January 2020 - Ffmpeg-devel-irc

ffmpeg-devel.log.20200105
by burek 06 Jan '20

06 Jan '20

[00:00:00 CET] --- Sun Jan 5 2020

1 0

ffmpeg.log.20200105
by burek 06 Jan '20

06 Jan '20

[00:47:50 CET] <BtbN> Can the loudnorm filter be used on a live stream, or is it useless without doing two passes? [00:51:39 CET] <kepstin> it works ok on live stuff, but it's more likely to have to use the non-linear normalization mode. [00:52:08 CET] <kepstin> er, i guess the linear mode only works in two-pass, so it has to use the non-linear mode [00:53:28 CET] <BtbN> My intent is to apply it to a playlist of various unrelated videos, and prevent the usual "Holy shit why is this so loud suddenly" when it switches [00:53:47 CET] <kepstin> yeah, it should work fine for that. [00:58:32 CET] <noudle> im trying to get hwdec to work using mpv on a rpi3 and have compiled ffmpeg with '--enable-mmal --enable-omx --enable-omx-rpi' but mpv fails to use hwdec erroring out with the following, [00:58:37 CET] <noudle> [ffmpeg/video] h264_mmal: MMAL error 2 on control port [00:58:41 CET] <noudle> [ffmpeg/video] h264_mmal: Did not get output frame from MMAL. [01:02:29 CET] <furq> noudle: error 2 is ENOSPC [01:02:44 CET] <furq> maybe you're running out of gpu memory [01:03:13 CET] <noudle> uhm [01:03:23 CET] <noudle> sec let me try something [01:03:52 CET] <furq> also make sure nothing else is using mmal [01:04:24 CET] <noudle> furq: nothing else is running [01:04:34 CET] <noudle> but it can be im using not the full gpu mem available [01:04:40 CET] <noudle> need to test that sec [01:13:21 CET] <noudle> furq: any idea how much mem should be needed? [01:13:39 CET] <furq> no idea [01:13:54 CET] <furq> test with 512 i guess [01:14:31 CET] <noudle> 64 resulted in the same, testing 128 now [01:15:08 CET] <noudle> oh wow 128 do work [01:15:13 CET] <noudle> ty for the hint! [02:55:57 CET] <Hello71> isn't 2 ENOENT [02:59:18 CET] <furq> https://github.com/raspberrypi/firmware/blob/master/opt/vc/include/interfac… [03:20:18 CET] <Hello71> right [03:20:20 CET] <Hello71> but like [03:21:22 CET] <nicolas17> ENOENT the errno is 2 [03:43:09 CET] <noudle> whatever increasing the gpu mem did the trick [03:43:15 CET] <noudle> night [04:48:38 CET] <Soni> can ffmpeg encode DSD yet [04:51:11 CET] <nicolas17> what is DSD? [04:52:12 CET] <Soni> effectively, a fancy name for delta-sigma modulation [04:53:01 CET] <furq> an uncompressed audio format that there's no need to care about [04:53:14 CET] <nicolas17> looks like it can only decode [04:54:25 CET] <furq> dacs use delta-sigma internally so i think the point of it is to avoid the conversion from pcm [04:54:32 CET] <furq> so it doesn't really make sense to have an encoder in ffmpeg [04:55:04 CET] <furq> either way it doesn't sound any better unless you work for sony [04:55:34 CET] <Soni> furq: all you care is about what you hear, but you forget that there are programmers here [04:56:04 CET] <Soni> furq: ffmpeg is a highly modular synth and a DSD encoder would be a nice addition to its arse-anal [04:56:26 CET] <nicolas17> make one :) [04:57:02 CET] <furq> some dacs, rather [04:58:12 CET] <Soni> furq: besides, I need something that's fast enough to use in real-time with a 155MHz 8-bit DAC [04:59:14 CET] <Soni> (as it turns out, PCM + dithering is... not) [08:31:11 CET] <MoziM> if a video with no audio is 1.9G then shouldn't all of its extracted keyframes be roughly equivalent to 1.9G as well? [10:17:31 CET] <yang> Hello, what is the best way to capture Analog VCR video? I have a script that is capturing to .avi (windows format), but I would prefer to capture to mp4 instead - ffmpeg -f v4l2 -standard PAL -thread_queue_size 512 -i /dev/video0 -f alsa -thread_queue_size 512 -i hw:2,0 -vcodec libx264 -preset superfast -crf 25 -s 720x576 -r 25 -aspect 4:3 -acodec libmp3lame -b:a 128k -channels 2 -ar 48000 out.avi [10:17:43 CET] <yang> Could I get an example for capturing .mp4 ? [10:19:07 CET] <durandal_1707> .mp4 is awful for capturing [10:19:26 CET] <yang> oh, which one is good and doesnt use much space ? .webm ? [10:19:47 CET] <yang> .avi stream is about 200 MB / 10 min. [10:20:34 CET] <yang> But since I'd prefer to play streams on Linux, .avi isn't optimal [10:46:44 CET] <yang> Do you think .avi is ok to capture video ? [11:39:11 CET] <MoziM> if a video is DTS_FGT what sort of audio codec should i remux it with to be as lossless as possible? [15:03:11 CET] <utack> i did not quite manage to figure it out from the manual, but is there a way to use ffprobe to find the maximum number of consecutive bframes used in a stream? [15:05:14 CET] <DHE> yes, but you have to actaully go and count them yourself. maybe some kind of awk script on the output with -show_frames [15:07:10 CET] <utack> DHE i am trying to piece that together, how would one tell show_frames to only output pict_type? [15:08:47 CET] <cehoyos> The option loglevel verbose is supposed to tell you the number of reference frames, it may help [15:10:39 CET] <utack> i have added a grep now around it to find the number of bframes, not elegant but works :D [16:37:23 CET] <yang> I am interested to record with Matroska open codec [16:37:29 CET] <yang> The current ffmpeg command line for .avi is like this: [16:37:31 CET] <yang> ffmpeg -f v4l2 -standard PAL -thread_queue_size 512 -i /dev/video0 -f alsa -thread_queue_size 512 -i hw:2,0 -vcodec libx264 -preset superfast -crf 25 -s 720x576 -r 25 -aspect 4:3 -acodec libmp3lame -b:a 128k -channels 2 -ar 48000 out.avi [16:37:36 CET] <yang> How do I make the command optimal to be used with matroska ? Just replace out.avi with out.mkv ? [16:38:10 CET] <yang> And secondary, is there a program/tool that reveals which format is the video file being recorded with ? [16:38:25 CET] <cehoyos> What is "Matroska open codec" [16:38:26 CET] <cehoyos> ? [16:38:49 CET] <cehoyos> If you don't want to write an avi file but an mkv file, use out.mkv instead of out.mkv [16:38:57 CET] <cehoyos> This will not change the "codec" but the container [16:39:45 CET] <cehoyos> FFmpeg reveals which format (and codec) a multimedia file has, yes: Just use "ffmpeg -i input" [16:39:56 CET] <yang> ok [16:39:58 CET] <yang> thank you [16:49:10 CET] <yang> How could I replace the command to record audio in an open format like FLAC ? [16:49:23 CET] <yang> or OPUS [16:51:34 CET] <cehoyos> Replace libmp3lame with the codec of your choice, note that mp3 and opus have a very different behaviour than flac and that opus may not play with older players. [16:53:21 CET] <MoziM> i have a video file that's an mkv and contains 1 english audio stream, if i want to add a secondary audio stream (russian dub) do i have to create a new file or can a second audio stream simply be "inserted" into an existing an mkv [16:53:27 CET] <MoziM> *existing mkv file [16:53:30 CET] <MoziM> ? [16:53:37 CET] <yang> Yes, can i do "-acodec flac" ? [16:58:02 CET] <cehoyos> mkvmerge