclanche> [lavf]   video/0: 4491 packets, 419481216 bytes
[23:27:36 CEST] <jleclanche> [lavf]   audio/1: 0 packets, 0 bytes
[00:00:01 CEST] --- Sun Jul  8 2018