July 2024 - ffmpeg-user - lists.ffmpeg.org

Creating Motion JPEG 2000 Images
by Aaron Boxer 29 Jul '24

29 Jul '24

Hello, I am able to play back mj2 clips with ffmpeg; can I also create mj2 files by converting from say mp4 ? Thank you! Aaron

3 4

How to extract pts_time fast?
by Moi15 Moi 27 Jul '24

27 Jul '24

Hi, My goal is to extract the timestamps of a video using ffprobe. Currently, I do it with mkvextract like this: mkvextract FILE.mkv timestamps_v2 0:timestamps.txt. I thought of extracting the timestamps like this: ffprobe -select_streams 0 -show_entries frame=pts_time -print_format json VIDEO.mkv. But I quickly realized that ffprobe takes a long time to execute this command (around 1 minute for a 7-minute video). That's why I tried to find a quicker way to extract the timestamps. I tried to use the timestamps of the packets like this: ffprobe -select_streams 0 -show_entries packet=pts_time -print_format json VIDEO.mkv. It is much faster (less than 1 second for a 7 minutes video). >From what I know, the pts_time provided with the packet may differ from the frame one because it's possible but uncommon for multiple frames to be carried in one packet, or for a frame to be split (as slices) across multiple packets. So, here is my question: is there a "heuristic" to know if the packet timestamps will be the same as the frame timestamps? My idea was that if the number of frames is different from the number of packets, then I need to use the frame timestamps. I thought of counting the frames and packets like this: ffprobe -select_streams 0 -show_entries stream=nb_read_frames,nb_read_packets -count_frames -count_packets -print_format json VIDEO.mkv. moi15moi

1 0

pict_type="I"
by Mark Filipak 26 Jul '24

26 Jul '24

I just ran this: ffprobe -sexagesimal -select_streams v -show_frames -of flat -i e:\00305.m2ts Since 00305.m2ts is AVC encoded, there are I slices in it but not I frames -- I P & B frames are MPEG-1 & MPEG-2. So, what does ffprobe mean when it reports this: frames.frame.0.pict_type="I"? Just asking... Anyone know? --Mark.

2 5

Confusion over stream specifier on -map, -codec, and -filter_complex
by gqqnb2005 26 Jul '24

26 Jul '24

Basically, I want to figure out how to specify streams for the option -map, -codec, and -filter_complex. ffmpeg -i INPUT -map 0:1 out.wav Is 0:1 a stream specifier ( https://ffmpeg.org//ffmpeg.html#Stream-specifiers-1)? Which form is it? ffmpeg -i INPUT -map 0 -c copy -c:v:1 libx264 -c:a:137 libvorbis OUTPUT v:1 and a:137 seem to follow the syntax of stream specifier because a and v are stream_type. ffmpeg -i in.flac -filter_complex 'acrossover=split=1500[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav Is '[LOW]' in the syntax of stream specifier?

3 6

Audio out of sync
by iburman＠compuscience.com 24 Jul '24

24 Jul '24

I have multiple mp4 files taken with a cellphone which I concatenate. Playing the concatenated file in a chrome browser, at some point in the middle of the file I start to see the audio starting to get out of sync, and the problem increases as we move to the end of the file. It isn't a large difference but noticeable. Some facts 1 - I tested the file in chrome, both when placed in a web site (through the internet) as well as from my local drive. Both show the async issue. 2 - If I use the player controls to stop the video in the middle, advance the cursor a little bit, and restart playing, the audio is back in sync. In other words, the async only happens when playing the file from start to end without interruption. 3 - Playing the video in media player doesn't have this problem. What could be causing this and how can I fix it? t I was hopping that the async command in ffmpeg could magically fix the problem, but it didn't make a difference. I tried to use it in the two ways below: ffmpeg -i inputFile -async 1 outputFile and ffmpeg -async 1 -i inputFile outputFile Although the problem only happens in chrome, I believe the issue is in the file itself. Most of my files work perfectly fine in chrome and media player, it is just this one that presents this problem. Any ideas on how to find the cause and how to fix it? I think that the fact described in (2) above, where the problem goes away when I stop the player and restart playing from a new position, shows that the time stamps seem to be fine, but maybe chrome doesn't use them the same way as media player, and it just accumulates some small errors, considering that this is a concatenation of multiple small files. Thanks