may be able to "insert" the audio track (I don't know), FFmpeg always creates a new file, you can choose if you copy the audio track (you should always do that if possible) or if you want to re-encode [17:01:08 CET] <MoziM> cehoyos: mkvtoolnix includes mkvmerge correct? [17:02:51 CET] <yang> cehoyos: is it necessary to use -b:a with flac? [17:03:27 CET] <cehoyos> You cannot use b:a with flac but as said, you should only use flac if you have a very specific reason, use mp3 if you have no cluse. [17:03:31 CET] <cehoyos> You cannot use b:a with flac but as said, you should only use flac if you have a very specific reason, use mp3 if you have no clue. [17:05:16 CET] <yang> cehoyos: I'd like to have audio and video codecs in an open format [17:05:34 CET] <cehoyos> Why? [17:06:09 CET] <cehoyos> And why do you think that mp3 which has a very old, known specification, is not old? [17:06:13 CET] <cehoyos> And why do you think that mp3 which has a very old, known specification, is not open? [17:06:19 CET] <yang> >> "hat should work, yes, though I would strongly recommend not to use MP3 nowadays. There are tons of better audio codecs out there (better in the sense of same quality [17:06:24 CET] <yang> for much lower bit rate). Use Opus for state-of-the-art or AAC if the resulting file must be playable on hardware devices (which often don't support Opus)." [17:12:04 CET] <cehoyos> Why do you ask if you know better? If your player supports opus, opus is the best codec. AAC is not more free than mp3 but has better quality. Best quality is not to re-encode. [17:12:16 CET] <cehoyos> Why do you ask if you know better? If your player supports opus, opus is the best codec. AAC is not more free than mp3 but has better quality. Best quality is always not to re-encode. [17:15:53 CET] <yang> does flac require re-encoding when capturing stream ? [17:16:02 CET] <yang> and mp3 doesn't ? [17:16:56 CET] <cehoyos> When you capture, it depends on what format your capturing driver provides, most likely uncompressed audio that you should compress. [17:17:05 CET] <cehoyos> If your player supports opus, opus is the best choice. [17:17:18 CET] <cehoyos> If your player does not support opus but aac, aac is the second best choice. [17:17:25 CET] <cehoyos> If your player supports only mp3, use mp3 [17:18:11 CET] <cehoyos> It is likely that you will not be able to hear a difference between the three as long as the bitrate is high enough. [17:18:23 CET] <cehoyos> 128k is not a low bitrate [17:42:38 CET] <kepstin> 128 might be kinda low for mp3, depending on content. it'd suggest going a *bit* higher there. [18:17:10 CET] <yang> does flac require 128k setting [18:19:45 CET] <furq> you can't set the bitrate of flac [18:19:53 CET] <yang> ok [18:53:13 CET] <kilobyte_ch> I try to optimize an mp4 for streaming with ffmpeg. The mp4 has 6 audio streams, 1 video stream and 1 subtitle stream. I use the following parameters `-movflags faststart -c:v copy -c:a copy -c:s copy -c:d copy -c:t copy`. I get an optimized mp4 out of it with all streams but I don't see the subtitle text in vlc on the optimized mp4. The entry is there but there is no text. On the source file it is working. [18:55:44 CET] <furq> is the source mp4 [18:58:47 CET] <kilobyte_ch> furq: yes, source is mp4 [19:00:09 CET] <kilobyte_ch> Not sure if it maybe messes the subtitles up with the faststart movflag [19:03:28 CET] <kilobyte_ch> Ok, even with just `-c:v copy -c:a copy -c:s copy -c:d copy -c:t copy -map 0` the subtitles are broken afterwards. [19:04:29 CET] <ritsuka> does the subtitles play in another player? [19:05:27 CET] <kilobyte_ch> ritsuka: On the source the subtitles work in vlc and gnome video player, on the output file the subtitles don't work on both players [19:08:07 CET] <pink_mist> I didn't know mp4 actually supported subtitle streams [19:08:43 CET] <ritsuka> kilobyte_ch: what's the format of the subtitles stream? [19:08:57 CET] <kilobyte_ch> ritsuka: vlc says tx3g [19:10:47 CET] <ritsuka> I can confirm there is something wrong, just tried to remux a mp4 [19:11:01 CET] <ritsuka> in quicktime player and mpv the subtitles work [19:11:06 CET] <ritsuka> but not in vlc [19:12:19 CET] <ritsuka> oh nv, they actually work [19:12:37 CET] <ritsuka> it's just vlc being stupid when you seek around in the file [19:15:57 CET] <kilobyte_ch> hmm, in the gnome video player the subtitles are coming up for a fraction of a second (output file) [20:02:01 CET] <pink_mist> wouldn't surprise me if that was a bug in the gnome video player [20:12:58 CET] <q3cpma> Hello, does anyone know if a lossless DSD -> PCM conversion is possible? I'd like to use sox afterward to get it to CD "quality". [20:14:46 CET] <durandal_1707> sox is dead [20:22:41 CET] <q3cpma> Doesn't really matter. I finally settled for DSDIFF -> pcm_s32le -> pcm_s16le [21:03:35 CET] <void09> trying to join 5 pieces of a video and at the point where they are joined, the playback stutters a little. any idea how to fix this ? I used ffmpeg -f concat -safe 1 -i 1.txt -c copy output.mkv [21:04:08 CET] <void09> I cut the pieces with ffmpeg also, and the video part is frame accurate [21:07:46 CET] <autofsckk> Hello, im trying to put subtitles into a movie, i found this ffmpeg -i video.avi -vf subtitles=subtitle.srt out.avi but the output video has a very bad quality, how can i keep the same size/quality of the output video? or at least a decent one? [21:11:17 CET] <pink_mist> keeping it the same is impossible since you're outputting to avi and thus need to burn in the subs by reencoding the video, so you'll never get the same quality unless you want to have absolutely huge files [21:11:49 CET] <autofsckk> pink_mist: i see, but is there a way to make it better? [21:12:20 CET] <pink_mist> I mean, if you don't need avi, you could just remux it to mkv instead, then the subs don't need to be burned into the video frames [21:12:39 CET] <pink_mist> but if you need avi, you'll need to figure out some codec and codec options that are acceptable to you [21:13:06 CET] <autofsckk> pink_mist: i would like to put the movie in a usb so i can see it from a tv [21:13:20 CET] <pink_mist> that doesn't really answer anything [21:13:26 CET] <pink_mist> some tvs support mkv with subs [21:13:29 CET] <pink_mist> others don't [21:15:10 CET] <pink_mist> other tvs support putting an avi file next to a srt file, and it will do the right thing [21:15:13 CET] <pink_mist> others don't [21:16:14 CET] <autofsckk> this one does not support mkv files, im not sure if it reads srt files, but i think it doesnt [21:16:56 CET] <pink_mist> then we're back to: 21:12 <pink_mist> but if you need avi, you'll need to figure out some codec and codec options that are acceptable to you [21:17:32 CET] <pink_mist> it's a matter of trial and error, though I'm sure there are blog posts out there that can help you figure such things out faster [00:00:02 CET] --- Mon Jan 6 2020

