Recherche avancée

Sur d’autres sites (10907)

• FFmpeg decode video to lossless frames hardware acceleration CUDA

  29 décembre 2020, par Seva Safris

  I have 1 second H.265-encoded videos at 30 fps coming into a server for processing. The server needs to decode the videos into individual frames (lossless). These videos are coming in very quickly, so performance is of utmost importance. The server has a H.265 compatible Nvidia GPU, and I have built ffmpeg with support for CUDA. The following is the configuration output from ffmpeg :

  


  ffmpeg version N-100479-gd67c6c7f6f Copyright (c) 2000-2020 the FFmpeg developers
  built with gcc 8 (Ubuntu 8.4.0-3ubuntu2)
  configuration: --enable-nonfree --enable-cuda-nvcc --enable-nvenc --enable-opencl --enable-shared
                 --enable-pthreads --enable-version3 --enable-avresample --enable-ffplay --enable-gnutls
                 --enable-gpl --disable-libaom --disable-libbluray --disable-libdav1d 
                 --disable-libmp3lame --enable-libopus --disable-librav1e --enable-librubberband
                 --enable-libsnappy --enable-libsrt --enable-libtesseract --enable-libtheora
                 --enable-libvidstab --enable-libvorbis --enable-libvpx --enable-libwebp --enable-libx264
                 --enable-libx265 --enable-libxml2 --enable-libxvid --enable-lzma --enable-libfontconfig
                 --enable-libfreetype --enable-frei0r --enable-libass --enable-libopencore-amrnb
                 --enable-libopencore-amrwb --disable-libopenjpeg --enable-librtmp --enable-libspeex
                 --enable-libsoxr --disable-videotoolbox --disable-libjack --disable-indev=jack
                 --extra-cflags=-I/usr/local/cuda/include --extra-ldflags=-L/usr/local/cuda/lib64


  


  I decode the videos into PNGs, and am using the following command :

  


  ffmpeg -y -vsync 0 -hwaccel cuvid -hwaccel_output_format cuda -hwaccel_device 0 -c:v hevc_cuvid \
       -i 0.mp4 -vf hwdownload,format=nv12 -q:v 1 -qmin 1 -qmax 1 -start_number 0 f%d.png


  


  This command successfully leverages the hardware acceleration for the H.265 decode. But, the PNG encode is done by the CPU.

  


  Does CUDA have support for encoding of lossless images ? The format does not need to be PNG, but it does need to be lossless. CUDA has a nvJPEG Library, but JPEG is a lossy format. Is there a similar image encoding library in CUDA for a lossless format (that is also integrated with ffmpeg) ?

  


  Edit : Some more context....

  


  I am currently using PNGs because of their compression-ability. These images are 2560x1280 in size, btw. On one hand, it is this compression that costs the CPU cycles. On the other hand, I am also limited by the throughput of how fast (and how much aggregate data) can I upload these frames to the upstream consumer. So it's basically a tradeoff between :

  


  

  1. We want to extract these frames as quickly as possible.

  


  2. We want efficiency regarding the image size.

  


  


• On ALAC’s Open Sourcing

  1er novembre 2011, par Multimedia Mike — Codec Technology

  Apple open sourced their lossless audio codec last week. Pretty awesome ! I have a theory that, given enough time, absolutely every codec will be open source in one way or another.

  I know I shouldn’t bother reading internet conversation around any news related to multimedia technology. And if I do read it, I shouldn’t waste any effort getting annoyed about them. But here are some general corrections :

  • ALAC is not in the same league as — nor is it a suitable replacement for — MP3/AAC/Vorbis or any other commonly used perceptual audio codec. It’s not a matter of better or worse ; they’re just different families of codecs designed for different purposes.
  • Apple open sourced ALAC, not AAC– easy mistake, though there’s nothing to ‘open source’ about AAC (though people can, and will, argue about its absolute ‘open-ness’).
  • There’s not much technical room to argue between ALAC and FLAC, the leading open source lossless audio compressor. Both perform similarly in terms of codec speeds (screamingly fast) and compression efficiency (results vary slightly depending on source material).
  • Perhaps the most frustrating facet is the blithe ignorance about ALAC’s current open source status. While this event simply added an official “open source” status to the codec, ALAC has effectively been open source for a very long time. According to my notes, the ALAC decoding algorithm was reverse engineered in 2005 and added into FFmpeg in March of the same year. Then in 2008, Google — through their Summer of Code program — sponsored an open source ALAC encoder.

