Recherche avancée

Sur d’autres sites (9543)

• Can not build ffmpeg with gpu acceleration on macOS

  21 février 2018, par Kirill Serebriakov

  I’m trying to use my GPU for video encoding/decoding operations on macOS.

  • OS : MacOS 10.12.5 (Sierra) //hackintosh if it matters
  • CUDA Toolkit 8.0 installed
  • NVidia GTX 1080 with latest web driver

  Followed this guides :

  • https://trac.ffmpeg.org/wiki/CompilationGuide/MacOSX
  • https://developer.nvidia.com/ffmpeg

  Config :

  ./configure --enable-cuda --enable-cuvid --enable-nvenc \
  
  --enable-nonfree --enable-libnpp \
  
  --extra-cflags=-I/Developer/NVIDIA/CUDA-8.0/include \
  
  --extra-ldflags=-L/Developer/NVIDIA/CUDA-8.0/lib

  Got this error :

  ERROR: cuvid requested, but not all dependencies are satisfied: cuda

  config.log - full configure log

  I did not install Video Codec SDK (not sure how to make it on macOS, just thought that it may come with cuda toolkit) and according to this page I have a lot of limitations on OSX.

  Is it possible on macOS ? Or this will work only for linux/windows ?

• Saving animation using FFMPEG is cutting off subplot

  13 juin 2017, par Aklys

  I’m using the following code to save an animated figure with two plots :

  c_anim = animation.FuncAnimation(c_fig, c_animate, init_func=c_init, repeat=True,
  
                                   frames=len(main_df.reset_index()['DATE'].tolist()),
  
                                   interval=graph_interval_speed, blit=True)
  
  plt.rcParams['animation.ffmpeg_path'] = '.\\FFMPEG\\bin\\ffmpeg.exe'
  
  plt.rcParams['animation.bitrate'] = -1
  
  FFwriter = animation.FFMpegWriter(fps=30, codec='libx264', extra_args=['-s', '1080:1920', '-aspect', '16:9'])
  
  c_anim.save('basic_animation.mp4', writer=FFwriter, dpi=100)

  But the video file it outputs cuts off the second plot. The animation works fine with plt.show() and the figure or plot have not been given specific sizes.

  I’m using the following versions :

  • python = 3.6.1
  • matplotlib = 1.5.3
  • ffmpeg = N-82225-gb4e9252
  • OS = Windows 10

  Please advise me if further information is needed as it’s the first time I’ve attempted to record an animated plot.

  How do I adjust the aspect and resolution to work with the figure itself so nothing is cut off in the video ?

• Android OpenCV - How can i read video files with using jni ?

  20 juin 2017, par Tiga

  I am developing an application using Android Opencv.

  This app, which I am developing, offers two operations.

  • The frame read from the camera is passed to Jni using native function
    Mat.getNativeObjAddr (), and the new image is returned through
    javaCameraView’s onCameraFrame() function
  • It reads a video clip inside Storage, processes each frame the same
    as # 1, and returns the resulting image via the onCameraFrame()
    function.

  First,function is implemented as simple as the following and works normally :

     @Override
  
      public Mat onCameraFrame(CameraBridgeViewBase.CvCameraViewFrame inputFrame) 
  
      {
  
          if(inputFrame!=null){
  
                  Detect(inputFrame.rgba().getNativeObjAddr(), boardImage.getNativeObjAddr());
  
              }
  
              return boardImage;
  
          }
  
      }

  However, the problem occurred in the second operation.
  As far as I know, the files inside Java Storage are not readable by jni.

  I already tried FFmpegMediaPlayer or MediaMetadataRetriever through Google search. However, the getFrameAtTime () function provided by this MetadataRetriever took an average of 170ms when grabbing a bitmap to a specific frame of 1920 * 1080 image. What I have to develop is to show the video results in real time at 30 fps. In # 1, the native function Detect () takes about 2ms to process one frame.

  For these reasons, I want to do this.

  java sends a video’s path (eg : /storage/emulated/0/download/video.mp4) to jni, and native functions process the video one frame at a time, and display the result image on ’onCameraFrame’.

  Is there a proper way ? I look forward to your reply. Thank you !