1 0

ffmpeg-devel.log.20200104
by burek 05 Jan '20

05 Jan '20

[00:00:00 CET] --- Sat Jan 4 2020

1 0

ffmpeg.log.20200104
by burek 05 Jan '20

05 Jan '20

[04:03:48 CET] <void09> anyone still around here? [04:04:44 CET] <void09> I discovered ffmpeg duplicates a frame when i cut two pieces, first: -ss start -to a_time, second: -ss a_time -to end [04:05:08 CET] <void09> last frame of first pieces is the first frame of second piece [04:06:48 CET] <void09> this happens in 2 cases. I checked the input and there's only one frame there, so it should not be in both pieces ? [04:07:04 CET] <void09> by cutting i mean encoding to lossless, not with -copy [04:35:43 CET] <while> void09: sometimes I find the ffmpeg has bugs that are easier to workaround then to open a ticket and wait for a patch [04:36:05 CET] <void09> I cannot work around this [04:36:16 CET] <while> what is your end objective? [04:36:21 CET] <void09> as I need it to work at scale, not just for this one example [04:36:48 CET] <void09> end objective is to encode pieces of a video, source being http, and begin/end timestamps given [04:37:18 CET] <void09> and I cannot have duplicated frames when I rejoin the encoded pieces, for obvious reasons [04:37:35 CET] <while> can you remove the extra frame from your encoded video? [04:37:41 CET] <void09> of course not [04:38:07 CET] <void09> you can only remove any frames at will from non-keyframe based video stream [04:38:12 CET] <while> could you encode into a bunch of .bmp files, and remove and offset them as to remove the offending frame? [04:38:44 CET] <void09> while: i could workaround it a few times I guess, but as I said, I need this method to work at scale [04:38:46 CET] <kepstin> void09: can you use the segment muxer instead for your use case? [04:38:53 CET] <void09> I cannot spend all day scanning for duplicae frames [04:39:28 CET] <while> void09: you can automate that, I think ffmpeg has something for it, I haven't looked recently though [04:39:47 CET] <while> like a way to 'rate' how similar 2 frames/pictures are [04:39:54 CET] <void09> kepstin: I'd rather have it working with the -ss -to instead. not sure how the segment muxer works [04:40:11 CET] <void09> I was more interested in why this problem occurs [04:40:19 CET] <kepstin> void09: also, are you using -to as an input option or output option? [04:40:32 CET] <void09> kepstin: I tried both ways, both of them produce the same result [04:40:44 CET] <void09> input is an mkv file made from a bluray [04:40:53 CET] <void09> bluray remux [04:41:23 CET] <kepstin> so, is the purpose to do a distributed video encode? [04:41:33 CET] <void09> so I am not sure if this is a ffmpeg bug or some weird edge case, the input having some particular properties that confuses ffmpeg [04:41:41 CET] <void09> kepstin: yes [04:42:04 CET] <kepstin> i'd suggest using the segment muxer then. it'll require more disk space since you'll have to store the segments, but it should work much cleaner [04:42:23 CET] <void09> uh oh [04:42:41 CET] <void09> why doesn't it work cleanly as it is ? [04:43:02 CET] <kepstin> not sure, would have to look into timestamp debugging output from ffmpeg running on your input file [04:43:13 CET] <void09> I'd rather not use the segment muxer as that would mean i have to change the logic of the system I set up, and it will take a long time [04:43:15 CET] <kepstin> but segment muxer is guaranteed to do a clean split between segments [04:43:26 CET] <void09> well that's the thing [04:43:31 CET] <void09> I cannot have segments [04:43:42 CET] <void09> because the keyframes are random in the input [04:43:53 CET] <void09> i mean, they have no corelation with the scenes [04:44:06 CET] <kepstin> oh, right, now i remember you talking about that before [04:44:07 CET] <void09> so i'd have to calculate offsets for the segments anyway [04:44:14 CET] <void09> oh yes, I have too much free time :D [04:44:16 CET] <void09> still at it [04:44:27 CET] <kepstin> and my response remains "don't worry, it doesn't make a big enough difference" [04:44:35 CET] <void09> I wrote some software to split segments and organize workers.. [04:44:46 CET] <void09> what doesn't make a difference? [04:45:11 CET] <kepstin> assuming the segments are big enough, missing scene boundaries makes barely any difference [04:45:24 CET] <void09> well you don't understand [04:45:37 CET] <void09> every 100kb matters :) [04:45:56 CET] <void09> and it's not about missing scene boudnaries [04:46:06 CET] <void09> well, I guess it is [04:46:19 CET] <kepstin> (besides, unless you're doing gop splitting using the same method as the encoder you're using, it'll be mismatched anyways) [04:46:36 CET] <void09> nope, I cannot have that. I am doing the distributed encoding in the first place jut because I want to be able to encode at maximum settings [04:46:55 CET] <void09> kepstin: I am using pyscenedetect, and it works pretty much flawlessly [04:47:47 CET] <void09> I use a threshold such that I make sure it does not split a scene in two segments. even though it might merge a few scenes as one [04:48:10 CET] <kepstin> anyways, if you're doing all this work, why not use ffmpeg apis directly? then you can look at the unmodified input timestamps and pick exactly when to stop :/ [04:48:22 CET] <void09> kepstin: I'm not a real programmer :\ [04:48:50 CET] <void09> and there's no ffmpeg library for my language.. well there is one but kind of oudated. headers [04:49:21 CET] <void09> so yeah, I'd rather invest time bugging ffmpeg devs to fix this [04:49:56 CET] <kepstin> well, you're gonna need to provide debug output from the ffmpeg command with timestamp information, as i said. [04:50:11 CET] <void09> not sure how to do that [04:51:15 CET] <void09> are you a dev kepstin, i forgot ? [04:51:21 CET] <kepstin> you could also try using the trim filter to set the end point, it might work differently. try the -t option, see if it behaves more like you want. play around with the different -copyts options, etc. [04:51:52 CET] <kepstin> last time you asked that, i said "define dev", and it became a whole long thing. i don't feel like doing that again. [04:52:02 CET] <void09> I completely forgot [04:52:36 CET] <void09> would a ffmpeg command with the remote file be of any use to you, to take a look ? [04:53:13 CET] <kepstin> i have written patches that have gotten merged into ffmpeg, and I volunteer to help people out here when i have time, but this isn't the main thing i do. [04:53:59 CET] <void09> it would be cool if you had a few minutes to take a look at it. it's only ~20 seconds or so of video [04:56:14 CET] <void09> or at least tell me how do i get that debug info to file a bug report [04:57:01 CET] <kepstin> please make commands that demonstrates the issue with two short clips that have a shared frame (preferably only a few seconds if possible), and run the ffmpeg commands with the options "-v debug -debug_ts 1" - it'll generate a lot of output, you'll want to save it to a file, or consider using the -report option. [04:57:49 CET] <kepstin> and then, on your own, try a few of the other things i mentioned. [05:00:03 CET] <void09> maybe the problem will not occur if i cut it in segments [19:37:31 CET] <cpplearner> Guys, I want to get an error report by feeding my HDTV record to ffprobe. IMO, just by making ffprobe reading all the bytes would do the trick, but I'm not sure the exact options to use. [19:38:50 CET] <cpplearner> How can I make ffprobe persists until the end? [19:40:27 CET] <cehoyos> (Complete, uncut console output missing.) If with "HDTV" record you mean a transport stream, then errors are unavoidable. [19:40:46 CET] <cpplearner> Yeah, I mean a transport stream. [19:41:29 CET] <cpplearner> So, what I want to do is just to see if my recording contains some errors. For that purpose, Windows has a great application called mpeg2repair, but linux has no counterpart. [19:42:17 CET] <cpplearner> So, I'll just try to use ffprobe for that, and I know that when reading a transport stream ff-family by default reports errors. [19:42:23 CET] <cehoyos> FFmpeg is not a repair tool for transport streams [19:42:33 CET] <cpplearner> No, I don't mean to repair. [19:42:44 CET] <furq> just use ffmpeg with null output [19:42:52 CET] <furq> unless you actually need the ffprobe output at the end [19:43:50 CET] <cpplearner> mpeg2repair can do report & repair erros, but the formar is the one that I'm interested in. [19:44:07 CET] <cpplearner> So, with null option ffmpeg reads to the end? [19:44:15 CET] <furq> ffmpeg always reads to the end [19:44:29 CET] <furq> ffprobe will do it with -count_frames but there's no advantage to using it over ffmpeg [19:45:00 CET] <cpplearner> Just a quick quesiton: what's the core difference between ffprobe & ffmpeg? [19:45:25 CET] <furq> different options [19:46:31 CET] <cpplearner> Okay, Thanks for the solution. I'll try with null options. [19:46:37 CET] <cpplearner> Have a nice day. [19:46:51 CET] <furq> -i foo -f null - [20:17:17 CET] <AiNA_TE> does -b:v get ignored when using crf ? [20:17:26 CET] <Mavrik> pretty much [20:18:17 CET] <furq> depends on the encoder [20:18:31 CET] <furq> with libvpx -b:v is the bitrate cap [20:18:50 CET] <furq> it's just ignored with x264/5 though [20:18:58 CET] <AiNA_TE> im using x264 and want to limit max bitrate [20:19:06 CET] <furq> you want bufsize and maxrate then [20:19:12 CET] <AiNA_TE> ok, so i need to fall back to 2pasS? [20:19:14 CET] <AiNA_TE> oh [20:19:19 CET] <AiNA_TE> ok i will try that [21:49:45 CET] <noudle> following https://trac.ffmpeg.org/wiki/ExtractSubtitles how do i know what to pass to -txt_format ? [21:51:33 CET] <noudle> and how can i specify which subtitles i want to extract? [22:07:33 CET] <noudle> nvm [00:00:00 CET] --- Sun Jan 5 2020

