January 2020 - Ffmpeg-devel-irc

ffmpeg-devel.log.20200130
by burek 31 Jan '20

31 Jan '20

[00:00:00 CET] --- Thu Jan 30 2020

1 0

ffmpeg.log.20200130
by burek 31 Jan '20

31 Jan '20

[00:06:54 CET] <gregorycu> I'm trying to listen for a RTMP stream, and pump out the raw frame data to file descriptors. At the moment, at the start of the stream, it takes 800 ms for data to start to flow. Is there some way to reduce this delay? [00:59:23 CET] <gregorycu> For the record, it's about 40 frames @ 60 fps, or about 650 ms. [01:21:55 CET] <HickHoward> >doesnt ffmpegs ac3 encoder base on the reference ac3 thing from 1990 [01:22:04 CET] <HickHoward> someone on discord said this [01:33:20 CET] <furq> ok [03:09:40 CET] <cehoyos> HickHoward: This sounds unlikely, the ac3 encoder was part of the original commit, Dolby didn't like the code very much, if they had a copyright claim, and didn't make it, that would be hard to explain [03:39:20 CET] <HickHoward> >Dolby didn't like the code very much [03:39:22 CET] <HickHoward> dan [03:39:23 CET] <HickHoward> damn [03:40:07 CET] <HickHoward> anyway i just put that out there to put down some doubts regarding certain codec implementations someone seems to be having [03:40:21 CET] <HickHoward> i mean [03:40:29 CET] <HickHoward> okay not the best way to put it but anyway [03:40:53 CET] <kingsley> I'm happy to report permission to bench mark on a Talos II computer that uses IBM's cool new POWER9v2 cpus. [03:42:03 CET] <kingsley> I've been bench marking how fast the "libvpx" codec can render .webm. [03:43:57 CET] <cehoyos> kingsley: We are not the developers of libvpx, interest here for libvpx performance tests are probably limited... [03:44:32 CET] <cehoyos> HickHoward: I believe the fact that Dolby never tried to have to ac3 code removed from FFmpeg seems (to me) like a good indication that there was no copyright violation. [03:45:03 CET] <cehoyos> And the fact that for a long time, there was only an ac3 encoder, no decoder, also supports the theory that this was not copied code. [03:46:23 CET] <kingsley> Do any encoders besides "libvpx" support multi-threading for faster rendering? [03:49:36 CET] <cehoyos> libx264,libx265,libaom (they are all not developed here) [03:50:21 CET] <cehoyos> "ffmpeg -encoders" provides a list with threading information [03:51:21 CET] <kingsley> cehoyos: Thank you very much! I'll check them out! [03:51:21 CET] <cehoyos> To test FFmpeg, the mpeg4 encoder comes to mind [03:52:33 CET] <cehoyos> But if this is to prove that a (Power) cpu with 256 cores runs faster than an (Intel) cpu with 20 cores, FFmpeg is likely the wrong tool: Video coding is scalable over cores, but only to a (very) limited degree, depending on the resolution [03:52:39 CET] <kingsley> Feel free to run and/or tweak my humble benchmark. It is [03:53:00 CET] <kingsley> $ threads_per_renderer=8 ; time ffmpeg -y -f lavfi -i mandelbrot=s=1920x1080 -loglevel fatal -threads "$threads_per_renderer" -row-mt 1 -frames 50 -c:v libvpx /tmp/bench.webm [03:53:19 CET] <cehoyos> Yes, this tests libvpx performance. [03:53:55 CET] <cehoyos> Note that I suspect mandelbrot has some computational complexity, testsrc2 would be a "faster" input [03:55:08 CET] <kingsley> For what it's worth, my *preliminary* benchmarks of libvpx on a POWER9v2 with 2 CPUs, 22 cores/CPU and 4 threads/core are [03:55:28 CET] <kingsley> One renderer using one thread: 1 frame per second (FPS). [03:55:42 CET] <kingsley> One renderer using 8 threads: 3 FPS. [03:56:08 CET] <kingsley> And 110 renderers using 2 threads each: 60 FPS. [03:56:55 CET] <furq> the -encoders output isn't accurate for codecs with auto threading [03:56:57 CET] <kingsley> So my *preliminary" results suggest 2-8 threads are best. [03:57:15 CET] <furq> like libx264 [03:58:23 CET] <furq> but yeah libvpx generally has poor multithreading [03:59:08 CET] <furq> x264 or something with proper frame threading will do a lot better, although you still get diminishing returns around 20 threads iirc [04:02:08 CET] <kingsley> furq: I researching your comments... [04:02:26 CET] <cehoyos> 20 would be impressive... [04:03:00 CET] <furq> i mean severe diminishing returns [04:03:07 CET] <furq> can't find the benchmark where i saw this though [04:04:35 CET] <furq> if i had 256 cores and a workflow to split, encode and concat, i'd probably start benchmarking from 8 threads per instance [04:18:13 CET] <kingsley> cehoyos: Thank you for suggesting "testsrc2" as input. I'm intrigued by distinguishing between fractal math (Mandelbrot) and rendering. [04:21:49 CET] <kingsley> I particularly like its recognition of ffmpeg's "-threads <n>" option for bench marking how well it parallelizes. [04:25:42 CET] <furq> you should probably benchmark with a non-synthetic clip [04:29:12 CET] <HickHoward> that's quite interesting to hear that in regards to ac3 [04:29:46 