Recherche avancée

Sur d’autres sites (5958)

• ffmpeg extracts only part of 1 audio stream from .ts file

  6 juillet 2020, par Shakalakah

  Here is what ffprobe input.ts shows :

  


    libavutil     56. 55.100 / 56. 55.100
  libavcodec    58. 93.100 / 58. 93.100
  libavformat       58. 47.100 / 58. 47.100
  libavdevice       58. 11.100 / 58. 11.100
  libavfilter         7. 86.100 /  7. 86.100
  libswscale          5.  8.100 /  5.  8.100
  libswresample       3.  8.100 /  3.  8.100
  libpostproc       55.  8.100 / 55.  8.100
Input #0, mpegts, from 'D:\Downloads\TEST\audio.ts':
  Duration: 02:22:29.67, start: 0.000000, bitrate: 51 kb/s
  Program 1
    Stream #0:0[0x101]: Audio: aac (LC) ([15][0][0][0] / 0x000F), 48000 Hz, stereo, fltp, 83 kb/s


  


  So the length of .ts file is 2h:22m:29.67s and looks like it contains only audio (AAC) : when i open it in Windows Media Player i can jump to any position within the 2h:22m range and hear sound playing, and there is no video on the screen. This makes me think the extracted audio will also be of 2h:22m length, but i get only a 51m:14s file.

  I tried the following commands :

  1) ffmpeg -i input.ts -vn -acodec copy outputaudio.aac

  2) ffmpeg -i input.ts -map 0:v -map 0:a -c copy outputaudio.aac (as far as i understood this is "extract ALL audiotracks" command)

  3) ffmpeg -i input.ts -ss 00:00:00 -t 02:22:29.6 -q:a 0 -map a outputaudio.aac (to force extraction to the full length, i.e. 2h:22m)

  4) ffmpeg -i input.ts -map 0:a outputaudio.aac -map 0:v outputonlyvideo.avi (i heard this is an alternative way to force extraction of full-length audio through simultaneous extraction of video. Though looks like my .ts has no video, i decided to try this command too and got an error message : Stream map '0:v' matches no streams)

  What am i doing wrong ? Which alternative commands can i try ?

  Is it possible, that the real length of audio stream is 51m:14s only ? But why i can listen to all 2h:22m length in Windows Media Player - could 51m:14s piece be somehow looped inside .ts to create an impression of 2h:22m length ?

  


• Detecting the value scale of statistics returned from ffprobe

  15 septembre 2023, par Farski

  I'm using ffprobe to detect max and min levels for various audio files. An example of the command I'm using is :

  


  ffprobe -v error -f lavfi -i amovie=my_song.mp3,asetnsamples=n=4410,astats=metadata=1:reset=1 -show_entries frame_tags=lavfi.astats.Overall.Max_level,lavfi.astats.Overall.Min_level -of json

  


  The max/min level values returned use different scales, depending on the format of the input file.

  


  For example, an MP3 file may return fractional values from -1.0 to 1.0 representing a percent of maximum level. A signed 16 bit WAV file returns values in the range -32,768 to 32767. A signed 32 bit FLAC file uses the range -2,147,483,648 to 2,147,483,647. In these cases, the bit size of the values matches the bit depth of the audio file.

  


  In other cases, such as a signed 8 bit WAV file, the results are returned using a scale that does not match the input file, such as 16 bit scale (-32,768 to 32767).

  


  I'm trying to determine if there's anyway to detect which format or scale ffprobe is using when the results are returned, besides trying to do it heuristically. I haven't been able to find any other value that gets returned which specifically reflects the number system being used to generate these levels values. sample_fmt does, in some cases, match, but in cases such as a s8 WAV file, sample_fmt would return s8, which does not match the number format of the returned levels values (s16).

  


  If it's not possible to request this information from ffprobe JIT, is there anywhere in the code base that would describe how it determines which scale to use ?

  


• How can I extract PGS/SUP subtitles from mkv container via libav

  24 août 2024, par Elia

  Trying to write a C code that does the equivalent of the following ffmpeg command :

  


  ffmpeg -i video.mkv -map 0:"$STREAM_INDEX" -c:s copy subtitles.sup


  


  I have managed to read the media container, locate the subtitle streams, and filter the relevant one based on stream metadata. However I only succeeded in getting the bitmap data, without the full PGS header and sections.

  


  Tried dumping packet->data to a file hoping it includes full sections, but the data doesn't start with the expected PGS magic header 0x5047.
  
Also using avcodec_decode_subtitle2 doesn't seem to be useful as it only gives bitmap data without the full PGS header and sections.

  


  What steps are supposed to happen after finding the relevant subtitle stream that allows extracting it as raw data that exactly matches the output of the ffmpeg command above ?

  


  Some sample code :

  


      while(av_read_frame(container_ctx, packet) == 0) {
      if(packet->stream_index != i) {
        continue;
      }

      // Try 1 - dump packet data to file
      fwrite(packet->data, 1, packet->size, fout);

      // Try 2 - decode subtitles
      int bytes_read = 0, success = 0;
      AVSubtitle sub;

      if ((bytes_read = avcodec_decode_subtitle2(codec_context, &sub, &success, packet)) < 0) {
        fprintf(stderr, "Failed to decode subtitles %s", av_err2str(bytes_read));
        goto end;
      }
      fprintf(stdout, "Success! Status:%d - BytesRead:%d NumRects:%d\n", success, bytes_read, sub.num_rects);

      if (success) {
        // Only managed to extract bitmap data here via sub.rects
      }

      av_packet_unref(packet);
    }