1 0

ffmpeg-devel.log.20200103
by burek 04 Jan '20

04 Jan '20

[00:00:00 CET] --- Fri Jan 3 2020

1 0

ffmpeg.log.20200103
by burek 04 Jan '20

04 Jan '20

[05:49:10 CET] <while> hi [15:27:13 CET] <LFSVeteran> with ffmpeg -f s32le -ac 2 -ar <samplerate> -i http://ip:port/stream -acodec pcm_s32le output.wav [15:27:25 CET] <LFSVeteran> do I still need init_resampler in transcode_aac.c ? [15:40:08 CET] <DHE> if you need to convert between formats, yes. same number of channels and same format helps, but that still leaves the sample rate in the air. [15:40:34 CET] <LFSVeteran> samplerate stays the same [15:41:15 CET] <furq> yeah you don't need it [17:16:34 CET] <Mavrik> hmm, how's libx265 these days? performance-wise? [19:41:49 CET] <digital-_-mystik> hello.. I would like to convert a DTS-HD Master Audio codec to AC3 in an mkv container. when using this command, `ffmpeg i input.mkv -c:v copy -c:a ac3 -b:a 320k /path/to/usb/output.mkv`, does the original file remain untouched and it just copies/converts to a new container while leaving the video the same? [19:42:20 CET] <DHE> basically yes [19:43:05 CET] <digital-_-mystik> would you say that is the fastest way to go about it without demuxing, converting the audio, remuxing, and then putting on a usb drive? [19:43:35 CET] <DHE> that's probably fine. it's a single command that does it all in one shot [19:43:43 CET] <digital-_-mystik> sounds goods [19:43:48 CET] <digital-_-mystik> good* [19:43:50 CET] <digital-_-mystik> thanks [20:07:39 CET] <barhom> nvidia gpu's have "GPU Utilization", "nvenc Utilization" and "nvdec Utilization". I am assuming if I run yadif_cuda filter that would not really touch the nvenc/nvdec util but increase the gpu util. Am I thinking wrong? [20:25:25 CET] <Mavrik> That's the theory, yes. [20:27:26 CET] <furq> barhom: nvdec has its own deinterlacing stuff that's supposedly pretty good [20:27:53 CET] <barhom> furq: Is that different from yadif_cuda? [20:28:01 CET] <furq> yeah [20:28:07 CET] <furq> -deint 2 -i foo iirc [00:00:00 CET] --- Sat Jan 4 2020