CET] <HickHoward> with that being said i believe ac-3 encoding/decoding support holds up fairly well within ffmpeg even today [04:35:22 CET] <furq> you probably shouldn't use ac3 at all any more [04:35:27 CET] <furq> but if you have to then the ffmpeg encoder is pretty good [04:41:44 CET] <kingsley> furq: If you happen to have the time, and are so inclined, please feel free to elaborate on how I might benchmark with a non-synthetic clip. Bonus points for a command line example! [04:49:14 CET] <furq> http://ultravideo.cs.tut.fi/#testsequences [04:49:18 CET] <furq> something like that [05:50:15 CET] <Media_Thor> looking for a method to comapre two audio tracks: original and one with added translation on top of original. Is there any method to differenciate two ? [06:06:34 CET] <EoE> usually when comparing two tracks my strategy is to invert one, and then mix down to a single track in audacity, so any parts that are the same combine to 0, and differences are left, in audacity [06:08:11 CET] <EoE> looks like you could do this in ffmpeg using the https://ffmpeg.org/ffmpeg-filters.html#aeval filter, and then mixing them together [06:09:08 CET] <EoE> the amerge filter should do that [06:09:15 CET] <EoE> see https://trac.ffmpeg.org/wiki/AudioChannelManipulation [12:27:51 CET] <Regor> hi friends how can i play m3u8 or .mpd files on cmus which uses ffmpeg backened ? i didnt find any cmus room so i am asking here [14:09:25 CET] <GenTooMan> Greetings I noticed ffmpeg uses all the available processors by default, is it possible to control the number of processors it uses for threading? like make -j <#> does? [14:09:54 CET] <JEEB> -threads before -i is decoders, after -i is encoders [16:37:33 CET] <FireBurn> Is there a rough guestimate as to when the next release of ffmpeg will be? I'm guessing 4.3 or 5.0 [20:27:59 CET] <barg> does ffmpeg have an option to write back to the same file? [20:28:12 CET] <JEEB> no [20:28:55 CET] <JEEB> ffmpeg.c is not a "patching" application, after all [20:30:00 CET] <Hello71> I mean... you can use the sponge utility [20:30:02 CET] <JEEB> and while in theory removing a file should leave it around until the file handle is closed, it just makes more sense to first output to a separate file name and then handle the renaming otherwise (scripted or otherwise) afterwards [20:30:07 CET] <Hello71> but probably you should use exiftool instead [21:01:52 CET] <barg> ta [21:01:59 CET] <barg> ta=thanks [21:37:53 CET] <giaco> ffmpeg clip is buggy. I've tried both -ss + -to and -ss + -t, but the resulting file has a wrong number of seconds in it [21:39:09 CET] <giaco> ffmpeg -i file.wav -ss 00:20:00 -to 01:20:00 results in a file of 3625 seconds [21:39:41 CET] <giaco> audacity exports that just right [21:52:33 CET] <giaco> using (-c copy) [22:09:42 CET] <durandal_1707> giaco: use filters / atrim instead [22:11:58 CET] <giaco> durandal_1707: I've just found out that is the other way around. Is the duration of the file loaded in audacity that differs from the one detected from ffmpeg [22:12:27 CET] <giaco> durandal_1707: surprisingly 30 seconds are missing but I can't say where: head and tail sounds in place [22:16:44 CET] <giaco> durandal_1707: Filtergraph 'trim=1000:2000' was defined for audio output stream 0:0 but codec copy was selected. [22:20:04 CET] <durandal_1707> drop codec copy [22:20:16 CET] <durandal_1707> it is pointlesss for wav and pcm [22:20:51 CET] <giaco> durandal_1707: not really, I'm failing in reproducing the same file format without it [22:21:19 CET] <durandal_1707> giaco: what codec you use? [22:21:31 CET] <giaco> durandal_1707: ffprobe on the original file is Stream #0:0: Audio: adpcm_ima_wav ([17][0][0][0] / 0x0011), 48000 Hz, stereo, s16p, 384 kb/s [22:22:01 CET] <durandal_1707> that can have any block size, so not really ffmpeg fault [22:22:50 CET] <durandal_1707> besides audacity last time I checked hadnt had codec copy [22:23:31 CET] <giaco> durandal_1707: exactly, I used audacity to check for file duration as I though ffmpeg was in error, but actually it seems the other way around [22:24:00 CET] <giaco> if I open that wav with audacity the total duration differs. Exiftool and ffmpeg agrees on the duration, audacity days -30s [22:24:26 CET] <giaco> no errors/warnings whatsoever displayed [22:30:12 CET] <giaco> durandal_1707: where can I read more about that format? [22:31:18 CET] <durandal_1707> giaco: about wav? [22:36:33 CET] <giaco> durandal_1707: about "that can have any block size". I need something like this su succeed (ffmpeg -i input.wav -filter:a "trim=1000:2000" -c:a adpcm_ima_wav output.wav) [22:36:52 CET] <giaco> it spits "Cannot connect video filter to audio input" [22:37:22 CET] <furq> atrim [22:37:31 CET] <giaco> oh, wait, my fault [22:37:38 CET] <giaco> thanks [22:37:47 CET] <giaco> I just realized it while pasting [00:00:00 CET] --- Fri Jan 31 2020

