Recherche avancée

Sur d’autres sites (6652)

• delphi firemonkey + FFmpeg Fill image/Tbitmap with data of AVFRAME->pixelformat->YUV420P

  9 février 2020, par coban

  I have managed to create a simple Video player using SDL2 + FFmpeg libraries with Delphi VCL. It’s about the same as ffplay.exe but not a Console app.
  I’ve noticed that FFmpeg (I might be wrong) converts/scales (sws_scale) source pixelformat(any) -> to destination -> YUV420P faster than to any other format.

  What I want to achieve is some kind of a (video)surface, where over I can put other components, like for example a TProgressbar. SDL has a function sdl_createwindowfrom which can turn a tpanel into video(surface/window) where it is possible to put any component over it. But this function is only for windows.

  Maybe I am looking in the wrong direction to achieve what I want, if so, any hint is welcome.
  I was thinkin of drawing the data retrieved in pixelformat yuv420p to a TBitmap of a Timage, this way I won’t need SDL2 library, and I will be able to put any other component above, in this case, Timage. Or another component which might be faster.

  It seems like I need to convert the YUV420P into BGRA format, because TBitmap does not seem to support any YUV format, worse is FIREMONKEY tbitmap is always BGRA format, changing to other format is not possible.

  In first case, I need a function to convert yuv420 to BGRA, can anyone help with this, is there a component/package/function for this which I could use ? Or maybe is it anyhow possible to use yuv420p format directly without converting ?
  I tried to convert some SDL2 functions from SDL2 source (C/C++) to Delphi functions but it’s to complicate for me, specially with my knowledge of C/C++. In SDL2 there are methods/functions implemented for converting RGB <-> YUV. (Why did I ever start Delphi programming ? my mistake).

  BTW, I already tried TMediaplayer, it’s drawing video(picture) above everything, nothing else than the video is visible.

  ---

  I’ve made an attempt, what I don’t understand is where to get/what is "y_stride, uv_stride and rgb_stride"
  Some variable declarations and/or assignments can be incorrect, need to debug the values, but first I need to know what to pass for the above variables.

  ---

      procedure STD_FUNCTION_NAME(width, height:Cardinal;Y, U, V:PByte; Y_stride, UV_stride:Cardinal;
  
                            RGB:PByte;     RGB_stride:Cardinal;yuv_type:YCbCrType;
  
                          YUV_FORMAT,RGB_FORMAT:Word);
  
  var param:PYUV2RGBParam;
  
    y_pixel_stride,
  
    uv_pixel_stride,
  
    uv_x_sample_interval,
  
    uv_y_sample_interval:Word;
  
  
  
    x, ys:Cardinal;
  
    y_ptr1,y_ptr2,u_ptr,v_ptr:PByte;
  
    rgb_ptr1,rgb_ptr2:PByte;
  
  
  
    u_tmp,v_tmp,r_tmp,
  
    g_tmp,b_tmp:Cardinal;
  
    y_tmp:Integer;
  
  begin
  
  param := @(YUV2RGB[integer( yuv_type)]);
  
  if YUV_FORMAT = YUV_FORMAT_420
  
  then begin
  
    y_pixel_stride      := 1;
  
    uv_pixel_stride     := 1;
  
    uv_x_sample_interval:= 2;
  
    uv_y_sample_interval:= 2;
  
  end;
  
  if YUV_FORMAT = YUV_FORMAT_422
  
  then begin
  
    y_pixel_stride        := 2;
  
    uv_pixel_stride       := 4;
  
    uv_x_sample_interval  := 2;
  
    uv_y_sample_interval  := 1;
  
  end;
  
  if YUV_FORMAT = YUV_FORMAT_NV12
  
  then begin
  
    y_pixel_stride        := 1;
  
    uv_pixel_stride       := 2;
  
    uv_x_sample_interval  := 2;
  
    uv_y_sample_interval  := 2;
  
  end;
  
  
  
  
  