1 0

ffmpeg-devel.log.20200102
by burek 03 Jan '20

03 Jan '20

[00:00:00 CET] --- Thu Jan 2 2020

1 0

ffmpeg.log.20200102
by burek 03 Jan '20

03 Jan '20

[00:20:47 CET] <barg> I have a 'video' I made it in premiere pro by importing 150MB of 128Kbps audio and adding some images. When I export it, it exports to 7GB. Let's say I want to use ffmpeg to make it a more reasonable size.. what settings should I use with ffmpeg? [00:21:45 CET] <barg> presumably the video codec is ridiculous overkill.. is there a video codec and settings associated with it that would be appropriate for images.. like I show a different image every 5min in the video [00:23:54 CET] <DHE> you probably want to crank the framerate way down, use qp or crf mode, and a fairly wide keyframe interval [00:24:55 CET] <barg> how do I do that? [00:25:06 CET] <barg> like what kind of command? [00:25:12 CET] <barg> given blahfile.mp4 [00:25:15 CET] <DHE> keyframe interval is -g measured in frames [00:25:18 CET] <cehoyos> You could start with "ffmpeg -i input -pix_fmt yuv420p out.mov" and check if you like the result [00:25:41 CET] <barg> so is mp4 a bad idea? [00:26:00 CET] <DHE> mp4 is a container. it says little about what codecs and parameters are inside it [00:26:33 CET] <barg> is mp4 with libx264 a bad idea? [00:26:55 CET] <cehoyos> You could start with "ffmpeg -i input -pix_fmt yuv420p out.mp4 and check if you like the result [00:26:59 CET] <DHE> no, that's what I would expect you to use. it's just a matter of configuring it appropriately [00:27:22 CET] <barg> i'm wary though of not specifying -vcodec as i've heard that the default codecs are bad? [00:27:37 CET] <cehoyos> You could start with "ffmpeg -i input -pix_fmt yuv420p out.mp4 and check if you like the result [00:28:56 CET] <furq> the default codecs for mp4 are x264 and aac [00:29:04 CET] <furq> and the default settings for both of those are ok [00:29:48 CET] <furq> using defaults is fine as long as you know what they are [07:44:16 CET] <Casper> Hi there. Gentoo here, trying to emerge ffmpeg-9999 (aka git version) and I get "ERROR: ffnvcodec requested but not found" does anyone have an idea how to fix it? Or should I ask in #ffmpeg-devel? [07:45:02 CET] <BtbN> Install/Update ffnvcodec. [07:46:39 CET] <Casper> which appear to not be something available o.O [07:47:01 CET] <BtbN> Well, that'd explain why it's not found. [07:47:52 CET] <Casper> or it's included in something else... [07:48:07 CET] <BtbN> Why would it be included in something else? [07:50:07 CET] <Casper> dunnot [07:50:25 CET] <while> BtbN: In my experience it is common to find git projects where some dependencies are not mentioned, in such cases, you have to go through the painful tasks of finding what they are and install them [07:50:51 CET] <BtbN> ffnvcodec is not a new thing. It exists since at least two releases. [07:51:34 CET] <BtbN> And looking at https://github.com/gentoo/gentoo/tree/master/media-libs/nv-codec-headers, they are simply out of date on their end. [07:52:00 CET] <Casper> ah so it's nv-codec-headers ... [07:53:30 CET] <BtbN> Whatever they decided to call it [07:55:33 CET] <Casper> I'll check that tomorrow, thanks will see if I downgrade instead... I forgot why I went with the git one.. [07:56:28 CET] <BtbN> Their versions of the SDK are seriously old, and will need updated sooner or later no matter what. [07:58:22 CET] <Casper> yeah... I think within a few years I may have to be forced to move to another distro... gentoo have some packages that need maintainers... [07:59:01 CET] <BtbN> Just file a bug and it will probably be bumped... [08:00:21 CET] <Casper> I just downgraded to the 'stable' package [08:03:53 CET] <Casper> ffmpeg-4.1.3 there! yay [08:05:23 CET] <Casper> thanks [08:05:46 CET] <Casper> atleast now I can do a full upgrade without being painfull... [08:06:11 CET] <Casper> time for bed, thanks again [09:33:14 CET] <hiihiii> hi [09:34:13 CET] <hiihiii> I have a bunch of video inputs with still frames at the end. I'm trying to use the select filter to trim off these ends [09:34:24 CET] <hiihiii> any ideas are appreciated [09:38:00 CET] <BtbN> Why not just use -t? [09:39:22 CET] <hiihiii> I have lots of files and I don't know when the still for each file happens [09:39:44 CET] <BtbN> How would the select filter help you then? [09:39:47 CET] <hiihiii> just curious if I can play with the select filter to achive this [09:39:57 CET] <hiihiii> currently re-readin docs [09:40:12 CET] <BtbN> You will need to know the precise duration no matter what [09:40:53 CET] <BtbN> In theory a filter that discards the last N frames is possible, by buffering N frames and discarding the buffer on EOF, but I don't think one exists yet. [09:41:24 CET] <hiihiii> idk maybe something to tell ffmpeg "hey we've incounterd a long sequence of still frames maybe we should stop" [09:41:40 CET] <hiihiii> oh ok [09:42:22 CET] <BtbN> Does mpdecimate work for your case? [09:45:11 CET] <hiihiii> maybe let me check [09:52:09 CET] <hiihiii> I'm not supposed to use mpdecimate with -vsync 0 [09:52:29 CET] <hiihiii> ? [10:26:50 CET] <hiihiii> select works just fine [10:26:55 CET] <hiihiii> thx [10:27:09 CET] <BtbN> How did you manage to select the last N frames? [10:27:15 CET] <BtbN> Without knowing the duration? [10:27:42 CET] <hiihiii> I didn't select the last N frames [10:27:48 CET] <hiihiii> per say [10:28:05 CET] <hiihiii> the ending of each video is a still scene [10:28:17 CET] <hiihiii> just told it to detect that and discard it [10:29:05 CET] <hiihiii> select=gt(scene,0.003) [10:29:56 CET] <hiihiii> although, for some reason I can play the output in my media player [10:30:08 CET] <hiihiii> ffplay plays it just well [10:30:55 CET] <hiihiii> I'm on a machine with no codecs though maybe that's why [10:31:32 CET] <hiihiii> I've also tried playing it on my phone. same problem [10:31:41 CET] <hiihiii> have to use vlc [10:32:36 CET] <hiihiii> is there something I need to run