Recherche avancée

Sur d’autres sites (12009)

• Revert "avformat/hlsenc : compute video_keyframe_size after write keyframe"

  1er mai 2021, par Steven Liu

  Revert "avformat/hlsenc : compute video_keyframe_size after write keyframe"
  This reverts commit b5ca8f2c66954614d81579082025f580efc0cffc.
  This commit will make new problem about tickets : 9193,9205
  
  It flush data into file with init file context together,
  
  and it can get keyframe size, maybe need more method to get keyframe
  
  size.
  Signed-off-by : Steven Liu <liuqi05@kuaishou.com>
  

  • [DH] libavformat/hlsenc.c

• "Critical error detected c0000374" when running av_packet_unref or av_frame_unref

  15 mai 2021, par Shivang Sharma

  I am trying to read and decode frames which is happening nicely but when its reaching part of un-referencing frame or packet using av_packet_unref and av_frame_unref it is giving error during second frame or third frame sometimes .

  


  Error (Copied from visual studio output window) :

  


  Critical error detected c0000374
Libav.exe has triggered a breakpoint


  


  Here is some code of reading and decoding which is giving error :

  


  void Decoder::streamNextFrame(int type = 0)
{
    while (av_read_frame(this->fileFormatCtx, this->latestpacket) >= 0) {
        if (this->audioDecoder->activeAudioStream != nullptr) {
            if (this->latestpacket->stream_index == this->audioDecoder->activeAudioStream->index) {
                avcodec_send_packet(this->audioDecoder->activeStreamDecoder, this->latestpacket);
                err = avcodec_receive_frame(this->audioDecoder->activeStreamDecoder, this->decodedFrame);
                if (err == AVERROR(EAGAIN)) {
                    av_frame_unref(this->decodedFrame);
                    av_packet_unref(this->latestpacket);
                    continue;
                }

                {
                    int currentIndex = (int)this->audioFrames->size();
                    this->audioFrames->resize((int)this->audioFrames->size() + 1);
                    int nb = 0;
                    this->audioFrames->at(currentIndex).pts = (int)this->decodedFrame->pts;
                    if (this->audioDecoder->activeStreamDecoder->sample_fmt != AV_SAMPLE_FMT_S16) {
                        nb = 2048 * this->audioDecoder->activeStreamDecoder->channels;
                        printf("%i\n", nb);
                        this->audioFrames->at(currentIndex).data.resize(nb);
                        if (!swr_is_initialized(swr)) {

                            swr_alloc_set_opts(swr, this->audioDecoder->activeStreamDecoder->channel_layout, AV_SAMPLE_FMT_S16, this->audioDecoder->activeStreamDecoder->sample_rate, this->audioDecoder->activeStreamDecoder->channel_layout, this->audioDecoder->activeStreamDecoder->sample_fmt, this->audioDecoder->activeStreamDecoder->sample_rate, 0, nullptr);
                            swr_init(swr);
                        }

                        uint8_t* buffer = this->audioFrames->at(currentIndex).data.data();
                        swr_convert(swr, &buffer, nb, (const uint8_t**)this->decodedFrame->extended_data, this->decodedFrame->nb_samples);
                    }
                    else {
                        nb = this->decodedFrame->nb_samples * this->audioDecoder->activeStreamDecoder->channels;
                        this->audioFrames->at(currentIndex).data = std::vector(*this->decodedFrame->extended_data, *this->decodedFrame->extended_data + (uint8_t)nb);
                    }

                    this->audioFrames->at(currentIndex).buffersize = nb;
                }

                if (err == AVERROR_EOF) {
                    this->audioDecoder->streamEnded = true;
                    av_frame_unref(this->decodedFrame);
                    av_packet_unref(this->latestpacket);
                    break;
                }
                else if (err >= 0) {
                    this->audioDecoder->streamEnded = false;
                }

                if (type != 0) {
                    av_packet_unref(this->latestpacket);
                    av_frame_unref(this->decodedFrame);
                    break;
                }
                av_packet_unref(this->latestpacket);
                av_frame_unref(this->decodedFrame);
            }
        }
        else {
            printf("No active audio stream is set\n");
            if(type!=0)
            break;
        }
    }

}


  


  I have removed some of code which was concerning the video and was not giving problem I think.

  


  Some Information about above code :

  


  this->audioFrames is a pointer to vector with following type.&#xA;std::vector<audioframeformat>* "AudioFrameFormat" is struct defined as following&#xA;&#xA;struct AudioFrameFormat {&#xA;        std::vector data = {};&#xA;        int pts = 0;&#xA;        int buffersize = 0;&#xA;    };&#xA;&#xA;&#xA;swr is a private class member allocated in constructor&#xA;</audioframeformat>

  


  Call Stack looks like :

  


  

  


  I am getting from call stack is that I am not taking care of my heap memory.

  


  Can someone please explain where problem is and, why some times it run till third frame and some time till second frame of the audio stream ?

  


  And please tell how can I improve this code.

  


• ffmpeg : "Impossible to convert between the formats" when using NVIDIA GPU hardware acceleration

  24 mai 2021, par MorenoGentili

  I'm using ffmpeg from the command line on Windows 10 and I wanted to gave GPU acceleration a try to improve execution times. A simple cut command like this works fine and its execution time is reduced by 60-70%. Awesome.

  


  ffmpeg -hwaccel cuvid -c:v h264_cuvid -ss 00:00:10 -i in.mp4 -c:v h264_nvenc out.mp4


  


  Now I tried to use the -filter_complex flag to overlay, fade and translate a png image over a video. The working non-GPU enhanced command is this one :

  


  ffmpeg -i in.mp4 -loop 1 -t 75 -i overlay.png -filter_complex "[1:v]fade=t=in:st=30:d=0.3:alpha=1,fade=t=out:st=35.7:d=0.3:alpha=1[png1];[0:v][png1]overlay=x='if(gte(t,30), (t-30)*10, NAN)'" -movflags +faststart out.mp4


  


  Then, I added the GPU-related flags to the command like this.

  


  ffmpeg -hwaccel cuvid -c:v h264_cuvid -i in.mp4 -loop 1 -t 75 -i overlay.png -filter_complex "[1:v]fade=t=in:st=30:d=0.3:alpha=1,fade=t=out:st=35.7:d=0.3:alpha=1[png1];[0:v][png1]overlay=x='if(gte(t,30), 60-tanh((t-30)*30/5)*60, NAN)'" -movflags +faststart -c:v h264_nvenc out.mp4


  


  But it won't work. I get this error.

  


  Impossible to convert between the formats supported by the filter 'graph 0 input from stream 0:0' and the filter 'auto_scaler_0'
Error reinitializing filters!
Failed to inject frame into filter network: Function not implemented
Error while processing the decoded data for stream #0:0


  


  I don't even know what it means. Can I actually run ANY ffmpeg command on the GPU with GPU acceleration ? I've found some information about the hwupload_cuda flag but I'm not sure if I should use it and how. My attempts have failed so far.

  


  Any advice on how I should modify the command to make it work on the GPU ?