Recherche avancée

Sur d’autres sites (5567)

• ffmpeg, opus encoded sound in webm does not work [on hold]

  1er août 2017, par Mockarutan

  I’m having trouble getting Opus encoded sound in the webm container to work. I’m using libopus in ffmpeg.

  The file does work in VLC. But not in ffplay or on YouTube. If I take the raw wav data in a wav file and then convert it to Opus/webm with the ffmpeg.exe that comes pre-compiled. It works in VLC, ffplay and YouTube.

  So ffmpeg can obviously do it correctly, I must be doing something wrong in my code.

  The file my code produces : https://drive.google.com/file/d/0B16rIXjPXJCqcU5HVllIYW1iODg/view?usp=sharing

  Edit, More details that I forgot in my frustration : The file can be opened by ffplay and uploaded to youtube (when I interlace it with VP9 video). But the sound is just "ticks", example : https://www.youtube.com/watch?v=j_ShBbuizeo&feature=youtu.be

  I have read though all example codes that I know of from ffmpeg, but all of them is in the old API, not the send/receive api, so a big part of the code does not apply anymore. This codes works with all other Codes I’ve tested, including H.264+AAC in mp4, VP8+Opus in ogg and raw PCM F32LE in wav. I would have gone with VP8+Opus in ogg if the license was as straight forward as the webm license

  I’ve looked though the source for the ffmpeg.exe command line tool and coped everything applicable in to my code base.

  Here is my code : https://pastebin.com/4c999Uz2

  #include "encoder.h"
  
  #include <algorithm>
  
  #include <iterator>
  
  
  
  extern "C"
  
  {
  
  #include "libavcodec/avcodec.h"
  
  #include "libavdevice/avdevice.h"
  
  #include "libavfilter/avfilter.h"
  
  #include "libavformat/avformat.h"
  
  #include "libavutil/avutil.h"
  
  #include "libavutil/imgutils.h"
  
  #include "libswscale/swscale.h"
  
  #include "libswresample/swresample.h"
  
  
  
      enum InfoCodes
  
      {
  
          ENCODED_VIDEO,
  
          ENCODED_AUDIO,
  
          ENCODED_AUDIO_AND_VIDEO,
  
          NOT_ENOUGH_AUDIO_DATA,
  
      };
  
  
  
      enum ErrorCodes
  
      {
  
          RES_NOT_MUL_OF_TWO = -1,
  
          ERROR_FINDING_VID_CODEC = -2,
  
          ERROR_CONTEXT_CREATION = -3,
  
          ERROR_CONTEXT_ALLOCATING = -4,
  
          ERROR_OPENING_VID_CODEC = -5,
  
          ERROR_OPENING_FILE = -6,
  
          ERROR_ALLOCATING_FRAME = -7,
  
          ERROR_ALLOCATING_PIC_BUF = -8,
  
          ERROR_ENCODING_FRAME_SEND = -9,
  
          ERROR_ENCODING_FRAME_RECEIVE = -10,
  
          ERROR_FINDING_AUD_CODEC = -11,
  
          ERROR_OPENING_AUD_CODEC = -12,
  
          ERROR_INIT_RESMPL_CONTEXT = -13,
  
          ERROR_ENCODING_SAMPLES_SEND = -14,
  
          ERROR_ENCODING_SAMPLES_RECEIVE = -15,
  
          ERROR_WRITING_HEADER = -16,
  
          ERROR_INIT_AUDIO_RESPAMLER = -17,
  
      };
  
  
  
      AVCodecID aud_codec_comp_id = AV_CODEC_ID_OPUS;
  
      AVSampleFormat sample_fmt_comp = AV_SAMPLE_FMT_S16;
  
  
  
      AVCodecID aud_codec_id;
  
      AVSampleFormat sample_fmt;
  
  
  
      char* compressed_cont = "webm";
  
  
  
      AVCodec *aud_codec = NULL;
  
      AVCodecContext *aud_codec_context = NULL;
  
      AVFormatContext *outctx;
  
      AVStream *audio_st;
  
      AVFrame *aud_frame;
  
      SwrContext *audio_swr_ctx;
  
      uint8_t **dst_data = NULL;
  
      AVRational conv_time_base;
  
  
  
      int aud_frame_counter;
  
      int dst_nb_samples, src_nb_samples, max_dst_nb_samples;;
  
      int src_rate, dst_rate;
  
      int dst_nb_channels;
  
      int dst_linesize;
  
  
  
      void write_error_to_file(const char* name, int value)
  
      {
  
          int buf_size = 100;
  
          char* buf = new char[buf_size];
  