1 0

ffmpeg-devel.log.20200129
by burek 30 Jan '20

30 Jan '20

[00:00:00 CET] --- Wed Jan 29 2020

1 0

ffmpeg.log.20200129
by burek 30 Jan '20

30 Jan '20

[00:43:44 CET] <cehoyos> Feel free to test if any two streams are synchronized, the command line and console output look sane to me now [00:44:32 CET] <cehoyos> At least there should be no increasing desync now [01:05:59 CET] <nickname123> hi [01:09:50 CET] <plitter> ok, ill do some more testing [01:12:09 CET] <nickname123> i get a list of Fontconfig errors with the static build from https://johnvansickle.com/ffmpeg/ but not with ffmpeg from the Ubuntu repository [01:12:34 CET] <nickname123> is it possible that the static builds don't work on Ubuntu 19.10/19.04? [01:15:44 CET] <nickname123> fontconfig version is "2.13.1-2ubuntu2" is installed on my computer, if that helps [02:19:16 CET] <Conjecture> Is there any codec better than flac for lossless audio compression? [02:20:00 CET] <c3r1c3-Win> Are there any other lossless audio codecs out there? [02:20:20 CET] <Conjecture> I believe there are [02:21:45 CET] <pink_mist> some people like wavpack [02:21:52 CET] <pink_mist> personally I prefer flac [02:24:43 CET] <Conjecture> pink_mist, Didn't you used to be in #Slackware? [02:24:52 CET] <Conjecture> Sorry, ##Slackware [02:25:38 CET] <der_richter> there is also tak and ape [02:25:39 CET] <pink_mist> yes, but I disagreed with how the ops discriminated against some users and did nothing against other users who did much worse things [02:25:49 CET] <pink_mist> so I moved to ##slackware-help [02:27:41 CET] <der_richter> there is also alac [02:28:04 CET] <another> define "better" [02:29:01 CET] <pink_mist> wavpack is fun in that you get two files, one much smaller, and if you only use that one, it's actually lossy (but still pretty good), and then the second which you can combine with the smaller one to get lossless [02:29:16 CET] <Conjecture> Better compression [02:29:18 CET] <Conjecture> Smaller files [02:29:31 CET] <Conjecture> Though that would probably have to be an average and depend on the file type [02:29:32 CET] <pink_mist> at least that was how I saw it used years ago [02:35:52 CET] <furq> Conjecture: https://xiph.org/flac/comparison.html [02:37:13 CET] <pink_mist> oh yeah, monkey's audio was also popular [02:37:38 CET] <another> i think i saw ape compress better [02:37:59 CET] <another> but support is much more limited [02:38:51 CET] <furq> there's also the small matter of it taking 40x as much cpu to decode [02:40:23 CET] <Conjecture> furq, Thanks [02:41:36 CET] <another> all things considered, i think flac is the overall best lossless codec atm [02:43:10 CET] <furq> https://clbin.com/M22pz [02:43:14 CET] <furq> these are the ones that matter [02:44:10 CET] <furq> but yeah anything other than flac is a waste of time [02:46:03 CET] <pink_mist> lol, LEAD Screen Capture is included because L[EAD S]creen matches your regex :P [02:47:12 CET] <furq> oh yeah [02:47:24 CET] <furq> https://clbin.com/Ey8ow [02:47:25 CET] <furq> there you go [02:49:33 CET] <Conjecture> Thanks for the list [08:34:20 CET] <th3_v0ice> Why would av_interleave_write_frame not write 1 last video frame to the file? I tried using the NULL packet to flush but that doesnt seem to help. FFmpeg CLI writes the last frame properly. [08:39:07 CET] <th3_v0ice> I send 361 video AVPackets to the muxer and when I decode the file I get 360 video frames. [11:11:34 CET] <Mavrik> th3_v0ice, did you flush ALL of the components? Video encoder and the muxer? [11:18:27 CET] <th3_v0ice> Mavrik, yes. It seems that setting the AVPacket.duration solved the problem for me. Maybe decoders are not outputting the frame properly when duration of the AVPacket is set to 0. [11:48:48 CET] <newbieG> How