Recherche avancée

Sur d’autres sites (7955)

• How to optimize FFMPEG/ Editing video ?

  6 mars 2016, par user6964

  I have the next commands for editing video but all the process take a long time. But with the same quality of the original video.

  //First cut original video
  
  exec("ffmpeg -i $video_path_main -ss $first_time1 -t $first_time2 -s 476x268 -r 10 -b 2000k -r 30 -g 100 -ar 22050 -ab 48000 -ac 1 -strict -2 $name_first");
  
  exec("ffmpeg -i $video_path_main -ss $second_time1 -t $second_time2 -s 476x268 -r 10 -b 2000k -r 30 -g 100 -ar 22050 -ab 48000 -ac 1 -strict -2 $name_second");
  
  
  
  $name_edit_second = uniqid() . '.mp4'; //Then editing the second video
  
  exec("ffmpeg -i $name_second -s 476x268 -r 10 -b 2000k -r 30 -g 100 -ar 22050 -ab 48000 -ac 1 -strict -2 -vf movie='" . $image_name . " [watermark]; [in] [watermark] overlay=308:43"."' $name_edit_second");
  
  
  
  //Then merge video file mp4 with Mencoder
  
  $name_total_1 = uniqid() . '.mp4';
  
  exec("mencoder -oac pcm -ovc xvid -vf scale -xvidencopts bitrate=460 -o $name_total_1 ".$name_first.' '.$name_edit_second);
  
  
  
  //Then convert the video to 3 formats that is necessary in my Player.
  
  $name_total = uniqid();
  
  
  
  //Of MP4 a FLV
  
  exec("ffmpeg -i $name_partial -f flv -s 476x268 -r 10 -b 2000k -r 30 -g 100 -ar 22050 -ab 48000 -ac 1 $name_total.flv");
  
  
  
  //Of MP4-Mencoder a MP4-FFMPEG
  
  exec("ffmpeg -i $name_partial -s 476x268 -r 10 -b 2000k -r 30 -g 100 -ar 22050 -ab 48000 -ac 1 -strict -2 $name_total.mp4"));
  
  
  
  //Of MP4 a WEBM
  
  exec("ffmpeg -i $name_partial -acodec libvorbis -s 476x268 -r 10 -b 2000k -r 30 -g 100 -ar 22050 -ab 48000 -ac 2 -f webm $name_total.webm");

  I don’t know if some of parameters take much time for all the process. Or if one of this command take much time.

  Note : Some videos have more than 2 parts of their original videos.

  ---

  UPDATE

  Maybe the parameter -theards 1 help me in NO take a lot of resources of the CPU. Also, I need to optimize the re-encoding because with only 8 users take the 100% of resources.

  I run FFMPEG in a other server that return the video edited to other server where stay my application.

  Sorry for my english.

• Convert .m4a to PCM using libavcodec

  17 décembre 2013, par gmcc051

  I'm trying to convert a .m4a file to raw PCM file so that I can play it back in Audacity.

  According to the AVCodecContext it is a 44100 Hz track using the sample format AV_SAMPLE_FMT_FLTP which, to my understanding, when decodeded using avcodec_decode_audio4, I should get two arrays of floating point values (one for each channel).

  I'm unsure of the significance of the AVCodecContext's bits_per_coded_sample = 16

  Unfortunately Audacity plays the result back as if I have the original track is mixed in with some white noise.

  Here is some sample code of what I've been done. Note that I've also added a case for a track that uses signed 16bit non-interleaved data (sample_format = AC_SAMPLE_FMT_S16P), which Audacity plays back fine.

  int AudioDecoder::decode(std::string path)
  
  {
  
    const char* input_filename=path.c_str();
  
  
  
    av_register_all();
  
  
  
    AVFormatContext* container=avformat_alloc_context();
  
    if(avformat_open_input(&container,input_filename,NULL,NULL)<0){
  
      printf("Could not open file");
  
    }
  
  
  
    if(avformat_find_stream_info(container, NULL)<0){
  
        printf("Could not find file info");
  