  
  
          int ret = av_strerror(value, buf, buf_size);
  
  
  
          FILE *f;
  
          fopen_s(&amp;f, name, "w");
  
          if (f != NULL)
  
          {
  
              if (ret != 0)
  
                  fprintf(f, "Error erroring: , \n", ret);
  
              else
  
                  fprintf(f, "Error code: %s\n", buf);
  
  
  
              fclose(f);
  
          }
  
      }
  
  
  
      void write_value_to_file(const char* name, float value)
  
      {
  
          FILE *f;
  
          fopen_s(&amp;f, name, "w");
  
          if (f != NULL)
  
          {
  
              fprintf(f, "Value: %f\n", value);
  
  
  
              fclose(f);
  
          }
  
      }
  
  
  
      char* concat(const char *s1, const char *s2)
  
      {
  
          char *result = (char*)malloc(strlen(s1) + strlen(s2) + 1);
  
  
  
          strcpy(result, s1);
  
          strcat(result, s2);
  
  
  
          return result;
  
      }
  
  
  
      int setup_audio_codec()
  
      {
  
          aud_codec_id = aud_codec_comp_id;
  
          sample_fmt = sample_fmt_comp;
  
  
  
          // Fixup audio codec
  
          if (aud_codec == NULL)
  
          {
  
              aud_codec = avcodec_find_encoder(aud_codec_id);
  
              avcodec_register(aud_codec);
  
          }
  
  
  
          if (!aud_codec)
  
              return ERROR_FINDING_AUD_CODEC;
  
  
  
          return 0;
  
      }
  
  
  
  
  
      int initialize_audio_stream(AVFormatContext *local_outctx, int sample_rate, int per_frame_audio_samples, int audio_bitrate)
  
      {
  
          aud_codec_context = avcodec_alloc_context3(aud_codec);
  
          if (!aud_codec_context)
  
              return ERROR_CONTEXT_CREATION;
  
  
  
          /* select other audio parameters supported by the encoder */
  
          aud_codec_context->bit_rate = audio_bitrate;
  
          aud_codec_context->sample_rate = sample_rate;
  
          aud_codec_context->sample_fmt = sample_fmt;
  
          aud_codec_context->channel_layout = AV_CH_LAYOUT_STEREO;
  
          aud_codec_context->channels = av_get_channel_layout_nb_channels(aud_codec_context->channel_layout);
  
  
  
          aud_codec_context->codec = aud_codec;
  
          aud_codec_context->codec_id = aud_codec_id;
  
  
  
          AVRational time_base;
  
          time_base.num = per_frame_audio_samples;
  
          time_base.den = aud_codec_context->sample_rate;
  
          aud_codec_context->time_base = time_base;
  
  
  
          int ret = avcodec_open2(aud_codec_context, aud_codec, NULL);
  
  
  
          if (ret &lt; 0)
  
              return ERROR_OPENING_AUD_CODEC;
  
  
  
          local_outctx->audio_codec = aud_codec;
  
          local_outctx->audio_codec_id = aud_codec_id;
  
  
  
          audio_st = avformat_new_stream(local_outctx, aud_codec);
  
  
  
          avcodec_parameters_from_context(audio_st->codecpar, aud_codec_context);
  
  
  
          conv_time_base.num = aud_codec_context->frame_size;
  
          conv_time_base.den = aud_codec_context->sample_rate;
  
  
  
          aud_frame = av_frame_alloc();
  
          aud_frame->nb_samples = aud_codec_context->frame_size;
  
          aud_frame->format = aud_codec_context->sample_fmt;
  
          aud_frame->channel_layout = aud_codec_context->channel_layout;
  
          aud_frame->sample_rate = aud_codec_context->sample_rate;
  
  
  
          int buffer_size;
  
          if (aud_codec_context->frame_size == 0)
  
          {
  
              buffer_size = per_frame_audio_samples * 2 * 4;
  
              aud_frame->nb_samples = per_frame_audio_samples;
  
          }
  
          else
  
          {
  
              buffer_size = av_samples_get_buffer_size(NULL, aud_codec_context->channels, aud_codec_context->frame_size,
  
                  aud_codec_context->sample_fmt, 0);
  
          }
  
  
  
          if (av_sample_fmt_is_planar(sample_fmt))
  
              ret = av_frame_get_buffer(aud_frame, buffer_size / 2);
  
          else
  
              ret = av_frame_get_buffer(aud_frame, buffer_size);
  
  
  
          if (!aud_frame || ret &lt; 0)
  
              return ERROR_ALLOCATING_FRAME;
  
