Recherche avancée

Sur d’autres sites (9435)

• Calling ffmpeg api from Oracle

  1er mai 2012, par TenG

  I have installed ffmpeg and ffmpeg-devel packages on Linux.

  Oracle 11g is installed and running.

  The database stores media files, and for better streaming we need to convert them to AVI format.

  For ease of integration, we would like to do this conversion in the database.

  Now, the simplest option is to write a wrapper for the ffmpeg command line utility, and enable a PLSQL procedure to call this.

  However this would require the following steps :

  1. Read video BLOB
  2. Write to a OS file
  3. Call ffmpeg wrapper giving file name from (2) and output file name
  4. Load output file from 3 into a BLOB in PLSQL

  I would like to if possible write a C routine (using the Oracle External Library feature) which accepts the input as the BLOB (OciLOBLocator), calls the appropriate libavformat functions presenting the LOB, and write the return to a LOB (again OciLOBLOcator) which is what the PLSQL layer then uses as the AVI file.

  The other advantage of this is it avoids the undesirable impact of issuing a OS command from within Oracle.

  The problem I have is that the examples given for ffmpeg show the processing of data from files, whereas I need the libraries to process the LOBs.

  The alternative is to see if the OrdVideo data type in Oracle does this kind of conversion by using setformat and process.

• How can you pass YUV frames from FFmpeg to OpenGL ES ?

  24 mai 2012, par TheRock

  Has anybody tried to use FFmpeg to decode a video frame, then display it in OpenGL ES in iOS 5.0 ?

  I tried to modify the GLCameraRipple example from Apple, but I always get a -6683 error from CVOpenGLESTextureCacheCreateTextureFromImage().

  Here is my decode code :

  ...
  
  
  
  convertCtx = sws_getContext(codecCtx->width, codecCtx->height, codecCtx->pix_fmt, 
  
                              codecCtx->width, codecCtx->height, PIX_FMT_NV12, 
  
                              SWS_FAST_BILINEAR, NULL, NULL, NULL);
  
  
  
  srcFrame = avcodec_alloc_frame();
  
  dstFrame = avcodec_alloc_frame();
  
  width = codecCtx->width;
  
  height = codecCtx->height;
  
  
  
  outputBufLength = avpicture_get_size(PIX_FMT_NV12, width, height);
  
  outputBuf = malloc(outputBufLength);
  
  
  
  avpicture_fill((AVPicture *)dstFrame, outputBuf, PIX_FMT_NV12, width, height);
  
  
  
  ...
  
  
  
  avcodec_decode_video2(codecCtx, srcFrame, &gotFrame, pkt);
  
  
  
  ...
  
  
  
  sws_scale(convertCtx, 
  
            (const uint8_t**)srcFrame->data, srcFrame->linesize, 
  
            0, codecCtx->height, 
  
            dstFrame->data, dstFrame->linesize);

  Here is my code for display :

      CVPixelBufferRef pixelBuffer;
  
      CVPixelBufferCreateWithBytes(kCFAllocatorDefault, [videoDecoder width], [videoDecoder height], 
  
                                   kCVPixelFormatType_420YpCbCr8BiPlanarVideoRange, 
  
                                   dstFrame->data[0], dstFrame->linesize[0], 0, 0, 0, 
  
                                   &pixelBuffer);
  
  
  
      ...
  
  
  
  CVReturn err;
  
  int textureWidth = CVPixelBufferGetWidth(pixelBuffer);
  
  int textureHeight = CVPixelBufferGetHeight(pixelBuffer);
  
  
  
  if (!videoTextureCache)
  
  {
  
      NSLog(@"No video Texture cache");
  
  }
  
  
  
  CVPixelBufferLockBaseAddress(pixelBuffer, 0);
  
  
  
  // Y-plane
  
  err = CVOpenGLESTextureCacheCreateTextureFromImage(kCFAllocatorDefault, 
  
                                                     videoTextureCache, 
  
                                                     pixelBuffer, 
  
                                                     NULL, 
  
                                                     GL_TEXTURE_2D, 
  
                                                     GL_RED_EXT, 
  
                                                     textureWidth, 
  
                                                     textureHeight,
  
                                                     GL_RED_EXT, 
  
                                                     GL_UNSIGNED_BYTE, 
  
                                                     0, 
  
                                                     &lumaTexture);
  
  if (err)
  
  {
  
      NSLog(@"Error at CVOpenGLESTextureCacheCreateTextureFromImage %d", err);
  
  }
  
  
  
  glBindTexture(CVOpenGLESTextureGetTarget(lumaTexture), CVOpenGLESTextureGetName(lumaTexture));
  
  glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
  
  glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
  
  
  
  // UV-plane
  
  err = CVOpenGLESTextureCacheCreateTextureFromImage(kCFAllocatorDefault, 
  
                                                     videoTextureCache, 
  
                                                     pixelBuffer, 
  
                                                     NULL, 
  
                                                     GL_TEXTURE_2D, 
  
                                                     GL_RG_EXT, 
  
                                                     textureWidth / 2, 
  
                                                     textureHeight / 2,
  
                                                     GL_RG_EXT, 
  
                                                     GL_UNSIGNED_BYTE, 
  
                                                     1, 
  
                                                     &chromaTexture);
  
  if (err)
  
  {
  
      NSLog(@"Error at CVOpenGLESTextureCacheCreateTextureFromImage %d", err);
  
  }
  
  
  
  glBindTexture(CVOpenGLESTextureGetTarget(chromaTexture), CVOpenGLESTextureGetName(chromaTexture));
  
  glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
  
  glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
  
  
  
  CVPixelBufferUnlockBaseAddress(pixelBuffer, 0);

  I know that the code is not complete but it should be enough to understand my problem.

  Could anybody please help me or show me some working example with this approach ?

• Légendes et interfaces tactiles

  8 juin 2013

  Le système de visualisation de légendes n’est pas optimisé du tout pour les systèmes tactiles (iOS / Android par exemple).

  Cet exemple est intéressant : http://io9.com/heres-what-pangea-looks-like-mapped-with-modern-politi-509812695

  Utiliser un bouton show / hide des légendes pour les afficher au lieu de ne se baser que sur le hover.

  Être fluide au redimensionnement de la fenêtre.