find the position of metadata in a mp4 video file like its in start or at the end? [11:52:03 CET] <Mavrik> MP4 should always have metadata at start right? [11:52:07 CET] <Mavrik> My memory is a bit rusty [11:52:47 CET] <BtbN> It commonly has it at the end of the file. [11:52:58 CET] <BtbN> Since it can only be generated after the file is complete. [11:53:08 CET] <BtbN> Moving it to the front is an extra post-processing step. [11:57:36 CET] <newbieG> Mavrik, It should but I have some mp4 file which are not streaming its only playing after downloading [11:58:04 CET] <newbieG> Thats why I thought maybe it has metadata at the end [11:58:58 CET] <BtbN> Players will just request the end via range requests and stream the file just fine usually. [11:59:26 CET] <BtbN> But if in doubt, just qt-faststart the file. [11:59:26 CET] <newbieG> BtbN, Can we pinpoint the metadata position? [11:59:42 CET] <BtbN> Open it in a hex editor and search for the MOOV atom [12:00:29 CET] <JEEB> you can look at how your mp4/mov files look by pointing f.ex. L-SMASH's boxdumper at it with `boxdumper --box FILE |less` [12:05:21 CET] <newbieG> BtbN, <bh:00><bh:00><bh:00><bh:18>ftypmp42<bh:00><bh:00><bh:00><bh:00>mp42mp41<bh:00><bh:02>a<bh:b7>moov<bh:00><bh:00><bh:00>lmvh [12:05:32 CET] <newbieG> This is the starting of metadata [12:06:07 CET] <newbieG> ? [13:00:03 CET] <familiyaF> Hi, I am transcoding via VAAPI on AMD and Intel GPU, the same file I have found that AMD GPU 3x time taken by Intel GPU. Is there a way I can increase the performance observed using AMD GPU [13:01:11 CET] <familiyaF> Intel GPU: Intel Corporation UHD Graphics 620 rev 07 [13:01:46 CET] <familiyaF> AMD GPU: Ellesmere Radeon RX 580 rev e7 [13:12:17 CET] <JaredCE> hi all. i'm using a 3rd party app that makes use of ffmpeg and ffprobe... currently ffprobe seems to be outputting this error: In MFMediaID_IdentifyFile, got Error 20266 (1218268970) FFPROBE generated a zero-length output file. wonder if anyone has any advice [13:13:14 CET] <JEEB> take out the wrapper and call ffprobe yourself with -v verbose or so on the file? see if it outputs anything? [13:13:34 CET] <JEEB> ffprobe -v verbose -i BLAH [13:13:36 CET] <JEEB> should do it just fine [13:13:47 CET] <JaredCE> ok cool [13:13:51 CET] <JaredCE> let me try that [14:29:15 CET] <DHE> termos: the http code calls the tcp code, so take its defaults as well. options for tcp should also work [14:57:23 CET] <termos> ah nice thanks! looks like it's 5s then (5,000,000 microseconds) [16:53:03 CET] <matt1122> Wanted to ask again before I spend a lot of time on this ... does anyone know if it's possible using the FFMPEG cmd to change the output aspect ration on the fly, like when the user rotates their phone (The input is live RTMP) ... however right now the output aspect ratio stays stuck at whatever the original aspect ratio was at start [16:53:39 CET] <matt1122> I don't want to spend a lot of time on this unless someone with more experience believes it's possible :) [22:46:37 CET] <Etua> Hello, I wanted to calculate Chromaprint ID in my C program and being short on time I just popen fpcalc from libchromaprint-tools but now I wonder whether using one of ffmpeg libraries for that purpose could be prettier without adding a few hundred lines to the code. Is it possible to use library with it's demuxers the easy 'black box' way like with fpcalc so without caring about file format or specific parameters and only take advantage of chromaprint [22:46:39 CET] <Etua> muxer? [22:51:27 CET] <Etua> I understand it cannot be oneliner but at the same time ideally I would just need to specify the file path and get the result just like with fpcalc. How close to that ideal can I get? [00:00:00 CET] --- Thu Jan 30 2020

