Recherche avancée

Sur d’autres sites (8338)

• Compile the GPU version of FFMPEG to OpenCV

  12 juillet 2018, par Dyex719

  When I try to use GPU
  FFMPEG on opencv by using a command like :

  VideoCapture cap("rtsp link");
  
  
  
  for(;;)
  
  {
  
  mat frame;
  
  cap>>frame;
  
  }

  I don’t see any output on nvidia-smi despite compiling the GPU version of FFMPEG.

  I am trying to get FFMPEG to use GPU to improve the speed of my program. I got FFMPEG from cloning this and followed the steps here to get it working.

  I have validated the results using nvidia-smi dmon -i 0 and have seen that dec% increases when I run the command ffmpeg -c:v h264_cuvid -i test_video.mp4 test_output.mkv.

  I then compiled Opencv to use the GPU enabled FFMPEG, I check this by running getBuildInformation from opencv :

  Video I/O: 
  
  DC1394 1.x: NO 
  
  DC1394 2.x: YES (ver 2.2.4) 
  
  FFMPEG: YES 
  
  codec: YES (ver 58.20.103) 
  
  format: YES (ver 58.17.100) 
  
  util: YES (ver 56.18.102) 
  
  swscale: YES (ver 5.2.100) 
  
  resample: NO 
  
  gentoo-style: YES 
  
  GStreamer: NO 
  
  OpenNI: NO 
  
  OpenNI PrimeSensor Modules: NO 
  
  OpenNI2: NO 
  
  PvAPI: NO 
  
  GigEVisionSDK: NO 
  
  UniCap: NO 
  
  UniCap ucil: NO 
  
  V4L/V4L2: NO/YES 
  
  XIMEA: NO 
  
  Xine: NO 
  
  gPhoto2: NO

  I cross-verify that I’m using the same ffmpeg (by checking the version numbers) :

  ffmpeg version N-91273-g67747c8 Copyright (c) 2000-2018 the FFmpeg developers
  
    built with gcc 5.4.0 (Ubuntu 5.4.0-6ubuntu1~16.04.9) 20160609
  
    configuration: --enable-cuda --enable-cuvid --enable-nvenc --enable-nonfree --enable-libnpp --extra-cflags=-I/usr/local/cuda/include --extra-ldflags=-L/usr/local/cuda/lib64
  
    libavutil      56. 18.102 / 56. 18.102
  
    libavcodec     58. 20.101 / 58. 20.101
  
    libavformat    58. 17.100 / 58. 17.100
  
    libavdevice    58.  4.101 / 58.  4.101
  
    libavfilter     7. 25.100 /  7. 25.100
  
    libswscale      5.  2.100 /  5.  2.100
  
    libswresample   3.  2.100 /  3.  2.100

  Please advise me on how I should compile OpenCV to use the GPU version of FFMPEG. Thank you !

• ffmpeg for android out of memory

  15 juin 2018, par Rafael Lima

  I’m getting an weird error when enconding a video with ffmpeg on android

  Failed to inject frame into filter network: Out of memory

  Althoug is a pretty objective error, I do believe this is some "generic message" for another problem, because the video I used for testing is 17MB and less than 11 seconds length and I’ve sucessfully tested with videos bigger than 200MB and longer than 5min

  This problem happens to any video recorded on my phone (the videos were recorded and play fine on my phone but i cant reencode them... other video h264 were encoded fine on same phone)

  this is ffmpeg output

  ffmpeg version 4.0 Copyright (c) 2000-2018 the FFmpeg developers
  
    built with Android (4691093 based on r316199) clang version 6.0.2 (https://android.googlesource.com/toolchain/clang 183abd29fc496f55536e7d904e0abae47888fc7f) (https://android.googlesource.com/toolchain/llvm 34361f192e41ed6e4e8f9aca80a4ea7e9856f327) (based on LLVM 6.0.2svn)
  
    configuration: --prefix=/home/rafa/Desktop/m4/build --target-os=android --arch=i686 --cpu=i686 --cross-prefix=/home/rafa/Desktop/m4/ndk/toolchain/i686/bin/i686-linux-android- --enable-cross-compile --cc=/home/rafa/Desktop/m4/ndk/toolchain/i686/bin/clang --cxx=/home/rafa/Desktop/m4/ndk/toolchain/i686/bin/clang++ --pkg-config=/usr/bin/pkg-config --pkg-config-flags=--static --enable-pic --enable-gpl --enable-nonfree --enable-static --disable-shared --enable-ffmpeg --disable-ffplay --disable-ffprobe --disable-doc --disable-devices --enable-libx264 --enable-libfdk-aac --enable-bsf=aac_adtstoasc --disable-asm
  
    libavutil      56. 14.100 / 56. 14.100
  
    libavcodec     58. 18.100 / 58. 18.100
  
    libavformat    58. 12.100 / 58. 12.100
  
    libavdevice    58.  3.100 / 58.  3.100
  
    libavfilter     7. 16.100 /  7. 16.100
  
    libswscale      5.  1.100 /  5.  1.100
  
    libswresample   3.  1.100 /  3.  1.100
  
    libpostproc    55.  1.100 / 55.  1.100
  
  Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '/data/data/com.tomatedigital.toolsforinstagram/qqq.mp4':
  