  
  
          audio_swr_ctx = swr_alloc();
  
          if (!audio_swr_ctx)
  
              return ERROR_INIT_AUDIO_RESPAMLER;
  
  
  
          /* set options */
  
          av_opt_set_int(audio_swr_ctx, "in_channel_layout", aud_codec_context->channel_layout, 0);
  
          av_opt_set_int(audio_swr_ctx, "in_sample_rate", sample_rate, 0);
  
          av_opt_set_int(audio_swr_ctx, "in_frame_size", per_frame_audio_samples, 0);
  
          av_opt_set_sample_fmt(audio_swr_ctx, "in_sample_fmt", AV_SAMPLE_FMT_FLT, 0);
  
  
  
          av_opt_set_int(audio_swr_ctx, "out_channel_layout", aud_codec_context->channel_layout, 0);
  
          av_opt_set_int(audio_swr_ctx, "out_sample_rate", aud_codec_context->sample_rate, 0);
  
          av_opt_set_int(audio_swr_ctx, "out_frame_size", aud_codec_context->frame_size, 0);
  
          av_opt_set_sample_fmt(audio_swr_ctx, "out_sample_fmt", aud_codec_context->sample_fmt, 0);
  
  
  
          /* initialize the resampling context */
  
          if ((ret = swr_init(audio_swr_ctx)) &lt; 0)
  
          {
  
              return ERROR_INIT_AUDIO_RESPAMLER;
  
          }
  
  
  
          dst_rate = aud_codec_context->sample_rate;
  
          src_rate = sample_rate;
  
  
  
          src_nb_samples = per_frame_audio_samples;
  
          dst_nb_samples = aud_codec_context->frame_size;
  
  
  
          max_dst_nb_samples = av_rescale_rnd(src_nb_samples, dst_rate, src_rate, AV_ROUND_UP);
  
  
  
          dst_nb_channels = av_get_channel_layout_nb_channels(aud_codec_context->channel_layout);
  
  
  
          ret = av_samples_alloc_array_and_samples(&amp;dst_data, &amp;dst_linesize, dst_nb_channels, dst_nb_samples, sample_fmt, 0);
  
  
  
          aud_frame_counter = 0;
  
  
  
          return 0;
  
      }
  
  
  
      int initialize_audio_only_encoding(int sample_rate, int per_frame_audio_samples, int audio_bitrate, const char *filename)
  
      {
  
          int ret;
  
  
  
          avcodec_register_all();
  
          av_register_all();
  
  
  
          outctx = avformat_alloc_context();
  
  
  
          char* full_filename;
  
          char* with_dot = concat(filename, ".");
  
          full_filename = concat(with_dot, compressed_cont);
  
          ret = avformat_alloc_output_context2(&amp;outctx, NULL, compressed_cont, full_filename);
  
  
  
          free(with_dot);
  
  
  
          if (ret &lt; 0)
  
          {
  
              free(full_filename);
  
              return ERROR_CONTEXT_CREATION;
  
          }
  
  
  
          ret = setup_audio_codec();
  
          if (ret &lt; 0)
  
              return ret;
  
  
  
          // Setup Audio
  
          ret = initialize_audio_stream(outctx, sample_rate, per_frame_audio_samples, audio_bitrate);
  
          if (ret &lt; 0)
  
              return ret;
  
  
  
          av_dump_format(outctx, 0, full_filename, 1);
  
  
  
          if (!(outctx->oformat->flags &amp; AVFMT_NOFILE))
  
          {
  
              if (avio_open(&amp;outctx->pb, full_filename, AVIO_FLAG_WRITE) &lt; 0)
  
              {
  
                  free(full_filename);
  
                  return ERROR_OPENING_FILE;
  
              }
  
          }
  
  
  
  
  
          free(full_filename);
  
  
  
          ret = avformat_write_header(outctx, NULL);
  
          if (ret &lt; 0)
  
              return ERROR_WRITING_HEADER;
  
  
  
          return 0;
  
      }
  
  
  
      int process_encode_loop(AVFormatContext *local_outctx, AVCodecContext *codec_context, AVStream *stream, AVRational time_base, bool flush)
  
      {
  
          int ret;
  
  
  
          AVPacket pkt;
  
          av_init_packet(&amp;pkt);
  
          pkt.data = NULL;
  
          pkt.size = 0;
  
  
  
          while (true)
  
          {
  
              ret = avcodec_receive_packet(codec_context, &amp;pkt);
  
              if (!ret)
  