1 0

ffmpeg-devel.log.20200128
by burek 29 Jan '20

29 Jan '20

[00:00:00 CET] --- Tue Jan 28 2020

1 0

ffmpeg.log.20200128
by burek 29 Jan '20

29 Jan '20

[01:32:39 CET] <danilo82> hello, i want to remove Stream #0:1, Stream #0:2, Stream #0:5, Stream #0:6 from a video [01:42:14 CET] <relaxed> danilo82: ffmpeg -i input -map 0:0 -map 0:3 -map 0:4 -c copy output [02:05:04 CET] <Conjecture> Is flac not entirely lossless? [02:05:30 CET] <Conjecture> I was told there was something to do with padding that it did that made it such that it wasn't completely lossless [02:06:30 CET] <c_14> iirc it pads 16bit to 24bit if that's what you mean [02:09:24 CET] <c_14> I think that's just me misremembering shit though [02:11:52 CET] <c_14> you can always test any file you encode by decoding the flac and comparing to the raw input stream [02:18:11 CET] <furq> what does "not entirely lossless" mean [02:18:19 CET] <furq> it's either lossless or it isn't [02:23:20 CET] <pink_mist> well it doesn't support floating point samples, only integer samples, and no more than 32 bits per sample, and no more than 8 audio channels ... at least according to wikipedia :P [02:44:07 CET] <c3r1c3-Win> Just padding from 16 to 24 bits would still be lossless. [02:44:33 CET] <cehoyos> pink_mist: More than 32bits per sample are unlikely [02:44:44 CET] <pink_mist> yeah, that's true [02:54:36 CET] <furq> it doesn't support 32-bit int either [02:54:40 CET] <furq> if anyone actually uses that [02:55:10 CET] <furq> technically it does but practically it doesn't [02:57:09 CET] <furq> also none of that makes flac not lossless, it just means the preprocessing step might not be lossless [03:01:37 CET] <klaxa> just like gif, lol [03:05:45 CET] <linext> can anyone help me add titles to a .xspf playlist? [03:42:14 CET] <cehoyos> furq: Even FFmpeg's flac encoder supports 32bit integer [03:49:27 CET] <ponyrider> linext: does ffmpeg even do that? [03:49:40 CET] <linext> ponyrider, i figured it out [03:49:59 CET] <linext> no, i asked in VLC but no one is available to responde [03:50:04 CET] <furq> cehoyos: how do you get that to work [03:50:14 CET] <furq> if you pass it s32 it'll encode 24-bit [03:50:51 CET] <linext> https://i.imgur.com/8TDVyzR.png [03:51:09 CET] <furq> also last i checked libflac won't encode or decode it, so it's not really usable [03:51:28 CET] <linext> i've got an hdhomerun with a quad tuner [03:51:43 CET] <linext> ffmpeg does a good job recording tv like a DVR [03:52:41 CET] <cehoyos> furq: Sorry, I didn't check carefully enought [03:53:09 CET] <linext> i'm considering write a web app to schedule when ffmpeg records [03:53:18 CET] <furq> there is definitely an encoder that does it and iirc ffmpeg's flac decoder supports it [03:53:35 CET] <furq> maybe flake [03:54:36 CET] <furq> but if you actually need s32 for some perverse reason then you should probably just use wavpack [03:55:15 CET] <furq> or just use fltp wavpack [07:24:09 CET] <Conjecture> Thanks for the help with my flac lossless question [07:24:18 CET] <Conjecture> Good night [11:37:58 CET] <snooky> hi all [11:41:45 CET] <snooky> Hi there. I have a folder with subfolders on my nas. in this are films and tv shows. now i want to make a script (bash) which lists all files, searches, and brings all series to X MB file size and all films to YMB file size. so that in the end all films and tv shows have the same files. I need twopass for that, right? [11:42:58 CET] <snooky> then everyone knows the problem. effects. music. sounds and so on is loud. and the voices of the film quietly. sound effects extremely loud again. is there a way to bring the ffmpeg the audio files to the exact same volume? So that an explosion is as loud as the voices? [11:43:44 CET] <funnybunny2> Why is there an av_strerror and an av_make_error_string? The latter just calls the former. [11:44:20 CET] <funnybunny2> I guess legacy reasons? Which one should I use? [11:47:12 CET] <relaxed> snooky: copy the audio stream [11:47:43 CET] <snooky> and than? [11:48:39 CET] <snooky> i want a constant audio volume [11:48:57 CET] <relaxed> well, I wouldn't transcode my entire library to save space on a NAS [11:50:35 CET] <snooky> I have enough space. it's not about space. audio it annoys me that the effects are louder and the rest is quiet. I want the volume to be the same. ffmpeg normalize? [11:51:10 CET] <relaxed> look at the dynaudnorm filter [11:51:41 CET] <snooky> it's about. my NAS in online. so that I can access it wherever I am via vpn. since I travel a lot, that's practical. but it buffers too much when I have a film of 8/9 gb .. so it has to go down to 3 / 4gb so that the stream runs without buffers. [11:54:10 CET] <funnybunny2> Is there a way to know in advance how big of a buffer I need to allocate for my PCM samples? [11:55:48 CET] <funnybunny2> av_samples_get_buffer_size I guess [12:20:02 CET] <funnybunny2> av_samples_get_buffer_size is giving me a way-too small number compared to how many bytes I seem to be actually receiving [12:23:23 CET] <funnybunny2> I have a 1184-byte 2-channel .wav file which I'm decoding to s16 pcm. av_samples_get_buffer_size returns 192000, but if I do count += frame->linesize[0] * avcodec_context->channels; for each frame, I get a total of 2421760 [12:26:31 CET] <funnybunny2> Ah, I think I see what I messed up... [12:34:09 CET] <funnybunny2> OK [12:34:46 CET] <funnybunny2> I'm not sure how I'm supposed to know what to pass for nb_samples to av_samples_get_buffer_size. It says "the number of samples in a single channel" [12:36:46 CET] <funnybunny2> It would seem that I need the audio duration... [12:42:30 CET] <funnybunny2> Where is everyone? [12:52:02 CET] <JEEB> funnybunny2: $dayjob [12:52:38 CET] <DHE> asleep [13:11:18 CET] <funnybunny2> I'm in the wrong country [13:31:05 CET] <ha3mak> hi [13:31:16 CET] <ha3mak> I'm trying to create a HLS stream from my IP cam with ffmpeg [13:31:25 CET] <ha3mak> I'm doing it like this: ffmpeg -i rtsp://username:password@mycam.tld/videoMain -c:v libx264 -crf 21 -preset veryfast -g 25 -keyint_min 25 -sc_threshold 0 -vsync 2 -an -f hls -hls_time 6 -hls_list_size 3 -hls_flags delete_segments -hls_playlist_type event /tmp/hls/stream.m3u8 [13:31:46 CET] <ha3mak> It works but it never deletes a segment [13:32:08 CET] <ha3mak> It just appends the newer ones to the m3u8 and also keeps them on the disk [13:32:33 CET] <ha3mak> I'm running ffmpeg 4.2.2, I think it's the latest [13:32:59 CET] <ha3mak> Is it a bug or I'm doing someting bad? [14:01:14 CET] <ha3mak> Hmm, I found out if I don't use the -hls_playlist_type event option it works [14:01:59 CET] <BtbN> Sounds about right, since you have a live playlist from how you describe it [14:02:19 CET] <BtbN> An event playlist does not change, and only adds elements to the end [14:03:04 CET] <ha3mak> Thanks, I didn't knew it. My bad, sry :) [17:32:35 CET] <snooky> https://nopaste.xyz/?2c2f0b31a3b4c9de#VII7xFGCY+48jbAC/ngNGdGPWY9YpX9e9czV8… [17:32:36 CET] <snooky> hi all [17:33:01 CET] <snooky> https://nopaste.xyz/?061c33d1fc02a5f5#VttY0rS+YYLU1HF0Ou+5LIX98nOIisPswIpkA… [17:33:04 CET] <snooky> my script [17:33:07 CET] <snooky> what is wrong!? [17:34:07 CET] <BtbN> Well, what is it supposed to do? [17:34:30 CET] <BtbN> Immediate thing that springs to eye is that it looks like you are trying to do two pass encoding, while not using an encoder in the first pass. [17:35:48 CET] <snooky> oh yes [17:35:49 CET] <snooky> thanks [17:35:49 CET] <TanaPanda> why did you name yourself snooky [17:35:51 CET] <snooky> now i runs [17:36:50 CET] <snooky> ok the encodeing runs [17:37:04 CET] <snooky> now i want an strict volume level [17:37:50 CET] <snooky> so if I completely re-encode the film anyway. can I then immediately edit the sound so that the level is completely the same? So the effects are as loud as they are true? [17:43:14 CET] <BtbN> loudnorm also has a two-pass mode you can use [17:48:11 CET] <matt1122> Hello, I have live RTMP input (from red5pro) from iPhone that's either 1280x720 OR 720x1280 ... and it will constantly be changing between ... is it going to be possible to handle this w/ ffmpeg so that the output resolution changes on the fly? [17:49:19 CET] <JEEB> at least the API has no limitations with that. in the worst case you might need to reconfig your encoder, although various encoder wrappers might already be handling that in the function that receives your AVFrame [17:59:33 CET] <matt1122> Thx JEEB [18:36:51 CET] <matt1122> JEEB Is there a way to handle the resolution change (phone rotation) on the fly when using the FFMPEG command? [18:38:29 CET] <matt1122> Currently, when I rotate the phone the aspect ratio stays at whatever the initial aspect ratio was. [19:17:12 CET] <linext> anyone know what GOP stands for? https://imgur.com/a/1mXvhwG [19:17:16 CET] <linext> is it group of pictures? [19:24:43 CET] <matt1122> I think that it's GOP=group of pictures [19:30:27 CET] <linext> do you know what it does? [19:30:28 CET] <Mavrik> linext, yes, the interval between keyframes (roughly) [19:30:40 CET] <linext> ah, so 1 keyframe for 30 in my screenshot? [19:30:46 CET] <Mavrik> Yeah [19:30:55 CET] <Mavrik> Depending on encoder it might not be static [19:31:02 CET] <Mavrik> E.g. it might insert it sooner but never later. [19:31:35 CET] <linext> i plugged a chromecast into a ISV-HEM4 [19:31:43 CET] <linext> the picture is good at 720p [19:32:01 CET] <linext> no PDF manual or support [21:50:01 CET] <plitter> ffmpeg -video_size 1920x1080 -framerate 30 -f x11grab -i :0.0+0,0 -f pulse -i default -f v4l2 -input_format mjpeg -framerate 30 -video_size 640x480 -i /dev/video0 -map 0 -map 1 -map 2 output.mkv I'm having this command get mic, monitor and desktop, but the sound is off from the cam... I thought since I specified all the relevant input to start at the same time that it would also be synchronized... [21:50:02 CET] <plitter> is there something I'm doing wrong or is there something else that I have to do? [22:10:13 CET] <plitter> it also has a tendency to stop recording audio before the video stops... [22:32:21 CET] <cehoyos> plitter: Please provide the command line you tested together with the complete, uncut console output [22:50:48 CET] <plitter> cehoyos: https://clbin.com/kNbjZ [22:52:44 CET] <cehoyos> Try with preset ultrafast, your hardware seems too slow [22:53:57 CET] <cehoyos> And / or remove the map options, you seem to encode video twice [22:54:35 CET] <cehoyos> You want to encode two video streams? [22:54:45 CET] <cehoyos> Then it is normal that the hardware can't do it [22:55:25 CET] <cehoyos> And consider using -pix_fmt yuv420p (x2) if you want the output file to be playable on non-FFmpeg based players [22:58:19 CET] <plitter> my thought is to get all streams together to later put together into a video for youtube or similar, basically I want a face cam to be put on top of my screen recording with audio [23:02:04 CET] <plitter> cehoyos: https://clbin.com/m7x2m [23:02:12 CET] <plitter> I forgot to show my command... [23:02:53 CET] <cehoyos> speed=1.08x indicates that AV sync can be reached (after some time) [23:03:16 CET] <cehoyos> Sorry: [23:03:25 CET] <plitter> so I should let it run longer? [23:03:39 CET] <cehoyos> The third input has an incorrect start time 108690.846475 that makes synchronisation impossible [23:04:33 CET] <plitter> I saw some weird time stamps when I was testing with mpv as well... [23:04:37 CET] <cehoyos> There is an v4l2 input option -timestamps abs that you should test [23:05:26 CET] <plitter> ok, give me some minutes :) [23:17:37 CET] <plitter> I tried with -timestamps now but it fails, saying "Error setting option timestamps to value now" [23:39:50 CET] <cehoyos> I didn't suggest "now" as value for the option -timestamps... [23:44:19 CET] <plitter> I thought I was supposed to set my own time.... [23:47:15 CET] <plitter> https://clbin.com/RaWrF when using abs the audio is still coming after the video [00:00:00 CET] --- Wed Jan 29 2020