    }
  
    av_dump_format(container,0,input_filename,false);
  
  
  
    int stream_id=-1;
  
    int i;
  
    for(i=0;inb_streams;i++){
  
      if(container->streams[i]->codec->codec_type==AVMEDIA_TYPE_AUDIO){
  
          stream_id=i;
  
          break;
  
      }
  
    }
  
    if(stream_id==-1){
  
      printf("Could not find Audio Stream");
  
    }
  
  
  
    AVDictionary *metadata=container->metadata;
  
    AVCodecContext *ctx=container->streams[stream_id]->codec;
  
    AVCodec *codec=avcodec_find_decoder(ctx->codec_id);
  
  
  
    if(codec==NULL){
  
      printf("cannot find codec!");
  
    }
  
  
  
    if(avcodec_open2(ctx,codec,NULL)<0){
  
       printf("Codec cannot be found");
  
    }
  
  
  
    AVSampleFormat sfmt = ctx->sample_fmt;
  
  
  
    AVPacket packet;
  
    av_init_packet(&packet);
  
    AVFrame *frame = avcodec_alloc_frame();
  
  
  
    int buffer_size = AVCODEC_MAX_AUDIO_FRAME_SIZE+ FF_INPUT_BUFFER_PADDING_SIZE;;
  
    uint8_t buffer[buffer_size];
  
    packet.data=buffer;
  
    packet.size =buffer_size;
  
  
  
    FILE *outfile = fopen("test.raw", "wb");
  
  
  
    int len;
  
    int frameFinished=0;
  
  
  
    while(av_read_frame(container,&packet) >= 0)
  
    {
  
        if(packet.stream_index==stream_id)
  
        {
  
          //printf("Audio Frame read \n");
  
          int len=avcodec_decode_audio4(ctx, frame, &frameFinished, &packet);
  
  
  
          if(frameFinished)
  
          {       
  
            if (sfmt==AV_SAMPLE_FMT_S16P)
  
            { // Audacity: 16bit PCM little endian stereo
  
              int16_t* ptr_l = (int16_t*)frame->extended_data[0];
  
              int16_t* ptr_r = (int16_t*)frame->extended_data[1];
  
              for (int i=0; inb_samples; i++)
  
              {
  
                fwrite(ptr_l++, sizeof(int16_t), 1, outfile);
  
                fwrite(ptr_r++, sizeof(int16_t), 1, outfile);
  
              }
  
            }
  
            else if (sfmt==AV_SAMPLE_FMT_FLTP)
  
            { //Audacity: big endian 32bit stereo start offset 7 (but has noise)
  
              float* ptr_l = (float*)frame->extended_data[0];
  
              float* ptr_r = (float*)frame->extended_data[1];
  
              for (int i=0; inb_samples; i++)
  
              {
  
                  fwrite(ptr_l++, sizeof(float), 1, outfile);
  
                  fwrite(ptr_r++, sizeof(float), 1, outfile);
  
               }
  
             }            
  
          }
  
      }
  
  }
  
  fclose(outfile);
  
  av_close_input_file(container);
  
  return 0;   

  }

  I'm hoping I've just done a naive conversion (most/less significant bit issues), but at present I've been unable to figure it out. Note that Audacity can only import RAW float data if its 32bit or 64 bit float (big or little endian).

  Thanks for any insight.

• Revision af3519a385 : Change the use of a reserved color space entry This commit rename a reserved co

  7 novembre 2014, par Yaowu Xu

  Changed Paths :
   Modify /vp9/common/vp9_enums.h

  
  
  Change the use of a reserved color space entry

  This commit rename a reserved color space entry to BT_2020, it intends
  to provide support for VP9 bitstream to pass along the color space
  type defined in BT.2020(Rec.2020)

  please note this entry does not have any effect on encoding/decoding
  behavior, but allow applications to the pass the information along
  from encoding end to decoding end.

  Change-Id : I4678520e89141ea5e8900f7bd1c0e95b710b7091