              {
  
                  pkt.stream_index = stream->index;
  
                  av_packet_rescale_ts(&amp;pkt, time_base, stream->time_base);
  
                  av_interleaved_write_frame(local_outctx, &amp;pkt);
  
  
  
                  av_packet_unref(&amp;pkt);
  
              }
  
  
  
              if (ret == AVERROR(EAGAIN))
  
                  break;
  
              else if (ret == AVERROR_EOF)
  
                  break;
  
              else if (ret &lt; 0)
  
                  return ERROR_ENCODING_FRAME_RECEIVE;
  
              else if (flush == false)
  
                  break;
  
          }
  
  
  
          return 0;
  
      }
  
  
  
      int write_audio_frame(float_t *aud_sample)
  
      {
  
          int ret;
  
          if (dst_nb_samples > max_dst_nb_samples)
  
          {
  
              av_free(&amp;aud_frame->data[0]);
  
              ret = av_samples_alloc(aud_frame->data, &amp;dst_linesize, dst_nb_channels, dst_nb_samples, sample_fmt, 1);
  
              if (ret &lt; 0)
  
                  return ERROR_INIT_AUDIO_RESPAMLER;
  
  
  
              max_dst_nb_samples = dst_nb_samples;
  
          }
  
  
  
          ret = swr_convert(audio_swr_ctx, dst_data, dst_nb_samples, (const uint8_t **)&amp;aud_sample, src_nb_samples);
  
          if (ret &lt; 0)
  
          {
  
              return ERROR_INIT_AUDIO_RESPAMLER;
  
          }
  
  
  
          aud_frame->data[0] = (uint8_t*)dst_data[0];
  
          aud_frame->extended_data[0] = (uint8_t*)dst_data[0];
  
  
  
          aud_frame->pts = aud_frame_counter++;
  
  
  
          ret = avcodec_send_frame(aud_codec_context, aud_frame);
  
  
  
          if (ret &lt; 0)
  
          {
  
              write_error_to_file("AVERROR.txt", ret);
  
  
  
              return ERROR_ENCODING_SAMPLES_SEND;
  
          }
  
  
  
          ret = process_encode_loop(outctx, aud_codec_context, audio_st, conv_time_base, false);
  
  
  
          if (ret &lt; 0)
  
              return ERROR_ENCODING_SAMPLES_RECEIVE;
  
  
  
          return ENCODED_AUDIO;
  
      }
  
  
  
      int finish_audio_encoding()
  
      {
  
          //fclose(dst_file);
  
  
  
          AVPacket pkt;
  
          av_init_packet(&amp;pkt);
  
          pkt.data = NULL;
  
          pkt.size = 0;
  
  
  
          fflush(stdout);
  
  
  
          int ret = avcodec_send_frame(aud_codec_context, NULL);
  
          if (ret &lt; 0)
  
              return ERROR_ENCODING_FRAME_SEND;
  
  
  
          while (true)
  
          {
  
              ret = avcodec_receive_packet(aud_codec_context, &amp;pkt);
  
              if (!ret)
  
              {
  
                  if (pkt.pts != AV_NOPTS_VALUE)
  
                      pkt.pts = av_rescale_q(pkt.pts, aud_codec_context->time_base, audio_st->time_base);
  
                  if (pkt.dts != AV_NOPTS_VALUE)
  
                      pkt.dts = av_rescale_q(pkt.dts, aud_codec_context->time_base, audio_st->time_base);
  
  
  
                  av_write_frame(outctx, &amp;pkt);
  
                  av_packet_unref(&amp;pkt);
  
              }
  
              if (ret == -AVERROR(AVERROR_EOF))
  
                  break;
  
              else if (ret &lt; 0)
  
                  return ERROR_ENCODING_FRAME_RECEIVE;
  
          }
  
  
  
          av_write_trailer(outctx);
  
  
  
          return 0;
  
      }
  
  
  
      void cleanup()
  
      {
  
          if (aud_frame)
  
          {
  
              av_frame_free(&amp;aud_frame);
  
          }
  
          if (outctx)
  
          {
  
              for (int i = 0; i &lt; outctx->nb_streams; i++)
  
                  av_freep(&amp;outctx->streams[i]);
  
  
  
              avio_close(outctx->pb);
  
              av_free(outctx);
  
          }
  
  
  
          if (aud_codec_context)
  
          {
  
              avcodec_close(aud_codec_context);
  
              av_free(aud_codec_context);
  
          }
  
      }
  
  
  
      void fill_samples(float_t *dst, int nb_samples, int nb_channels, int sample_rate, float_t *t)
  