  //for(y=0; y<(height-(uv_y_sample_interval-1)); y+=uv_y_sample_interval)
  
  ys := 0;
  
  while ys < height-(uv_y_sample_interval-1) do
  
  begin
  
      y_ptr1  := Y+ys*Y_stride;
  
    y_ptr2  := Y+(ys+1)*Y_stride;
  
    u_ptr   := U+(ys div uv_y_sample_interval)*UV_stride;
  
    v_ptr   := V+(ys div uv_y_sample_interval)*UV_stride;
  
  
  
      rgb_ptr1:=RGB+ys*RGB_stride;
  
  
  
      if uv_y_sample_interval > 1
  
    then rgb_ptr2:=RGB+(ys+1)*RGB_stride;
  
  
  
  
  
      //for(x=0; x<(width-(uv_x_sample_interval-1)); x+=uv_x_sample_interval)
  
  x := 0;
  
  while x<(width-(uv_x_sample_interval-1)) do
  
      begin
  
          // Compute U and V contributions, common to the four pixels
  
  
  
          u_tmp := (( u_ptr^)-128);
  
          v_tmp := (( v_ptr^)-128);
  
  
  
          r_tmp := (v_tmp*param.v_r_factor);
  
          g_tmp := (u_tmp*param.u_g_factor + v_tmp*param.v_g_factor);
  
          b_tmp := (u_tmp*param.u_b_factor);
  
  
  
          // Compute the Y contribution for each pixel
  
  
  
          y_tmp := ((y_ptr1[0]-param.y_shift)*param.y_factor);
  
          PACK_PIXEL(RGB_FORMAT,y_tmp,r_tmp, g_tmp, b_tmp, rgb_ptr1);
  
  
  
          y_tmp := ((y_ptr1[y_pixel_stride]-param.y_shift)*param.y_factor);
  
          PACK_PIXEL(RGB_FORMAT,y_tmp,r_tmp, g_tmp, b_tmp, rgb_ptr1);
  
  
  
          if uv_y_sample_interval > 1
  
    then begin
  
      y_tmp := ((y_ptr2[0]-param.y_shift)*param.y_factor);
  
      PACK_PIXEL(RGB_FORMAT,y_tmp,r_tmp, g_tmp, b_tmp, rgb_ptr2);
  
  
  
      y_tmp := ((y_ptr2[y_pixel_stride]-param.y_shift)*param.y_factor);
  
      PACK_PIXEL(RGB_FORMAT,y_tmp,r_tmp, g_tmp, b_tmp, rgb_ptr2);
  
          end;
  
  
  
          y_ptr1 := y_ptr1 + 2*y_pixel_stride;
  
          y_ptr2 := y_ptr2 + 2*y_pixel_stride;
  
          u_ptr  := u_ptr  + 2*uv_pixel_stride div uv_x_sample_interval;
  
          v_ptr  := v_ptr  + 2*uv_pixel_stride div uv_x_sample_interval;
  
    x := x + uv_x_sample_interval
  
      end;
  
  
  
      //* Catch the last pixel, if needed */
  
      if (uv_x_sample_interval = 2) and (x = (width-1))
  
      then begin
  
          // Compute U and V contributions, common to the four pixels
  
  
  
          u_tmp := (( u_ptr^)-128);
  
          v_tmp := (( v_ptr^)-128);
  
  
  
          r_tmp := (v_tmp*param.v_r_factor);
  
          g_tmp := (u_tmp*param.u_g_factor + v_tmp*param.v_g_factor);
  
          b_tmp := (u_tmp*param.u_b_factor);
  
  
  
          // Compute the Y contribution for each pixel
  
  
  
          y_tmp := ((y_ptr1[0]-param.y_shift)*param.y_factor);
  
          PACK_PIXEL(RGB_FORMAT,y_tmp,r_tmp, g_tmp, b_tmp, rgb_ptr1);
  