1 0

ffmpeg-devel.log.20200127
by burek 28 Jan '20

28 Jan '20

[00:00:00 CET] --- Mon Jan 27 2020

1 0

ffmpeg.log.20200127
by burek 28 Jan '20

28 Jan '20

[13:51:33 CET] <funnybunny2> Is it true that av_register_all is deprecated? https://github.com/leandromoreira/ffmpeg-libav-tutorial/issues/29 The docs don't say that. If I remove the call my program doesn't work [13:57:05 CET] <ritsuka> funnybunny2: which version of ffmpeg are you using? [13:59:05 CET] <funnybunny2> ritsuka: Looks like 3.4.6, but I want to use the latest version. I guess Ubuntu comes with an old version [13:59:33 CET] <funnybunny2> I should switch my code to link against my own compiled version [14:02:46 CET] <funnybunny2> The documentation for avformat_alloc_context should also be updated to say what the function returns [14:35:34 CET] <funnybunny2> Why can avcodec_open2 fail? [14:36:03 CET] <funnybunny2> The documentation doesn't say it returns an AVERROR and I don't know what sort of error message to print [14:42:52 CET] <DHE> funnybunny2: static char errstr[1024]; int errstatus = av_strerror(avcodec_open2_error_return, errstr, sizeof(errstr)-1); if (errstatus < 0) printf("Unknown error\n"); else printf("Error message: %s\n", errstr); [14:43:53 CET] <funnybunny2> DHE: ? [14:46:16 CET] <DHE> av_strerror will convert an error return code to a user/dev friendly string. this is the code I use (sorta) to make use of it. [14:47:25 CET] <funnybunny2> DHE: I know that, but the documentation for avcodec_open2 doesn't say it returns an AVERROR. Looking at the source code, it actually does. But if it didn't, I wouldn't just want to print Unknown error [14:49:12 CET] <funnybunny2> I'd like the documentation to be updated so I can trust the function will not be changed. The documentation is unclear. I guess I'm just supposed to treat the source code as the documentation and assume there is no stability to ffmpeg [14:51:13 CET] <DHE> that's not the right way to describe it. AVERROR is a macro. AVERROR(errno) is its intended usage if an errno code is appropriate, else there are a variety of AVERROR_xxx constants [17:34:54 CET] <JaredCE> i've used ffmpeg to split a file up... how might i use it to recombine the file? [17:39:42 CET] <klaxa> concat muxer? [17:39:58 CET] <klaxa> https://trac.ffmpeg.org/wiki/Concatenate#demuxer [17:40:02 CET] <klaxa> *demuxer [17:43:09 CET] <JaredCE> klaxa that looks like it [21:51:48 CET] <LFSVeteran> more a generic C question.... AVAudioFifo has a field nb_buffers, how to reference nb_buffers if I have AVAudioFifo **fifo ? [22:02:22 CET] <klaxa> is **fifo->nb_buffers invalid? i think so, then you may need brackets? like *(*fifo)->nb_buffers ? [22:02:39 CET] <klaxa> err... *parenthesis [22:10:20 CET] <LFSVeteran> I still get the error "dereferencing pointer to incomplete type 'AVAudioFifo' {aka 'struct AVAudioFifo'}" [22:12:09 CET] <DHE> oh that. it's an abstract type. use av_audio_fifo_size() and/or av_audio_fifo_space() [22:12:09 CET] <c_14> isn't it just (*fifo)->nb_buffers ? [22:12:22 CET] <c_14> or that [22:12:38 CET] <LFSVeteran> also tried [22:12:50 CET] <c_14> you tried av_audio_fifo_size ? [22:13:57 CET] <LFSVeteran> it's not just this I want , I have also other vars in **var form I want to check [22:14:50 CET] <LFSVeteran> hence I'm searching for a generic solution [22:23:22 CET] <c_14> in general it should be (*ptr)->field https://gcc.godbolt.org/z/m5KoG8 [22:28:51 CET] <LFSVeteran> ypedef struct AVAudioFifo AVAudioFifo; [22:29:08 CET] <LFSVeteran> there seems not to be a real struct [22:29:22 CET] <LFSVeteran> at least the fields are not defined [22:29:48 CET] <LFSVeteran> can't find other definition of it in the headers [22:29:54 CET] <c_14> then it's private [22:30:13 CET] <c_14> and any access to its members is through av_ functions [22:31:17 CET] <LFSVeteran> checking relevant functions... [00:00:00 CET] --- Tue Jan 28 2020

