Recherche avancée

Sur d’autres sites (4771)

• What is difference between using ffmpeg GOP setting or x264opt keyint combination ?

  25 mars 2016, par scaryguy

  Here is a ffmpeg command :

  ffmpeg -i input.mp4 -flags  +global_header -c:v  libx264 -vf scale="864x486",setsar=1:1,setdar=16:9 -profile:v main -level 31 -g 50 -keyint_min 50 -sc_threshold 0 -b:v 700k -pix_fmt yuv420p -c:a libfdk_aac -ar 44100 -ac 2 -b:a 128k output2.mp4

  and here is the other one :

  ffmpeg -i input.mp4 -flags  +global_header -c:v  libx264 -vf scale="864x486",setsar=1:1,setdar=16:9 -x264opts keyint=50:min-keyint=50:no-scenecut -profile:v main -level 31 -b:v 700k -pix_fmt yuv420p -c:a libfdk_aac -ar 44100 -ac 2 -b:a 128k output2.mp4

  Are there any differences between using -g 50 -keyint_min 50 -sc_threshold 0 and -x264opts keyint=50:min-keyint=50:no-scenecut settings to gain constant keyframe interval ?

• Difference between single and double quotes in subprocess [Python 3.4]

  11 décembre 2016, par Vasilis

  I’m using Python 3.4 in Windows Server 12 and I have some Python code that executes the ffmpeg command bellow :

  ffmpeg -i input.mp4 -vf select='not(mod(n\,30)),setpts=N/((30)*TB)' -c:v rawvideo  -pix_fmt uyvy422 -y output.avi

  I use the following code to execute the external command :

  try:
  
      output = subprocess.check_output(command, stderr=subprocess.STDOUT, shell=True)
  
  except subprocess.CalledProcessError as exc:
  
      print ("Command %s failed with error code" % command, exc.returncode, exc.output, file=sys.stderr)

  When I pass the command enclosed in single quotes it successfully runs the command :

  command = 'ffmpeg -i input.mp4 -vf select="not(mod(n\,30)),setpts=N/((30)*TB)" -c:v rawvideo  -pix_fmt uyvy422 -y output.avi'

  When I pass the command as a string enclosed in double quotes it fails :

  command = "ffmpeg -i input.mp4 -vf select='not(mod(n\,30)),setpts=N/((30)*TB)' -c:v rawvideo  -pix_fmt uyvy422 -y output.avi"

  The error message is the following :

  [Eval @ 0000000eaf2fe040] Invalid chars ',setpts=N/((30)*TB)' at the end of expression 'not(mod(n,30)),setpts=N/((30)*TB)'  
  
  [Parsed_select_0 @ 0000000eb0d27ca0] Error while parsing expression 'not(mod
  
  (n,30)),setpts=N/((30)*TB)'  
  
  [AVFilterGraph @ 0000000eb0d0a5a0] Error initializing filter 'select' with args 'not(mod(n\\,30)),setpts=N/((30)*TB)'  
  
  Error opening filters!"

  So it appears that when using double quotes the slash / that is part of the setpts=N/((30)*TB) option is not interpreted correctly, while with double quotes there’s no problem. Note that both commands (either with double or single quotes in the select option) work fine when I run them directly from the command prompt.
  However, I’ve seen many people saying that from a technical perspective single and double quotes make no difference, e.g.

  • http://programmers.stackexchange.com/questions/155176/single-quotes-vs-double-quotes
  • http://programmers.stackexchange.com/questions/155176/single-quotes-vs-double-quotes

  Does slash parsing depend on the quotes around the string or this is just a behavior specific to the executable I’m running ?

• TSCC RLE Decoding : Best way to get difference between two frames

  25 janvier 2016, par Dylan Alvaro

  Hello video codec experts,

  I am working with a video codec which is tscc video codec. It uses RLE algorithm.

  I am looking for a way to get the differences between the current decoded frame and the previous frame and store those differences in an array (the best case scenario would be to store those frame differences in an openCV matrice cv::Mat).

  Currently, I am waiting to have 2 frames decoded, I keep in memory those two decoded frames and from the 3rd one, I am comparing the two previous frames to look for any difference between them and then I store the differences in a vector.
  This algorithm could be optimized because I’m using lost of resources and not taking advantage of the compressed frame...

  If anyone could help me find a way to improve the idea I have behind this it would be great :

  The decode function is the following. (It uses two control bytes at the beginning of the compressed RLE frame to analyse whether it is an end of frame, end of line, jump of pixels, etc...) :

  static int decode_rle(CamtasiaContext *c, unsigned int srcsize)
  
  {
  
      unsigned char *src = c->decomp_buf;
  
      unsigned char *output, *output_end;
  
      int p1, p2, line=c->height, pos=0, i;
  
      uint16_t pix16;
  
      uint32_t pix32;
  
  
  
      output = c->pic.data[0] + (c->height - 1) * c->pic.linesize[0];
  
      output_end = c->pic.data[0] + (c->height) * c->pic.linesize[0];
  
      while(src < c->decomp_buf + srcsize) {
  
          p1 = *src++;
  
          //Escape code
  
          if(p1 == 0) 
  
          {
  
              p2 = *src++;
  
              if(p2 == 0) { //End-of-line
  
                  output = c->pic.data[0] + (--line) * c->pic.linesize[0];
  
                  if (line < 0)
  
                      return -1;
  
                  pos = 0;
  
                  continue;
  
              } else if(p2 == 1) { //End-of-picture
  
                  return 0;
  
              } else if(p2 == 2) { //Skip
  
                  p1 = *src++;
  
                  p2 = *src++;
  
                  line -= p2;
  
                  if (line < 0)
  
                      return -1;
  
                  pos += p1;
  
                  output = c->pic.data[0] + line * c->pic.linesize[0] + pos * (c->bpp / 8);
  
                  continue;
  
              }
  
              // Copy data
  
              if (output + p2 * (c->bpp / 8) > output_end) {
  
                  src += p2 * (c->bpp / 8);
  
                  continue;
  
              }
  
              if (c->bpp == 16) 
  
              {
  
                  for(i = 0; i < p2; i++) 
  
                  {
  
                      pix16 = AV_RL16(src);
  
                      src += 2;
  
                      *(uint16_t*)output = pix16;
  
                      output += 2;
  
                  }
  
              }
  
              pos += p2;
  
          } 
  
          else 
  
          { //Run of pixels
  
              int pix[4]; //original pixel
  
              switch(c->bpp){
  
              case  8: pix[0] = *src++;
  
                       break;
  
              case 16: pix16 = AV_RL16(src);
  
                       src += 2;
  
                       *(uint16_t*)pix = pix16;
  
                       break;
  
              case 24: pix[0] = *src++;
  
                       pix[1] = *src++;
  
                       pix[2] = *src++;
  
                       break;
  
              case 32: pix32 = AV_RL32(src);
  
                       src += 4;
  
                       *(uint32_t*)pix = pix32;
  
                       break;
  
              }
  
              if (output + p1 * (c->bpp / 8) > output_end)
  
                  continue;
  
              for(i = 0; i < p1; i++) {
  
                  switch(c->bpp){
  
                  case  8: *output++ = pix[0];
  
                           break;
  
                  case 16: *(uint16_t*)output = pix16;
  
                           output += 2;
  
                           break;
  
                  case 24: *output++ = pix[0];
  
                           *output++ = pix[1];
  
                           *output++ = pix[2];
  
                           break;
  
                  case 32: *(uint32_t*)output = pix32;
  
                           output += 4;
  
                           break;
  
                  }
  
              }
  
              pos += p1;
  
          }
  
      }

  Thank you for your kind help !