  
  
          if uv_y_sample_interval > 1
  
    then begin
  
            y_tmp := ((y_ptr2[0]-param.y_shift)*param.y_factor);
  
      PACK_PIXEL(RGB_FORMAT,y_tmp,r_tmp, g_tmp, b_tmp, rgb_ptr2);
  
            //PACK_PIXEL(rgb_ptr2);
  
          end;
  
      end;
  
  ys := ys +uv_y_sample_interval;
  
  end;
  
  
  
  //* Catch the last line, if needed */
  
  if (uv_y_sample_interval = 2) and (ys = (height-1))
  
  then begin
  
      y_ptr1 :=Y+ys*Y_stride;
  
  u_ptr  :=U+(ys div uv_y_sample_interval)*UV_stride;
  
  v_ptr  :=V+(ys div uv_y_sample_interval)*UV_stride;
  
  
  
      rgb_ptr1:=RGB+ys*RGB_stride;
  
  
  
      //for(x=0; x<(width-(uv_x_sample_interval-1)); x+=uv_x_sample_interval)
  
  x := 0;
  
  while x < (width-(uv_x_sample_interval-1)) do
  
      begin
  
          // Compute U and V contributions, common to the four pixels
  
  
  
          u_tmp := (( u_ptr^)-128);
  
          v_tmp := (( v_ptr^)-128);
  
  
  
          r_tmp := (v_tmp*param.v_r_factor);
  
          g_tmp := (u_tmp*param.u_g_factor + v_tmp*param.v_g_factor);
  
          b_tmp := (u_tmp*param.u_b_factor);
  
  
  
          // Compute the Y contribution for each pixel
  
  
  
          y_tmp := ((y_ptr1[0]-param.y_shift)*param.y_factor);
  
          //PACK_PIXEL(rgb_ptr1);
  
    PACK_PIXEL(RGB_FORMAT,y_tmp,r_tmp, g_tmp, b_tmp, rgb_ptr1);
  
          y_tmp := ((y_ptr1[y_pixel_stride]-param.y_shift)*param.y_factor);
  
          //PACK_PIXEL(rgb_ptr1);
  
    PACK_PIXEL(RGB_FORMAT,y_tmp,r_tmp, g_tmp, b_tmp, rgb_ptr1);
  
  
  
          y_ptr1 := y_ptr1 + 2*y_pixel_stride;
  
          u_ptr  := u_ptr  + 2*uv_pixel_stride div uv_x_sample_interval;
  
          v_ptr  := v_ptr  + 2*uv_pixel_stride div uv_x_sample_interval;
  
  
  
    x := x + uv_x_sample_interval
  
      end;
  
  
  
      //* Catch the last pixel, if needed */
  
      if (uv_x_sample_interval = 2) and (x = (width-1))
  
      then begin
  
          // Compute U and V contributions, common to the four pixels
  
  
  
          u_tmp := (( u_ptr^)-128);
  
          v_tmp := (( v_ptr^)-128);
  
  
  
          r_tmp := (v_tmp*param.v_r_factor);
  
          g_tmp := (u_tmp*param.u_g_factor + v_tmp*param.v_g_factor);
  
          b_tmp := (u_tmp*param.u_b_factor);
  
  
  
          // Compute the Y contribution for each pixel
  
  
  
          y_tmp := ((y_ptr1[0]-param.y_shift)*param.y_factor);
  
          //PACK_PIXEL(rgb_ptr1);
  
    PACK_PIXEL(RGB_FORMAT,y_tmp,r_tmp, g_tmp, b_tmp, rgb_ptr1);
  
      end;
  
  end;

  end ;

