Recherche avancée

Sur d’autres sites (8627)

• Realtime video processing with javacv on Android

  7 août 2014, par moonie

  I wish to do realtime video processing on Android using opencv/javacv. My basic approach can be summarized as follow :

  onCreate(), generating a new JavaCameraView

  mCameraView = new JavaCameraView(this, mCameraIndex);
  
  mCameraView.setCvCameraViewListener(this);   
  
  setContentView(mCameraView); 

  when the user clicks a button, a ffmpegframe recorder is created

  try{
  
              final long currentTimeMillis = System.currentTimeMillis();
  
              final String appName=getString(R.string.app_name);
  
              final String galleryPath=Environment.getExternalStoragePublicDirectory(Environment.DIRECTORY_PICTURES).toString();
  
              final String albumPath=galleryPath+"/"+appName;
  
              final String photoPath=albumPath+"/"+currentTimeMillis+".avi";
  
              FFmpegFrameRecorder recorder=FFmpegFrameRecorder.createDefault(photoPath, 640, 360);
  
              recorder.setFrameRate(30);
  
              recorder.start();
  
              }catch(FrameRecorder.Exception e){
  
              e.printStackTrace();
  
              }

  then for each CameraFrame, use the following code to process and save the frame

      public Mat onCameraFrame(final CvCameraViewFrame inputFrame){
  
      final Mat rgba = inputFrame.rgba();
  
      //...apply some filters on rgba
  
      if(mIsTakingVideo){
  
          try{
  
            recorder.record(MatToIplImage(rgba,640,320));           
  
          }catch(FrameRecorder.Exception {
  
              e.printStackTrace();
  
              }       
  
      }
  
      return rgba;
  
  }
  
  
  
   public IplImage MatToIplImage(Mat m){
  
    Bitmap bmp=Bitmap.createBitmap(m.width(), m.height(), Config.ARGB_8888);
  
    Utils.matToBitmap(m, bmp);
  
    IplImage image=IplImage.create(m.width(),m.height(),IPL_DEPTH_8U,1);
  
    bmp.copyPixelsToBuffer(image.getByteBuffer());
  
    return image;}

  When the user clicks the button again, the recorder is stopped and released. However the video I create is an empty video. Any idea why ?

  EDIT :
  The video was empty because the format avi is not supported on my test phone...
  Now the video is not empty but it has the following errors

  1. The recorded video has two split windows instead of just one window
  2. The color is very different from the object I recorded
  3. The play rate is too fast.

• How can I rename a file from a txt file with Windows bat file ?

  12 septembre 2022, par user1264599

  I have a batch script that renames a file to input.mkv so it can be processed by a string of other commands in the bat file with a final file called ProcessedVideo.mkv. I capture the OG file name using "dir *.mkv /b>OG_FileName.txt" before being renamed.

  


  How can I rename the final processed mkv file to the name captured in the OG_FileName.txt and maybe add "_Added-Text.mkv" as the last part of my Batch Script ? (Adding text to the file name is not that important if it is too much trouble).

  


  I really thought this would be easy but I'm defeated.

  


• hw_base_encode : make recon_frames_ref optional

  30 août 2024, par Lynne

  hw_base_encode : make recon_frames_ref optional
  Vulkan supports some stupidly odd hardware, that unfortunately,
  
  most modern GPUs happen to be.
  
  The DPB images for encoders may be required to be preallocated
  
  all at once, and rather than be individual frames, be layers of
  
  a single frame.
  As the hw_base_encode code is written with the thought that either
  
  the driver or the device itself supports sane image allocation,
  
  Vulkan does not leave us with this option.
  So, in the case that the hardware does not support individual
  
  frames to be used as DPBs, make the DBP frames context optional,
  
  and let the subsystem manage this.
  

  • [DH] libavcodec/hw_base_encode.c