with select I'm not aware of [10:32:38 CET] <hiihiii> ? [10:33:10 CET] <hiihiii> my command is simple [10:33:45 CET] <hiihiii> -i input -vsync 0 -an -c:v libx264 -crf 30 -vf "select=gt(scene\,0.003)" output [10:34:34 CET] <hiihiii> without the video filter everything plays fine ?! [10:52:23 CET] <BtbN> the filter probably makes a mess of the timestamps [10:52:33 CET] <BtbN> try regenerating them afterwards [10:52:37 CET] <BtbN> setpts or something [10:53:10 CET] <BtbN> Or try getting rid of -vsync [11:54:41 CET] <shin-ts> Why can't we add plug-ins after compilation of FFmpeg binary. [11:56:05 CET] <pink_mist> what are plug-ins? [11:56:14 CET] <pink_mist> in terms of ffmpeg [11:57:00 CET] <leot> Hello folks! [11:57:12 CET] <shin-ts> pink_mist plug-ins or format plug-ins. [11:58:25 CET] <leot> AFAICS it seems that CVE-2019-13390 was fixed in ffmpeg-4.2.1 can it be added to https://www.ffmpeg.org/security.html the corresponding commit should be: https://git.ffmpeg.org/gitweb/ffmpeg.git/commit/1529dfb73a5157dcb8762051ec4… [11:58:47 CET] <shin-ts> Suppose you compiled an FFmpeg binary libaom-av1 but you wish to use rav1e in FFmpeg, you've to recompile entire binary yourself, again. [11:59:17 CET] <shin-ts> With plug-ins you can easily add those and work with it without thinking. [12:01:07 CET] <ritsuka> probably because it's fast enough to recompile the whole thing, and you don't have to worry about api changes breaking compiled plugins [12:04:31 CET] <shin-ts> The '-q' scale ranges 2 to 31. Is there any reason of why range is not 0 to 100? [12:05:36 CET] <shin-ts> the scale should be like (0 = worst quality, 100 = best quality) [12:26:27 CET] <shin-ts> The WebP decoder implementation is incomplete, it can't iterate animated WebP frames. [12:26:37 CET] <durandal_1707> nobody cares [12:27:43 CET] <shin-ts> what? [15:40:33 CET] <devinheitmueller> Good morning. Ive got a weird case where I have a filtergraph that includes crop followed by scale. However when I use the eval=frame mode for crop the scaler becomes confused. It looks like the crop filter arbitrarily changes the AVFilterLink when the crop properties change, and the scaler doesnt see the change because it expects the AVFilterLinks to never change after the pipeline is setup. [15:40:48 CET] <devinheitmueller> Is there any policy with regards to changing AVFilterLinks after the graph is constructed and running? [15:41:53 CET] <devinheitmueller> In principle we should probably be calling config_props() again to inform the downstream filter of the change, but there is no way for a filter to signal that back to the caller. [16:28:04 CET] <devinheitmueller> I can either modify the crop filter to not change its output links properties, so that it uses local context data to drive its operations instead of the AVFilterLink properties. That will cause the scaler to see the frame properties dont match the filter link and reconfigure. Alternatively I can change the scale filter to watch for changes to its filter link (i.e. cache the last values in the filter contex [16:28:04 CET] <devinheitmueller> and compare against the filterlink properties). [16:28:12 CET] <devinheitmueller> ^ Both of these solutions are pretty ugly. [17:02:38 CET] <ufk> hi! :) i'm trying to stream a single picture endlessly to a v4l2 device without success. i don't see errors but it stops immidiately. this is what I have so far: ffmpeg -loop 1 -framerate 10 -i binary-file.png -t 30 -pix_fmt yuv420p -f v4l2 /dev/video0 [17:02:41 CET] <ufk> what am I missing ? :) [17:13:25 CET] <ufk> ok removed -t 30 and it works [17:19:43 CET] <Hello71> >endlessly [17:19:46 CET] <Hello71> >limit time [17:19:48 CET] <Hello71> hm... [19:14:13 CET] <voices> hey, can i reduce a flicker (shutterspeed/lighting mismatch) with ffmpeg? here's an example of the technique https://youtu.be/YyQjFuFYchg?t=91 [19:21:23 CET] <voices> they do a 50% opacity overlay of the same clip and offset it by one frame or so [19:27:39 CET] <furq> voices: there's a deflicker filter but it doesn't do that [19:28:45 CET] <furq> you could probably replicate that specific technique with the overlay filter though [20:10:17 CET] <durandal11707> voices: tmix filter? [21:01:44 CET] <irgendwer4711> hi, how to get a corrected time code when trim copy a movie? [21:40:29 CET] <irgendwer4711> anyone? [22:17:02 CET] <tester0815> Hi guys, I'm new at IRC, never tried this before, hope I'll do fine [22:17:13 CET] <tester0815> I'm sure it's a stupid question, but I find no answer for this. [22:17:22 CET] <tester0815> So would be proud if you can help me: [22:17:33 CET] <tester0815> Tried to rewrap a .mts to mp4 using ffmpeg, with this command: ffmpeg -i test.mts -c copy -a copy test.mp4 [22:17:45 CET] <tester0815> Works fine, but after checking bitrate, the bitrate of mts ist something around 105MB/s and mp4 ist 25MB/s [22:18:25 CET] <tester0815> Is that correct? [22:19:35 CET] <tester0815> Thanks for your answers. [22:20:06 CET] <klaxa> not sure -a copy is a thing? [22:20:38 CET] <klaxa> but if you use -c copy the video streams of both files should be bit identical (save for format filters) [22:21:42 CET] <furq> tester0815: you probably want -map 0 [22:21:50 CET] <furq> otherwise it'll only copy one of each type of stream [22:23:15 CET] <tester0815> thanks, I'll try [22:31:19 CET] <irgendwer4711> how to get a corrected time code when trim copy a movie? [22:42:37 CET] <tester0815> furq I tried your suggestion but same video bitrate as before, mts 105Mbit/s, mp4 25 Mbit/s [22:43:20 CET] <devinheitmueller> Could also be the mts is a constant bitrate stream with null pid padding, which would all be stripped in creating the MP4. Hence the video and audio are copied exactly but the resulting file is much smaller. [22:44:00 CET] <devinheitmueller> @tester0815 Pastebin the output that shows the input, mapping, and ouput definition when running the ffmpeg command line. [22:44:04 CET] <furq> is the mp4 actually