      {
  
          int i, j;
  
          float_t tincr = 1.0 / sample_rate;
  
          const float_t c = 2 * M_PI * 440.0;
  
          /* generate sin tone with 440Hz frequency and duplicated channels */
  
          for (i = 0; i &lt; nb_samples; i++) {
  
              *dst = sin(c * *t);
  
              for (j = 1; j &lt; nb_channels; j++)
  
                  dst[j] = dst[0];
  
              dst += nb_channels;
  
              *t += tincr;
  
          }
  
      }
  
  
  
      int main()
  
      {
  
          int width = 1920;
  
          int height = 1080;
  
          int frame_rate = 30;
  
          float t = 0, tincr = 0, tincr2 = 0;
  
  
  
          int sec = 20;
  
  
  
          int tot = 3 * width * height;
  
          uint8_t* rgb24Data = new uint8_t[tot];
  
          float_t** aud_samples_planar;
  
          int16_t** aud_samples_s16;
  
          int src_samples_linesize;
  
          int src_nb_samples = 1024;
  
          int src_channels = 2;
  
          int sample_rate = 48000;
  
  
  
          uint8_t **src_data = NULL;
  
  
  
          int ret;
  
  
  
          initialize_audio_only_encoding(48000, src_nb_samples, 256000, "only_sound");
  
  
  
          ret = av_samples_alloc_array_and_samples(&amp;src_data, &amp;src_samples_linesize, src_channels,
  
              src_nb_samples, AV_SAMPLE_FMT_FLT, 0);
  
  
  
  
  
          for (size_t i = 0; i &lt; frame_rate * sec; i++)
  
          {
  
              fill_samples((float *)src_data[0], src_nb_samples, src_channels, sample_rate, &amp;t);
  
              write_audio_frame((float *)src_data[0]);
  
          }
  
  
  
          finish_audio_encoding();
  
  
  
          cleanup();
  
  
  
          return 0;
  
      }
  
  }
  
  </iterator></algorithm>

  Any any suggestion is appreciated, thanks in Advance !

  Some things I tried :

  I’ve looked at the header with the "MediaInfo" app built in to MVKTool :
  https://i.gyazo.com/3b29b41629a28bd526bf7637ce3f2601.png
  It all looks fine to me.

  I’ve also inspected the raw EBML file with EBML-Viewer (https://code.google.com/archive/p/ebml-viewer/) and in there I can se some difference between the files ;

  My file : https://i.gyazo.com/6fa8c540a2698a8a4d3421d363aede0a.png
  File produced with ffmpeg.exe : https://i.gyazo.com/04d60e64ff3c3040ea83e98cdf507530.png

  In my file it’s "Cluster" -> "BlockGroup" -> "Block", " ?"
  In the other it’s just "Cluster" -> "SimpleBlock"
  And in the webm specs, it says both are supported (https://www.webmproject.org/docs/container/)

  But I do not know much about these specific things, just looking for anything.

• ffmpeg ; opus encoded sound in webm does not work with ffplay or YouTube, only VLC [on hold]

  2 août 2017, par Mockarutan

  I’m having trouble getting Opus encoded sound in the webm container to work. I’m using libopus in ffmpeg.

  The file does work in VLC. But not in ffplay or on YouTube. If I take the raw wav data in a wav file and then convert it to Opus/webm with the ffmpeg.exe that comes pre-compiled. It works in VLC, ffplay and YouTube.

  So ffmpeg can obviously do it correctly, I must be doing something wrong in my code.

  The file my code produces : https://drive.google.com/file/d/0B16rIXjPXJCqcU5HVllIYW1iODg/view?usp=sharing

  Edit, More details that I forgot in my frustration : The file can be opened by ffplay and uploaded to youtube (when I interlace it with VP9 video). But the sound is just "ticks", example : https://www.youtube.com/watch?v=j_ShBbuizeo&feature=youtu.be

  I have read though all example codes that I know of from ffmpeg, but all of them is in the old API, not the send/receive api, so a big part of the code does not apply anymore. This code works with all other Codes I’ve tested, including H.264+AAC in mp4, VP8+Opus in ogg and raw PCM F32LE in wav. I would have gone with VP8+Opus in ogg if the license was as straight forward as the webm license

  I’ve looked though the source for the ffmpeg.exe command line tool and coped everything applicable in to my code base.