1 0

ffmpeg-devel.log.20200126
by burek 27 Jan '20

27 Jan '20

[00:00:00 CET] --- Sun Jan 26 2020

1 0

ffmpeg.log.20200126
by burek 27 Jan '20

27 Jan '20

[03:26:23 CET] <unixabg> Greetings, I am wanting to capture 2 or 3 fps with snapshots from an rtsp feed, I am trying something like: [03:26:32 CET] <unixabg> fmpeg -rtsp_transport tcp -i rtsp://blahblah -vf fps=fps=1/0.8 -strftime 1 %s-%04d.jpg [03:28:31 CET] <unixabg> but I only appear to be getting 1 fps. How can I snapshot 2 or 3 frames per second with a jpeg for each frame? [07:40:21 CET] <f00lest> How can I record my desktop audio using ffmpeg [07:41:04 CET] <f00lest> tried `ffmpeg -f alsa -i hw:0 -t 30 out.wav ` [07:41:17 CET] <f00lest> got `Unknown input format: 'alsa' ` [07:43:13 CET] <f00lest> I've also tried installing `sudo dnf install alsa-lib-devel` [07:43:19 CET] <f00lest> but still issue persists [16:42:14 CET] <LFSVeteran> can't figure out why frame_size is 0, everywhere I check it's zero [19:32:59 CET] <LFSVeteran> https://pastebin.com/L4UZfqHg [19:33:18 CET] <LFSVeteran> seems like trouble starts at opening the input [20:35:00 CET] <DHE> LFSVeteran: frame size of 0 means input size doesn't matter. sample_rate is samples per second, not bit rate or total samples (ie. 2 channel vs 1 doesn't mean double the sample_rate) [20:37:14 CET] <LFSVeteran> ffmpeg -f s32le -ac 2 -ar 192000 -i http://ip:port/stream -acodec pcm_s32le output.wav [20:37:35 CET] <LFSVeteran> sample_rate should be 192000 here? [20:38:08 CET] <LFSVeteran> and channels 2? [20:41:36 CET] <DHE> typically. but 192 kHz audio is pretty high. 96k is normally the highest I see [21:03:40 CET] <LFSVeteran> the problem is that frame_size is 0 and therefore the number of samples is 0 and allocate will fail [21:12:21 CET] <DHE> but it means there's no size requirement. so you can deliver samples as you have them rather than queuing and partitioning like you would need to do for AAC or such [21:14:32 CET] <LFSVeteran> My knowledge with the FFmpeg api is very limited....The resampler can be removed too but how to fix this in the example.... [21:27:29 CET] <LFSVeteran> but if the input framesize doesn't matter, why is the output framesize 0? [21:42:10 CET] <chaosblade> oh, lovely. managed to compile a static ffmpeg and libs with dynamic glibc (with the workaround in https://trac.ffmpeg.org/ticket/6431#comment:7) only to find out something in my combo/libs needs libmvec that is glib 2.22 and up. CentOS/RHEL 7.x is on glibc 2.17 :X [21:42:38 CET] <chaosblade> i wonder if just compiling on an older distro and/or with older glibc would work, but i imagine whatever needs mvec will complain [21:42:52 CET] <chaosblade> probably the nvidia support i included [21:43:23 CET] <chaosblade> maybe should also try cehoyos's suggestion for musl instead. [21:43:27 CET] <DHE> chaosblade: https://github.com/wheybags/glibc_version_header you may be interested in this then... [21:44:29 CET] <chaosblade> Hm. That's an interesting read. Thanks for the heads up! [21:45:44 CET] <cehoyos> chaosblade: I tested musl after suggesting it to you and it passes fate with current FFmpeg, should have a performance advantage [21:45:44 CET] <kepstin> that's an openmp support library i think? if you compile whatever is using it with an older gcc/glibc it'll probably just not use it (and the resulting code might be slower) [21:46:16 CET] <chaosblade> kepstin, yeah, that was what i was thinking. [21:46:37 CET] <cehoyos> This is of course you don't need to many dependencies which may also make compilation with glibc less troublesome [21:48:27 CET] <chaosblade> might be worth a try. was it just a ./configure level change or anything more special? [00:00:00 CET] --- Mon Jan 27 2020

1 0