a quarter of the size [22:44:20 CET] <tester0815> the file size is nearly same [22:44:32 CET] <furq> well yeah the bitrates can't possibly be that much different then [22:44:41 CET] <furq> so whatever tool you're using is probably misreporting it [22:45:01 CET] <tester0815> ffmpeg -i PRIVATE/AVCHD/BDMV/STREAM/00000.MTS -map 0 -vcodec copy -acodec copy -sn test2.mp4 [22:45:01 CET] <tester0815> ffmpeg version 4.2.2 Copyright (c) 2000-2019 the FFmpeg developers [22:45:01 CET] <tester0815> built with Apple clang version 11.0.0 (clang-1100.0.33.16) [22:45:03 CET] <tester0815> configuration: --prefix=/usr/local/Cellar/ffmpeg/4.2.2 --enable-shared --enable-pthreads --enable-version3 --enable-avresample --cc=clang --host-cflags='-I/Library/Java/JavaVirtualMachines/adoptopenjdk-13.0.1.jdk/Contents/Home/include -I/Library/Java/JavaVirtualMachines/adoptopenjdk-13.0.1.jdk/Contents/Home/include/darwin -fno-stack-check' --host-ldflags= --enable-ffplay --enable-gnutls --enable-gpl --enable-libaom --enable-libbluray - [22:45:05 CET] <tester0815> -enable-libmp3lame --enable-libopus --enable-librubberband --enable-libsnappy --enable-libtesseract --enable-libtheora --enable-libvidstab --enable-libvorbis --enable-libvpx --enable-libx264 --enable-libx265 --enable-libxvid --enable-lzma --enable-libfontconfig --enable-libfreetype --enable-frei0r --enable-libass --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenjpeg --enable-librtmp --enable-libspeex --enable-libsoxr [22:45:07 CET] <tester0815> --enable-videotoolbox --disable-libjack --disable-indev=jack [22:45:09 CET] <tester0815> libavutil 56. 31.100 / 56. 31.100 [22:45:11 CET] <tester0815> libavcodec 58. 54.100 / 58. 54.100 [22:45:13 CET] <tester0815> libavformat 58. 29.100 / 58. 29.100 [22:45:15 CET] <tester0815> libavdevice 58. 8.100 / 58. 8.100 [22:45:17 CET] <tester0815> libavfilter 7. 57.100 / 7. 57.100 [22:45:19 CET] <tester0815> libavresample 4. 0. 0 / 4. 0. 0 [22:45:21 CET] <tester0815> libswscale 5. 5.100 / 5. 5.100 [22:45:22 CET] <devinheitmueller> tester0815: STOP [22:45:23 CET] <furq> well this is going to take a while [22:45:23 CET] <tester0815> libswresample 3. 5.100 / 3. 5.100 [22:45:25 CET] <tester0815> libpostproc 55. 5.100 / 55. 5.100 [22:45:27 CET] <tester0815> Input #0, mpegts, from 'PRIVATE/AVCHD/BDMV/STREAM/00000.MTS': [22:45:29 CET] <tester0815> Duration: 00:02:46.08, start: 1.020000, bitrate: 26768 kb/s [22:45:31 CET] <tester0815> Program 1 [22:45:33 CET] <tester0815> Stream #0:0[0x1011]: Video: h264 (High) (HDPR / 0x52504448), yuv420p(progressive), 1920x1080 [SAR 1:1 DAR 16:9], 50 fps, 50 tbr, 90k tbn, 100 tbc [22:45:35 CET] <devinheitmueller> please do not paste console output directly into the IRC window. Please use pastebin. [22:45:35 CET] <tester0815> Stream #0:1[0x1100]: Audio: ac3 (AC-3 / 0x332D4341), 48000 Hz, 5.1(side), fltp, 448 kb/s [22:45:37 CET] <tester0815> Stream #0:2[0x1200]: Subtitle: hdmv_pgs_subtitle ([144][0][0][0] / 0x0090), 1920x1080 [22:45:39 CET] <tester0815> File 'test2.mp4' already exists. Overwrite ? [y/N] y [22:45:41 CET] <tester0815> [mp4 @ 0x7fa085009c00] track 1: codec frame size is not set [22:45:43 CET] <tester0815> Output #0, mp4, to 'test2.mp4': [22:45:45 CET] <tester0815> Metadata: [22:45:47 CET] <tester0815> encoder : Lavf58.29.100 [22:45:49 CET] <tester0815> Stream #0:0: Video: h264 (High) (avc1 / 0x31637661), yuv420p(progressive), 1920x1080 [SAR 1:1 DAR 16:9], q=2-31, 50 fps, 50 tbr, 90k tbn, 90k tbc [22:45:51 CET] <tester0815> Stream #0:1: Audio: ac3 (AC-3 / 0x332D4341), 48000 Hz, 5.1(side), fltp, 448 kb/s [22:45:53 CET] <tester0815> Stream mapping: [22:45:55 CET] <tester0815> Stream #0:0 -> #0:0 (copy) [22:45:57 CET] <tester0815> Stream #0:1 -> #0:1 (copy) [22:45:59 CET] <tester0815> Press [q] to stop, [?] for help [22:46:01 CET] <tester0815> frame= 8304 fps=6788 q=-1.0 Lsize= 516022kB time=00:02:46.04 bitrate=25458.0kbits/s speed= 136x [22:46:03 CET] <tester0815> video:506815kB audio:9082kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.024099% [22:46:05 CET] <tester0815> oh sorry, [22:46:09 CET] <tester0815> that's too much. [22:46:11 CET] <tester0815> Sorry, for that [22:46:15 CET] <tester0815> newb as its best [22:47:07 CET] <furq> anyway ffmpeg is reporting the input bitrate there as 25mbps [22:47:24 CET] <furq> so probably-mediainfo is misreporting it [22:47:41 CET] <devinheitmueller> Yeah, if the filesize is nearly the same and the output youve provided shows the content isnt being reencoded then its probably a reporting issue. [22:47:46 CET] <devinheitmueller> Yeah, what furq said. :_) [22:48:10 CET] <tester0815> okay, didn't find that message. [22:51:04 CET] <tester0815> Thanks guys, that's what i thought, so it's a losless rewrap. Thanks a lot. [22:51:40 CET] <tester0815> This is the summary of the input file, right: [22:52:18 CET] <tester0815> Input #0, mpegts, from 'PRIVATE/AVCHD/BDMV/STREAM/00000.MTS': Duration: 00:02:46.08, start: 1.020000, bitrate: 26768 kb/s [22:56:32 CET] <tester0815> Kay Guys thanks a lot for your answers and your time, bye cu next time [23:15:55 CET] <LFSVeteran> I want to implement "ffmpeg -f s32le -ac 2 -ar <samplerate> -i http://ip:port/stream -acodec pcm_s32le -ac 2 -ar <samplerate> output.wav" in c [23:16:05 CET] <LFSVeteran> but where to start? checked the examples [23:25:57 CET] <furq> LFSVeteran: probably transcode_aac.c [23:26:17 CET] <furq> you don't need -ac or -ar as output options, they'll be kept the same as the input by default [23:26:31 CET] <LFSVeteran> ah ok [23:26:35 CET] <furq> so you don't need to filter or resample [23:27:12 CET] <LFSVeteran> little confused about the streaming part [23:27:23 CET] <LFSVeteran> checking transcode_aac.c once more [23:32:33 CET] <furq> it's more or less the same but you'll need av_find_input_format("s32le") and an avdictionary containing the sample rate and channels as the third and fourth args to avformat_open_input [00:00:00 CET] --- Fri Jan 3 2020