  (Edit 3, reduced the code as much as I can)
  Here is my code : https://pastebin.com/HTuc0g8K

  Setup :

  int initialize(int sample_rate, int per_frame_audio_samples, int audio_bitrate, const char *filename)
  
      {
  
          int ret;
  
  
  
          avcodec_register_all();
  
          av_register_all();
  
  
  
          ret = avformat_alloc_output_context2(&outctx, NULL, "webm", filename);
  
  
  
          if (ret < 0)
  
              return ret;
  
  
  
          aud_codec = avcodec_find_encoder(aud_codec_id);
  
          avcodec_register(aud_codec);
  
  
  
          if (!aud_codec)
  
              return -1;
  
  
  
          // Setup Audio Stream
  
  
  
          aud_codec_context = avcodec_alloc_context3(aud_codec);
  
          if (!aud_codec_context)
  
              return -1;
  
  
  
          /* select other audio parameters supported by the encoder */
  
          aud_codec_context->bit_rate = audio_bitrate;
  
          aud_codec_context->sample_rate = sample_rate;
  
          aud_codec_context->sample_fmt = sample_fmt;
  
          aud_codec_context->channel_layout = AV_CH_LAYOUT_STEREO;
  
          aud_codec_context->channels = av_get_channel_layout_nb_channels(aud_codec_context->channel_layout);
  
  
  
          aud_codec_context->codec = aud_codec;
  
          aud_codec_context->codec_id = aud_codec_id;
  
  
  
  
  
          AVRational time_base;
  
          time_base.num = per_frame_audio_samples;
  
          time_base.den = aud_codec_context->sample_rate;
  
          aud_codec_context->time_base = time_base;
  
  
  
          ret = avcodec_open2(aud_codec_context, aud_codec, NULL);
  
  
  
          if (ret < 0)
  
              return ret;
  
  
  
          outctx->audio_codec = aud_codec;
  
          outctx->audio_codec_id = aud_codec_id;
  
  
  
          audio_st = avformat_new_stream(outctx, aud_codec);
  
  
  
          avcodec_parameters_from_context(audio_st->codecpar, aud_codec_context);
  
  
  
          conv_time_base.num = aud_codec_context->frame_size;
  
          conv_time_base.den = aud_codec_context->sample_rate;
  
  
  
          // Setup audio frame
  
          aud_frame = av_frame_alloc();
  
          aud_frame->nb_samples = aud_codec_context->frame_size;
  
          aud_frame->format = aud_codec_context->sample_fmt;
  
          aud_frame->channel_layout = aud_codec_context->channel_layout;
  
          aud_frame->sample_rate = aud_codec_context->sample_rate;
  
  
  
          int buffer_size;
  
          if (aud_codec_context->frame_size == 0)
  
          {
  
              buffer_size = per_frame_audio_samples * 2 * 4;
  
              aud_frame->nb_samples = per_frame_audio_samples;
  
          }
  
          else
  
          {
  
              buffer_size = av_samples_get_buffer_size(NULL, aud_codec_context->channels, aud_codec_context->frame_size,
  
                  aud_codec_context->sample_fmt, 0);
  
          }
  
  
  
          if (av_sample_fmt_is_planar(sample_fmt))
  
              ret = av_frame_get_buffer(aud_frame, buffer_size / 2);
  
          else
  
              ret = av_frame_get_buffer(aud_frame, buffer_size);
  
  
  
          if (!aud_frame || ret < 0)
  
              return ret;
  
  
  
          // Setup audio resampler
  
  
  
          audio_swr_ctx = swr_alloc();
  
          if (!audio_swr_ctx)
  
              return -1;
  
  
  
          /* set options */
  
          av_opt_set_int(audio_swr_ctx, "in_channel_layout", aud_codec_context->channel_layout, 0);
  
          av_opt_set_int(audio_swr_ctx, "in_sample_rate", sample_rate, 0);
  
          av_opt_set_int(audio_swr_ctx, "in_frame_size", per_frame_audio_samples, 0);
  
          av_opt_set_sample_fmt(audio_swr_ctx, "in_sample_fmt", AV_SAMPLE_FMT_FLT, 0);
  
  
  
          av_opt_set_int(audio_swr_ctx, "out_channel_layout", aud_codec_context->channel_layout, 0);
  
          av_opt_set_int(audio_swr_ctx, "out_sample_rate", aud_codec_context->sample_rate, 0);
  
          av_opt_set_int(audio_swr_ctx, "out_frame_size", aud_codec_context->frame_size, 0);
  
          av_opt_set_sample_fmt(audio_swr_ctx, "out_sample_fmt", aud_codec_context->sample_fmt, 0);
  
  
  
          /* initialize the resampling context */
  
          if ((ret = swr_init(audio_swr_ctx)) < 0)
  