1 0

Big missmatch In format duration (ffmpeg V6)
by Avraham Sikirov 22 Jul '24

22 Jul '24

Hey, I am trying to upgrade from ffmpeg v5 to v6, I have noticed one video that has big increase in format duration time This is the command I use: ffprobe -v error -print_format json -show_format -show_streams too-many-packets.mp4 Link to video and outputs: https://drive.google.com/drive/folders/1BT6NMqZ2pSd-shbu-hok-2-Gaihn0Av2?us…

3 5

Transform dvdsub to png
by Thomas Ledoux 22 Jul '24

22 Jul '24

Hi all, with the introduction of the new dvdvideo decoder, we are able to access successfully any video and transcode it. Moreover, we can extract subtitles in the native image-based subdvd format. The command ffmpeg -y -hide_banner -f dvdvideo -title 1 -i dvd.iso -vn -an -map 0:i:0x22 -c:s dvdsub title_01_fre_32.vob allows the extraction of the subtitle with the id 0x20 in dvdsub format. If I add the image2 demuxer, I'm able to extract each RLE frames with the PTS in the filename ffmpeg -y -hide_banner -f dvdvideo -title 1 -i dvd.iso -vn -an -copyts -map 0:i:0x20 -c:s dvdsub -f image2 -frame_pts true title_01_fre_32_%d.rle However, if I ask for a image format like png, with a command like this ffmpeg -y -hide_banner -f dvdvideo -title 1 -i dvd.iso -vn -an -copyts -map 0:i:0x20 -f image2 -frame_pts true title_01_fre_32_%d.png I got [sost#0:0 @ 000002187d8e73c0] Automatic encoder selection failed Default encoder for format image2 (codec none) is probably disabled. Please choose an encoder manually. [sost#0:0 @ 000002187d8e73c0] Error selecting an encoder If I explicitly ask for png with ffmpeg -y -hide_banner -f dvdvideo -title 1 -i dvd.iso -vn -an -copyts -map 0:i:0x20 -c:s png -f image2 -frame_pts true title_01_fre_32_%d.png [sost#0:0 @ 0000021eb03cc640] Invalid encoder type 'png' [sost#0:0 @ 0000021eb03cc640] Error selecting an encoder Indeed, PNG is not supposed to be used for subtitles. But in this case, subtitles are images so they could be transform in PNG. Am I forgetting a filter to make it react as if it was a video ? Thanks for any insight Thomas

4 5

MPEG-2 Streams
by Mark Filipak 21 Jul '24

21 Jul '24

I've attached a final version of MPEG-2 Streams. AVC Streams is in the works. If you see something wrong with MPEG-2 Streams, I'd appreciate an authoritative reference. Otherwise, please anticipate discussion to resolve controversy. I assure you it will be good discussion. Thank you --Mark.