    Metadata:
  
      major_brand     : mp42
  
      minor_version   : 0
  
      compatible_brands: isommp42
  
      creation_time   : 2018-06-09T04:28:55.000000Z
  
      location        : -08.0638-034.8723/
  
      location-eng    : -08.0638-034.8723/
  
      com.android.version: 7.0
  
    Duration: 00:00:09.06, start: 0.000000, bitrate: 17349 kb/s
  
      Stream #0:0(eng): Video: h264 (High) (avc1 / 0x31637661), yuv420p(tv, bt709), 1920x1080, 17088 kb/s, SAR 1:1 DAR 16:9, 30.04 fps, 30 tbr, 90k tbn, 180k tbc (default)
  
      Metadata:
  
        rotate          : 90
  
        creation_time   : 2018-06-09T04:28:55.000000Z
  
        handler_name    : VideoHandle
  
      Side data:
  
        displaymatrix: rotation of -90.00 degrees
  
      Stream #0:1(eng): Audio: aac (LC) (mp4a / 0x6134706D), 48000 Hz, stereo, fltp, 255 kb/s (default)
  
      Metadata:
  
        creation_time   : 2018-06-09T04:28:55.000000Z
  
        handler_name    : SoundHandle
  
  Stream mapping:
  
    Stream #0:0 -> #0:0 (h264 (native) -> h264 (libx264))
  
    Stream #0:1 -> #0:1 (copy)
  
  Press [q] to stop, [?] for help
  
  Error reinitializing filters!
  
  Failed to inject frame into filter network: Out of memory
  
  Error while processing the decoded data for stream #0:0
  
  Conversion failed!

• Compile the GPU version of FFMPEG on OpenCV

  22 juin 2018, par Dyex719

  When I try to use GPU
  FFMPEG on opencv by using a command like :

  VideoCapture cap("rtsp link");
  
  
  
  for(;;)
  
  {
  
  mat frame;
  
  cap>>frame;
  
  }

  I don’t see any output on nvidia-smi despite compiling the GPU version of FFMPEG.

  I am trying to get FFMPEG to use GPU to improve the speed of my program. I got FFMPEG from cloning https://git.ffmpeg.org/ffmpeg.git and followed the steps here https://developer.nvidia.com/ffmpeg to get it working.

  I have validated the results using nvidia-smi dmon -i 0 and have seen that dec% increases when I run the command ffmpeg -c:v h264_cuvid -i test_video.mp4 test_output.mkv.

  I then compiled Opencv to use the GPU enabled FFMPEG, I check this by running getBuildInformation from opencv :

  Video I/O:
DC1394 1.x: NO
DC1394 2.x: YES (ver 2.2.4)
FFMPEG: YES
codec: YES (ver 58.20.103)
format: YES (ver 58.17.100)
util: YES (ver 56.18.102)
swscale: YES (ver 5.2.100)
resample: NO
gentoo-style: YES
GStreamer: NO
OpenNI: NO
OpenNI PrimeSensor Modules: NO
OpenNI2: NO
PvAPI: NO
GigEVisionSDK: NO
UniCap: NO
UniCap ucil: NO
V4L/V4L2: NO/YES
XIMEA: NO
Xine: NO
gPhoto2: NO

  I cross-verify that I’m using the same ffmpeg (by checking the version numbers) :

  ffmpeg version N-91273-g67747c8 Copyright (c) 2000-2018 the FFmpeg developers
 built with gcc 5.4.0 (Ubuntu 5.4.0-6ubuntu1~16.04.9) 20160609
 configuration: --enable-cuda --enable-cuvid --enable-nvenc --enable-nonfree --enable-libnpp --extra-cflags=-I/usr/local/cuda/include --extra-ldflags=-L/usr/local/cuda/lib64
 libavutil      56. 18.102 / 56. 18.102
 libavcodec     58. 20.101 / 58. 20.101
 libavformat    58. 17.100 / 58. 17.100
 libavdevice    58.  4.101 / 58.  4.101
 libavfilter     7. 25.100 /  7. 25.100
 libswscale      5.  2.100 /  5.  2.100
 libswresample   3.  2.100 /  3.  2.100

  Please advise me on how I should compile OpenCV to use the GPU version of FFMPEG. Thank you !