Recherche avancée

Sur d’autres sites (16220)

• Opus strict flag

  1er novembre 2018, par KoLiBer

  I’m working on library that uses libavcodec library.

  I try to encode audio frames using opus codec but after avcodec_open2(...) i get this log

  [opus @ 0x2335f30] The encoder 'opus' is experimental but experimental codecs are not enabled, add '-strict -2' if you want to use it.

  also when trying to av_fill_audio_frame(..) i get this error

  Invalid argument

  My code :

  #include <libavcodec></libavcodec>avcodec.h>
  
  #include "libavutil/opt.h"
  
  #include <libavformat></libavformat>avformat.h>
  
  AVFrame *audio_frame = NULL;
  
  AVPacket *audio_packet = NULL;
  
  AVCodecContext *audio_encoder_codec_context = NULL;
  
  AVCodec *audio_encoder_codec = NULL;
  
  
  
  void init(){
  
      audio_frame = av_frame_alloc();
  
      audio_packet = av_packet_alloc();
  
      audio_encoder_codec = avcodec_find_encoder(AV_CODEC_ID_OPUS);
  
      audio_encoder_codec_context = avcodec_alloc_context3(audio_encoder_codec);
  
      audio_encoder_codec_context->time_base = (AVRational){1,25};
  
      audio_encoder_codec_context->sample_rate = audio_sample_rate;
  
      audio_encoder_codec_context->sample_fmt = (enum AVSampleFormat) audio_sample_format == 0 ? AV_SAMPLE_FMT_U8 : AV_SAMPLE_FMT_S16;
  
      audio_encoder_codec_context->channels = 1;
  
      audio_encoder_codec_context->channel_layout = AV_CH_LAYOUT_MONO;
  
      audio_encoder_codec_context->codec_type = AVMEDIA_TYPE_AUDIO;
  
      audio_encoder_codec_context->extradata = NULL;
  
      avcodec_open2(audio_encoder_codec_context,audio_encoder_codec,NULL);
  
      audio_frame_size = av_samples_get_buffer_size(
  
              NULL,
  
              audio_encoder_codec_context->channels,
  
              audio_sample_rate,
  
              audio_encoder_codec_context->sample_fmt,
  
              1
  
      );
  
      audio_frame_buffer = (uint8_t*) av_malloc((audio_frame_size)*sizeof(uint8_t));
  
  }
  
  
  
  void encode(uint8_t *frame,int frame_size,uint8_t **packet, int *packet_size){
  
      memcpy(audio_frame_buffer, frame, (size_t) frame_size);
  
      av_init_packet(audio_packet);
  
      int isFin = 0;
  
      int r = avcodec_fill_audio_frame(audio_frame,audio_encoder_codec_context->channels,audio_encoder_codec_context->sample_fmt,audio_frame_buffer,audio_frame_size,0);
  
      printf("ERROR = %s",av_err2str(r));
  
      avcodec_encode_audio2(audio_encoder_codec_context,audio_packet,audio_frame,&amp;isFin);
  
      *packet = audio_packet->data;
  
      *packet_size = audio_packet->size;
  
  }
  
  
  
  int main(){
  
      init();
  
      uint8_t *packet = NULL;
  
      int size = 0;
  
      encode(".....",44100,packet,size);
  
  
  
  }

• opus : Do not call vector_fmul_scalar on zero samples

  28 octobre 2015, par Kieran Kunhya

  opus : Do not call vector_fmul_scalar on zero samples
  The x86 variant of this function crashes in that specific case.
  CC : libav-devel@libav.org
  

  • [DH] libavcodec/opusdec.c

• opus/matroska : Adding support for DiscardPadding in muxer

  30 août 2013, par Vignesh Venkatasubramanian

  opus/matroska : Adding support for DiscardPadding in muxer
  Support for end trimming Opus in Matroska is implemented by using
  
  the DiscardPadding container element in the Block data. The last
  
  chunk is stored as a Block instead of SimpleBlock and the
  
  trimming information is stored and used to discard samples that
  
  were padded by the Opus codec. This patch adds support for muxing
  
  DiscardPadding element into the container with appropriate value.
  
  Matroska spec for the DiscardPadding element can be found here :
  
  http://matroska.org/technical/specs/index.html#DiscardPadding
  Signed-off-by : Vignesh Venkatasubramanian <vigneshv@google.com>
  
  Signed-off-by : Michael Niedermayer <michaelni@gmx.at>
  

  • [DH] libavformat/matroska.h
  • [DH] libavformat/matroskaenc.c