1 0

Reordering frames in corrupt GoPro videos
by Vincent Deconinck 21 Jul '24

21 Jul '24

Hi, A few months ago, my GoPro randomly started creating some corrupt (or partly-corrupt) video files, mixing up frames and causing jaggy playback.To illustrate the case, here is an excerpt I generated with "ffmpeg -c copy" from a corrupt source file : https://drive.google.com/file/d/1GD7YiVS2z8h0r6UdsPPIXFgnh9RukoeR/view?usp=… As you can see, frames are mixed up, but with a repeating pattern. By playing the file frame by frame, I observed that the mangling pattern repeats every 5 frames and that frames should be played in order 1-0-3-4-2 - 6-5-8-9-7 - 11-10-13-14-12 - 16-15-18-19-17 - etc. At first I only had two 5-minute corrupted files, so I resorted to : 1) extract each frame to PNG (with "ffmpeg -i in_bad_frame_order.MP4 frame%06d.png") 2) create a text file with the frames in the right order (Excel fiddling) 3) recreate a MP4 file from the PNG files (with "ffmpeg -r 30 -f concat -i full_list.txt -c:v libx264 -pix_fmt yuv420p out_no_audio.mp4") 4) extract audio from original (with "ffmpeg -i in_bad_frame_order.MP4 -vn -acodec copy audio.aac" 5) reinject audio into the rebuilt file (with "ffmpeg -i out_no_audio.mp4 -i audio.aac -c:v copy -c:a copy out_fixed.mp4") That worked pretty well and I was able to use the rebuilt video no prob, but it took around 24 hours to process 10 minutes of video (it is 4K30) and PNGs are huge. Now I got back from a trip and discovered that I have around 10-15 hours of video that are corrupt :-(. I threw away the GoPro, but I would like to recover the footage without spending 2-3 months with my PC burning CPU 24/7 just to rebuild the files the same way, so I wondered if I could fiddle with the PTS of the frames to make them play in the right order. I thus came up with a formula and tried to apply it using ffmpeg, with the following command (and output) : ffmpeg -i in_bad_frame_order.MP4 -bsf:v setts=pts='if(eq(mod(PTS,5),0),PTS+6,if(eq(mod(PTS,5),1),PTS+4,if(eq(mod(PTS,5),2),PTS+6,if(eq(mod(PTS,5),3),PTS+6,PTS+3))))' -c copy out.mp4 ffmpeg version 7.0.1-full_build-www.gyan.dev Copyright (c) 2000-2024 the FFmpeg developers built with gcc 13.2.0 (Rev5, Built by MSYS2 project) configuration: --enable-gpl --enable-version3 --enable-static --disable-w32threads --disable-autodetect --enable-fontconfig --enable-iconv --enable-gnutls --enable-libxml2 --enable-gmp --enable-bzlib --enable-lzma --enable-libsnappy --enable-zlib --enable-librist --enable-libsrt --enable-libssh --enable-libzmq --enable-avisynth --enable-libbluray --enable-libcaca --enable-sdl2 --enable-libaribb24 --enable-libaribcaption --enable-libdav1d --enable-libdavs2 --enable-libuavs3d --enable-libxevd --enable-libzvbi --enable-librav1e --enable-libsvtav1 --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxavs2 --enable-libxeve --enable-libxvid --enable-libaom --enable-libjxl --enable-libopenjpeg --enable-libvpx --enable-mediafoundation --enable-libass --enable-frei0r --enable-libfreetype --enable-libfribidi --enable-libharfbuzz --enable-liblensfun --enable-libvidstab --enable-libvmaf --enable-libzimg --enable-amf --enable-cuda-llvm --enable-cuvid --enable-dxva2 --enable-d3d11va --enable-d3d12va --enable-ffnvcodec --enable-libvpl --enable-nvdec --enable-nvenc --enable-vaapi --enable-libshaderc --enable-vulkan --enable-libplacebo --enable-opencl --enable-libcdio --enable-libgme --enable-libmodplug --enable-libopenmpt --enable-libopencore-amrwb --enable-libmp3lame --enable-libshine --enable-libtheora --enable-libtwolame --enable-libvo-amrwbenc --enable-libcodec2 --enable-libilbc --enable-libgsm --enable-libopencore-amrnb --enable-libopus --enable-libspeex --enable-libvorbis --enable-ladspa --enable-libbs2b --enable-libflite --enable-libmysofa --enable-librubberband --enable-libsoxr --enable-chromaprint libavutil 59. 8.100 / 59. 8.100 libavcodec 61. 3.100 / 61. 3.100 libavformat 61. 1.100 / 61. 1.100 libavdevice 61. 1.100 / 61. 1.100 libavfilter 10. 1.100 / 10. 1.100 libswscale 8. 1.100 / 8. 1.100 libswresample 5. 1.100 / 5. 1.100 libpostproc 58. 1.100 / 58. 1.100 Input #0, mov,mp4,m4a,3gp,3g2,mj2, from 'in_bad_frame_order.MP4': Metadata: major_brand : isom minor_version : 512 compatible_brands: isomiso2mp41 encoder : Lavf60.16.100 Duration: 00:00:07.03, start: 0.000000, bitrate: 43392 kb/s Stream #0:0[0x1](eng): Video: hevc (Main) (hvc1 / 0x31637668), yuvj420p(pc, bt709), 3840x2160 [SAR 1:1 DAR 16:9], 46444 kb/s, 29.97 fps, 29.97 tbr, 90k tbn (default) Metadata: handler_name : GoPro H.265 vendor_id : [0][0][0][0] encoder : GoPro H.265 encoder timecode : 16:17:19:08 Side data: displaymatrix: rotation of -180.00 degrees Stream #0:1[0x2](eng): Audio: aac (LC) (mp4a / 0x6134706D), 48000 Hz, stereo, fltp, 189 kb/s (default) Metadata: handler_name : GoPro AAC vendor_id : [0][0][0][0] Stream #0:2[0x3](eng): Data: none (tmcd / 0x64636D74), 0 kb/s Metadata: handler_name : GoPro H.265 timecode : 16:17:19:08 Stream mapping: Stream #0:0 -> #0:0 (copy) Stream #0:1 -> #0:1 (copy) Output #0, mp4, to 'out.mp4': Metadata: major_brand : isom minor_version : 512 compatible_brands: isomiso2mp41 encoder : Lavf61.1.100 Stream #0:0(eng): Video: hevc (Main) (hvc1 / 0x31637668), yuvj420p(pc, bt709), 3840x2160 [SAR 1:1 DAR 16:9], q=2-31, 46444 kb/s, 29.97 fps, 29.97 tbr, 90k tbn (default) Metadata: handler_name : GoPro H.265 vendor_id : [0][0][0][0] encoder : GoPro H.265 encoder timecode : 16:17:19:08 Side data: displaymatrix: rotation of -180.00 degrees Stream #0:1(eng): Audio: aac (LC) (mp4a / 0x6134706D), 48000 Hz, stereo, fltp, 189 kb/s (default) Metadata: handler_name : GoPro AAC vendor_id : [0][0][0][0] Press [q] to stop, [?] for help [out#0/mp4 @ 0000020e07576a80] video:37078KiB audio:162KiB subtitle:0KiB other streams:0KiB global headers:0KiB muxing overhead: 0.022741% size= 37248KiB time=00:00:06.99 bitrate=43607.6kbits/s speed=8.12x No error is thrown, but when playing back, the file looks exactly the same, as if the PTS were left unchanged. However, with a simpler formula, such as doubling the PTS (ffmpeg -i in_bad_frame_order.MP4 -bsf:v setts=pts='PTS*2' -c copy out.mp4), I can see that the video indeed plays twice as slow, so it seems the ffmpeg syntax is correct but the formula somehow gets ignored. Does anybody have an idea what is wrong ? Or can you advise another way to fix the files ? Thanks for your help, Vincent