          {
  
              return ret;
  
          }
  
  
  
          dst_rate = aud_codec_context->sample_rate;
  
          src_rate = sample_rate;
  
  
  
          src_nb_samples = per_frame_audio_samples;
  
          dst_nb_samples = aud_codec_context->frame_size;
  
  
  
          max_dst_nb_samples = av_rescale_rnd(src_nb_samples, dst_rate, src_rate, AV_ROUND_UP);
  
  
  
          dst_nb_channels = av_get_channel_layout_nb_channels(aud_codec_context->channel_layout);
  
  
  
          ret = av_samples_alloc_array_and_samples(&dst_data, &dst_linesize, dst_nb_channels, dst_nb_samples, sample_fmt, 0);
  
  
  
          aud_frame_counter = 0;
  
  
  
          if (ret < 0)
  
              return ret;
  
  
  
          av_dump_format(outctx, 0, filename, 1);
  
  
  
          if (!(outctx->oformat->flags & AVFMT_NOFILE))
  
          {
  
              ret = avio_open(&outctx->pb, filename, AVIO_FLAG_WRITE);
  
              if (ret < 0)
  
              {
  
                  return ret;
  
              }
  
          }
  
  
  
          ret = avformat_write_header(outctx, NULL);
  
          if (ret < 0)
  
              return ret;
  
  
  
          return 0;
  
      }

  Encoding and ending :

  int process_encode_loop(AVFormatContext *local_outctx, AVCodecContext *codec_context, AVStream *stream, AVRational time_base, bool flush)
  
      {
  
          int ret;
  
  
  
          AVPacket pkt;
  
          av_init_packet(&pkt);
  
          pkt.data = NULL;
  
          pkt.size = 0;
  
  
  
          while (true)
  
          {
  
              ret = avcodec_receive_packet(codec_context, &pkt);
  
              if (!ret)
  
              {
  
                  pkt.stream_index = stream->index;
  
                  av_packet_rescale_ts(&pkt, time_base, stream->time_base);
  
                  av_interleaved_write_frame(local_outctx, &pkt);
  
  
  
                  av_packet_unref(&pkt);
  
              }
  
  
  
              if (ret == AVERROR(EAGAIN))
  
                  break;
  
              else if (ret == AVERROR_EOF)
  
                  break;
  
              else if (ret < 0)
  
                  return ret;
  
              else if (flush == false)
  
                  break;
  
          }
  
  
  
          return 0;
  
      }
  
  
  
      int write_audio_frame(float_t *aud_sample)
  
      {
  
          int ret;
  
          if (dst_nb_samples > max_dst_nb_samples)
  
          {
  
              av_free(&aud_frame->data[0]);
  
              ret = av_samples_alloc(aud_frame->data, &dst_linesize, dst_nb_channels, dst_nb_samples, sample_fmt, 1);
  
              if (ret < 0)
  
                  return ret;
  
  
  
              max_dst_nb_samples = dst_nb_samples;
  
          }
  
  
  
          ret = swr_convert(audio_swr_ctx, dst_data, dst_nb_samples, (const uint8_t **)&aud_sample, src_nb_samples);
  
          if (ret < 0)
  
          {
  
              return ret;
  
          }
  
  
  
          aud_frame->data[0] = (uint8_t*)dst_data[0];
  
          aud_frame->extended_data[0] = (uint8_t*)dst_data[0];
  
  
  
          aud_frame->pts = aud_frame_counter++;
  
  
  
          ret = avcodec_send_frame(aud_codec_context, aud_frame);
  
  
  
          ret = process_encode_loop(outctx, aud_codec_context, audio_st, conv_time_base, false);
  
  
  
          if (ret < 0)
  
              return ret;
  
  
  
          return 0;
  
      }
  
  
  
      int finish_audio_encoding()
  
      {
  
          int ret = avcodec_send_frame(aud_codec_context, NULL);
  
          if (ret < 0)
  
              return ret;
  
  
  
          ret = process_encode_loop(outctx, aud_codec_context, audio_st, conv_time_base, true);
  
          if (ret < 0)
  
              return ret;
  
  
  
          av_write_trailer(outctx);
  
  
  
          return ret;
  
      }

  Main :

  void fill_samples(float_t *dst, int nb_samples, int nb_channels, int sample_rate, float_t *t)
  
      {
  
          int i, j;
  
          float_t tincr = 1.0 / sample_rate;
  
          const float_t c = 2 * M_PI * 440.0;
  