1 0

ffmpeg-devel.log.20200101
by burek 02 Jan '20

02 Jan '20

[00:00:00 CET] --- Wed Jan 1 2020

1 0

ffmpeg.log.20200101
by burek 02 Jan '20

02 Jan '20

[03:10:32 CET] <Media_Thor> Happy New Year FFmpeg-ers! [05:43:46 CET] <while> is there a way to break an audio stream into its composing sine waves, like say, into multiple streams each stream containing only 1 sine wave [06:30:16 CET] <DHE> while: you mean a fourier analysis? or just turning a single stream of multiple channels into many mono streams [06:45:08 CET] <while> yes fourier analysis, and synthesis [12:37:39 CET] <soub> Hi [12:37:57 CET] <soub> Ffmpeg is planning for gsoc 2020 [19:05:35 CET] <AABatteries> hey folks, I'm struggling with chaining filters - a have a pretty simple script that creates a padding and then overlays a watermark over it. It works fine, but now I decided to make sure that the watermark is scaled to 10% of the original video height (instead of hardcoding the value) and it stopped working. The padding is added, but the overlay is missing. Can anyone help me find where I messed up? I'm pretty new to ffmpeg [19:05:47 CET] <AABatteries> here's the code snippet: https://pastebin.com/cGwL35XM [19:06:19 CET] <AABatteries> I think I'm struggling with commas/semicolons [20:18:02 CET] <cousin_luigi> Greetings. [20:18:43 CET] <while> hi [20:18:51 CET] Action: cousin_luigi has a number of MPEG-TS avc/aac to concatenate, but he's failing at it [20:19:09 CET] <cousin_luigi> So far only the audio appears to be copied. [20:19:21 CET] <while> whats your command line look like? [20:19:38 CET] <cousin_luigi> ffmpeg -f concat -i filelist1.txt -c:v copy -c:a copy output.wav [20:19:45 CET] <cousin_luigi> crap [20:20:34 CET] <while> well, with how ffmpeg works, it makes assumptions based on the output file name, in this case its likely going to treat the output as an MS wav file [20:20:52 CET] <cousin_luigi> while: It's working now. [20:21:02 CET] <while> if you changed the output file name to output.wav.mp4 or output.wav.mkv, that may solve that [20:21:07 CET] <cousin_luigi> while: How can I set the language for the audio track? [20:21:18 CET] <while> like subtitles? [20:21:50 CET] <while> I mean, set the language in what capacity [20:22:05 CET] <cousin_luigi> while: For the audio track itself. [20:22:22 CET] <cousin_luigi> Even if there's just one, I like having the language shown in the player. [20:22:54 CET] <while> so in the audio stream headers [20:23:38 CET] <cousin_luigi> -metadata:s:a:0 language=foo is apparently what I need [20:24:06 CET] <while> yes, and I think that takes and ISO 639-2 code of the language, there is probably a table online [20:34:20 CET] <cousin_luigi> Thanks, my problem is apparently solved! [20:59:29 CET] <cousin_luigi> bbl [00:00:00 CET] --- Thu Jan 2 2020

1 0