• configure : Include quotes around pkg_version

  16 décembre 2024, par Joe Schiffler

  configure : Include quotes around pkg_version
  In some MSYS environments it can happen that the 3 argument syntax
  
  for pkg-config library specifications fails because somehow the
  
  expansion of pkg_version ends up with a redirection we guess.
  To avoid failures like in the referenced build[2], we quote it
  
  so the whole module including operators will be expanded into
  
  a single shell word and the single argument syntax for specifying
  
  the library for pkg-config will be used.
  The single argument syntax seems to be supported by the original
  
  pkg-config from the beginning more than 20 years[3].
  In the pkgconf implementation single argument syntax was supported
  
  pretty much from the beginning as well. The multiple argument syntax
  
  we used until this change, was not supported until a change[4] more
  
  than 10 years ago.
  References
  
  1. Build passing with quotes :
  
     https://github.com/JoeSchiff/pyav-ffmpeg/actions/runs/12358403929
  
  2. Build failing without quotes :
  
     https://github.com/JoeSchiff/pyav-ffmpeg/actions/runs/12360472377
  
  3. Earliest commit of the current pkg-config Git repo already mentions the single argument syntax :
  
     https://gitlab.freedesktop.org/pkg-config/pkg-config/-/commit/2ac96cbcc708d8945329fd1b2001386e1c895c64#124c0becfe68b1ef671f49ed2b9d24779ace126f_0_162
  
  4. pkgconf gets support for 3 argument syntax (pkgconf —exists liba = 1.2.3) :
  
     https://github.com/pkgconf/pkgconf/commit/793de6a06ca52fbfe906a269b5f2b2ba41739517
  Commit-message-by : Alexander Strasser <eclipse7@gmx.net>
  
  Signed-off-by : Joe Schiffler <joeschiffler3@gmail.com>
  
  Signed-off-by : Alexander Strasser <eclipse7@gmx.net>
  

  • [DH] configure

• using pocketsphinx_continuous with a .wav file

  3 avril 2013, par user2242131

  I am attempting to write an application that will allow a user to speak a small set of commands from a remote system and have them executed on my server. Using pocketsphinx to parse the spoken text. When run locally with the microphone, pocketsphinx_continuous works perfectly no matter how I slur the words. But when importing the audio file and using ffmpeg to downsample the audio to a single channel, 16 bit PCM file, it will parse the first word without difficulty. Then it will skip everything else and treat it as . I am confident that the problem is in the file format and not in the pocketsphinx configuration.

  Using command line
  ffmpeg -y -i Sound\AddSheet.wav -ac 1 -f s16le -acodec pcm_s16le -ar 16k AddTmp.wav
  in a batch file.

  The bottom of the output I get is :

  INFO: fsg_search.c(1407): Start node ADD.0:5:47
  
  INFO: fsg_search.c(1407): Start node <sil>.0:2:49
  
  INFO: fsg_search.c(1446): End node <sil>.126:128:305 (-486)
  
  INFO: fsg_search.c(1662): lattice start node <s>.0 end node <sil>.126
  
  INFO: ps_lattice.c(1352): Normalizer P(O) = alpha(<sil>:126:305) = -175371
  
  INFO: ps_lattice.c(1390): Joint P(O,S) = -176076 P(S|O) = -705
  
  000000000: ADD USER
  
  </sil></sil></s></sil></sil>

  Which is not the audio in the file. The words spoken in the file are "ADD SPREADSHEET", which works perfectly from the same microphone without the intervening .wav file.

  I have tried increasing the audio volume and decreasing the background noise using sox :

  sox -v 3.0 Sound\%1 Sound\%1-loud.wav ffmpeg -i Sound\%1-loud.wav -vn -ss 00:00:00 -t 00:00:01 -y Sound\%1-noiseaud.wav 
  
  sox Sound\%1-noiseaud.wav -n noiseprof Sound\%1-noise.prof 
  
  sox Sound\%1 Sound\%1-clean.wav noisered sound\noise.prof 0.21 
  
  ffmpeg -y -i Sound\%1-clean.wav -ac 1 -f s16le -acodec pcm_s16le -ar 16k AddTmp.wav

  with no noticeable effect on the final results.

  If you look at the output you will notice that fsg_search.c has found ADD as the start node, then silence for the remainder. Please help on this.