          /* generate sin tone with 440Hz frequency and duplicated channels */
  
          for (i = 0; i < nb_samples; i++) {
  
              *dst = sin(c * *t);
  
              for (j = 1; j < nb_channels; j++)
  
                  dst[j] = dst[0];
  
              dst += nb_channels;
  
              *t += tincr;
  
          }
  
      }
  
  
  
      int main()
  
      {
  
          int frame_rate = 30;
  
          int sec = 12;
  
          int bit_rate = 192000;
  
          float t = 0;
  
  
  
          int src_samples_linesize;
  
          int src_nb_samples = 1024;
  
          int src_channels = 2;
  
          int sample_rate = 48000;
  
  
  
          uint8_t **src_data = NULL;
  
  
  
          int ret;
  
  
  
          initialize(sample_rate, src_nb_samples, bit_rate, "sound_test.webm");
  
  
  
          ret = av_samples_alloc_array_and_samples(&src_data, &src_samples_linesize, src_channels,
  
              src_nb_samples, AV_SAMPLE_FMT_FLT, 0);
  
  
  
          for (size_t i = 0; i < frame_rate * sec; i++)
  
          {
  
              fill_samples((float *)src_data[0], src_nb_samples, src_channels, sample_rate, &t);
  
              write_audio_frame((float *)src_data[0]);
  
          }
  
          finish_audio_encoding();
  
  
  
          cleanup();
  
  
  
          return 0;
  
      }

  Edit 2, This code reproduces the issue exactly and is fully self contained, if you have the ffmpeg 3.3.x libraries. It’s tried with 3.3.1 and 3.3.2 is the same results.

  So what could I be missing ? I do not think something is wrong with the sample rates or any other specifications, else it would not work in VLC or an ogg file. I do think the audio stream itself if correct, just some part of the header or how the file is formatted (look further down for some EBML inspection) that is not correct.

  As explained earlier, the licence with VP9+Opus in webm is why I have these specifics. And the exact problem is that I want the audio stream produced to work well when I upload it to YouTube.

  Any suggestion is appreciated, thanks in Advance !

  Some other things I’ve tried :

  I’ve looked at the header with the "MediaInfo" app built in to MVKTool :
  https://i.gyazo.com/3b29b41629a28bd526bf7637ce3f2601.png
  It all looks fine to me.

  I’ve also inspected the raw EBML file with EBML-Viewer (https://code.google.com/archive/p/ebml-viewer/) and in there I can se some difference between the files ;

  My file : https://i.gyazo.com/6fa8c540a2698a8a4d3421d363aede0a.png
  File produced with ffmpeg.exe : https://i.gyazo.com/04d60e64ff3c3040ea83e98cdf507530.png

  In my file it’s "Cluster" -> "BlockGroup" -> "Block", " ?"
  In the other it’s just "Cluster" -> "SimpleBlock"
  And in the webm specs, it says both are supported (https://www.webmproject.org/docs/container/)

  But I do not know much about these specific things, just looking for anything.

• Seam carving

  13 février 2008, par Mikko Koppanen — Imagick, PHP stuff

  Today I was reading trough the ImageMagick ChangeLog and noticed an interesting entry. “Add support for liquid rescaling”. I rushed to check the MagickWand API docs and there it was : MagickLiquidRescaleImage ! After about ten minutes of hacking the Imagick support was done. Needless to say ; I was excited

  For those who don’t know what seam carving is check the demo here. More detailed information about the algorithm can be found here : “Seam Carving for Content-Aware Image Resizing” by Shai Avidan and Ariel Shamir

  To use this functionality you need to install at least ImageMagick 6.3.8-2 and liblqr. Remember to pass –with-lqr to ImageMagick configuration line. You can get liblqr here : http://liblqr.wikidot.com/. The Imagick side of the functionality should appear in the CVS today if everything goes as planned.

  Here is a really simple example just to illustrate the results of the operation. The parameters might be far from optimal (didn’t do much testing yet). The original dimensions of image are 500×375 and the resulting size is 500×200.

  Update : the functionality is pending until license issues are solved.

  php

  < view plain text >

  1. < ?php
  2.  
  3. /* Create new object */
  4. $im = new Imagick( ’test.jpg’ ) ;
  5.  
  6. /* Scale down */
  7. $im->liquidRescaleImage( 500, 200, 3, 25 ) ;
  8.  
  9. /* Display */
  10. header( ’Content-Type : image/jpg’ ) ;
  11. echo $im ;
  12.  
  13.  ?>

  The original image by flickr/jennconspiracy

  

  And the result :

  

  Update. On kenrick’s request here is an image which is scaled down to 300×300