  From the multimedia-savvy who are versed in these concepts, the conversation revolves around which would win in a fight, ALAC or FLAC ? And who between Apple and FFmpeg/Libav has a faster ALAC decoder ? The faster and more efficient ALAC encoder ? I contend that these issues don’t really matter. If you have any experience working with lossless audio encoders, you know that they tend to be ridiculously fast to both encode and decode and that many different lossless codecs compress at roughly the same ratios.

  As for which encoder is the fastest : use whatever encoder is handiest and most familiar, either iTunes or FFmpeg/Libav.

  As for whether to use FLAC or ALAC — if you’ve already been using one or the other for years, keep on using it. Support isn’t going to vanish. If you’re deciding which to use for a new project, again, perhaps choose based on software you’re already familiar with. Also, consider hardware support– ALAC enjoys iPod support, FLAC is probably better supported in a variety of non-iPod devices, though that may change going forward due to this open sourcing event.

  For my part, I’m just ecstatic that the question of moral superiority based on open source status has been removed from the equation.

  Code-wise, I’m interested in studying the official ALAC code to see if it has any corner-case modes that the existing open source decoders don’t yet account for. The source makes mention of multichannel (i.e., greater than stereo) configurations, but I don’t know if that’s in FFmpeg/Libav.

• Making a movie out of pictures in correct order

  6 novembre 2022, par astrogab

  Short version

  


  How can one combine files img1000.png, img5000.png, img10000.png, img11000.png in the right order into a movie ?

  


  Longer version

  


  I am using ffmpeg to make a movie out of snapshots of a simulation. There should be for instance 5 images per second. The names are :

  


  image0200.png
image0300.png
image0400.png
image0500.png
image1000.png
image1500.png
image2000.png
...
image8500.png
image9000.png
image9500.png
image10000.png
image15000.png


  


  i.e., they are sequential but there are irregular gaps in the numbers. The numbers are formatted according to '%04d' but go above 9999. I have tried

  


  ffmpeg -y -loglevel debug -nostats \
-r:v 5 -thread_queue_size 1024 -f image2 \
-pattern_type glob -i "*[0-9][0-9][0-9][0-9].png" \
-r:v 30 -preset veryslow -pix_fmt yuv420p -crf 28 \
-an  AMDG.mp4


  


  and many, many other variations but either only two frames end up being visible in the movie (even though the images are found when using -debug) or only the files up to image9500.png are used (and glob does not seem to allow [0-9]{4,} as for regex), or, with

  


      ffmpeg -y -loglevel debug -nostats \
    -r:v 5 \
    -thread_queue_size 1024 -f image2 -pattern_type glob \
       -i "image[0-9][0-9][0-9][0-9].png" \
    -r:v 5 \
    -thread_queue_size 1024 -f image2 -pattern_type glob \
       -i "image[1-9][0-9][0-9][0-9][0-9].png" \
    -r:v 30 -preset veryslow -pix_fmt yuv420p -crf 28 \
    -map 0 -map 1 \
    -an  AMDG.mp4


  


  there are apparently two streams in the output movie and only one of them is being played. (I realised in the process -map 0 -map 1 was needed in order for both input streams to be used.)

  


  In one of the variations of options I found (now I have lost what it was exactly !) all images were included but the order was not the desired one : image1000.png was shown before image10000.png. Apparently a newer version of ffmpeg (I have ffmpeg version 3.4.8-0ubuntu0.2) has the ability to sort like sort -V, so that image10000 come after image1000, but reinstalling ffmpeg is in general not a practical option. Also renaming the files is not practical and creating e.g. soft links with sequential names in the format '%05d' starting at 0 and in steps of 1 (so that -i '%05d' could be used) is of course not elegant.

  


  With the -concat filter as in https://unix.stackexchange.com/questions/77016/ffmpeg-pattern-type-glob-not-loading-files-in-correct-order, i.e.,

  


    ffmpeg -y -loglevel debug -nostats -r:v 5 \
    -thread_queue_size 1024 -f image2 -f concat \
    -safe 0 -i <(find . -maxdepth 1 -regex 'image*.png' \
       -exec echo "file $(pwd)/"{} \; | sort -V) \
    -r:v 30 -codec:v libx264 -preset veryslow -pix_fmt yuv420p -crf 28 \
    -an \
    AMDG.mp4


  


  the processing took a long time and made the whole system sluggish, while producing a movie of 60 kB showing only two different images.

  


  I have the impression that there are several issues at once... Thanks if you can help !