4 11

The future of video
by Mark Filipak 17 Jul '24

17 Jul '24

I interviewed at NVIDIA when it was a start-up. My interview wound up being an argument over sprite graphics, which they held in contempt. I did not get the job. Well, guess what? The so-called "slice prediction" in AVC is sprite graphics. It's real image but it's sprites. At Atari, we did sprites in hardware only because the 6502 was 8 bits running at 1.193[18..] MHz (i.e., 3.579[54..]/3), far too slow. Television - TVs will change a lot. The architecture of television has to change in order to make them simpler, more responsive, easer to drive. The entire concept of frames and of FPS is outmoded and will hopefully become outdated soon. To fill human field of vision, including peripheral vision, screen resolution needs to be approximately 12000x6000. If viewers are closer to the screen so that FOV+ is filled, then resolution needs to be proportionately greater. To match human persistence of vision, the refresh rate of individual pels actually needs to be no greater than 10 or 12 refreshes per second. Only the dots that change need be refreshed and only to the extent that they change. Of course, if sub-dots are electrically dynamic instead of electrically static, then TVs need to have some sort of dynamic refresh that's opaque to the input. Bulk screen refresh at the input is undesirable because it would flash at 10 or 12 Hz. TV dots need to be randomly addressable. TVs need to hide their native screen refresh interval, if any, preferably acting as closely as possible to analog. TV sub-dots need to act like memory so that their current values are readily available to use on updates. To take best advantage of compression, pels should be HSV, not RGB. TVs need to have built-in hardware to do HSV-to-R -G -B conversions that are weighted to match the optical properties of their RGB sub-dots. The best circuit way to set sub-dot levels is analog-R -G -B. The amount of circuitry to do that is trivial to nonexistent -- transistor gate capacitance alone may be enough to power LEDs from refresh to refresh -- we will eventually get to the point of actual R G B LED sub-dots, won't we? I'm probably all wrong. I've been surprised before. Video will probably continue to slowly crawl forward on its hands and knees. --Mark.

5 15