Recherche avancée

Sur d’autres sites (8341)

• How do I extract the first frame of a video stream as an image using avcpp ?

  19 mars 2023, par AurelaNYT

  I'm trying to extract the first frame of a video stream and save it as an image using the avcpp library (https://github.com/h4tr3d/avcpp). I have already written some code but I'm not sure how to extract the first frame from the stream and save it to a .JPG file. Can someone help me with this ?

  


  I have done the following for now and it works fine, however I am unable to understand how to go on with the conversion of the frame to an image :

  


  #include "iostream"
#include "avcpp/av.h"
#include "avcpp/codec.h"
#include "avcpp/frame.h"
#include "avcpp/packet.h"
#include "avcpp/codeccontext.h"
#include "avcpp/formatcontext.h"

class OpenVHCR {
public:

    static void getFirstFrame(const std::string& stream_uri, const bool debug_log) {
        av::init();
        av::Codec videoCodec;
        av::Stream videoStream;
        size_t videoStreamID = -1;
        std::error_code errorCode;
        av::FormatContext formatContext;
        av::VideoDecoderContext decoderContext;
        formatContext.openInput(stream_uri);
        if(formatContext.streamsCount() < 1){
            std::cerr << "[I/O] Failed to find a video steam in the URI provided." << std::endl;
        }else{
            if(debug_log){
                std::cout << "[I/O] Successfully found " <<  formatContext.streamsCount() << " streams in the URI provided." << std::endl;
            }
            formatContext.findStreamInfo(errorCode);
            if (errorCode) {
                std::cerr << "[AV] An unexpected error has occurred: " << errorCode.message() << std::endl;
                return;
            }
            for (size_t i = 0; i < formatContext.streamsCount(); ++i) {
                auto st = formatContext.stream(i);
                if (st.mediaType() == AVMEDIA_TYPE_VIDEO) {
                    videoStreamID = i;
                    videoStream = st;
                    break;
                }
            }
            if(!videoStream.isNull() && videoStream.isValid()){
                if(debug_log){
                    std::cout << "[I/O] Successfully opened a valid video stream." << std::endl;
                }
                decoderContext = av::VideoDecoderContext(videoStream);
                videoCodec = av::findDecodingCodec(decoderContext.raw()->codec_id);
                decoderContext.setCodec(videoCodec);
                decoderContext.setRefCountedFrames(true);
                decoderContext.open({{"threads", "1"}}, av::Codec(), errorCode);
                if(errorCode){
                    std::cout << "[AV] Failed to find a supported codec for the stream." << std::endl;
                    return;
                }else{
                    while (av::Packet streamPacket = formatContext.readPacket(errorCode)){
                        if(errorCode){
                            std::cout << "[AV] Failed to read  a packet from the stream." << std::endl;
                            return;
                        }else{
                            std::cout << "[AV] Successfully read a packet from the stream." << std::endl;
                            if(streamPacket.streamIndex() != videoStreamID){
                                continue;
                            }
                            //TODO: Find a way to get the first frame from the stream and save it as an image.
                            av::VideoFrame videoFrame = decoderContext.decode(streamPacket, errorCode);
                            std::cout << "[AV] Successfully received a frame from the video." << std::endl;
                            if (errorCode) {
                                std::cerr << "[AV] An unexpected error has occurred: " << errorCode.message() << std::endl;
                                return;
                            } else if (!videoFrame) {
                                std::cerr << "[AV] The received video frame seems to be empty." << std::endl;
                            }
                        }
                    }

                    formatContext.close();
                    decoderContext.close();
                }
            }else{
                std::cerr << "[I/O] Failed to find open a valid video stream." << std::endl;
            }
        }
    }
};



  


• FFmpeg auto rotate video when copying video stream without re-encoding

  25 février 2019, par Mahfujur Rahman

  I have been trying to convert an mp4 (portrait) file to mkv. As the mkv container support any video and audio stream, I am copying the audio and video stream to the output container. The command I’m using

  ffmpeg -y -i test.mp4 -vcodec copy -acodec copy test.mkv

  File info for test.mp4 is following

  Input #0, mov,mp4,m4a,3gp,3g2,mj2, from 'test.mp4':
  
    Metadata:
  
      major_brand     : mp42
  
      minor_version   : 0
  
      compatible_brands: isommp42
  
      creation_time   : 2019-02-23T11:18:50.000000Z
  
      com.android.version: 8.0.0
  
    Duration: 00:00:25.86, start: 0.000000, bitrate: 12270 kb/s
  
      Stream #0:0(eng): Video: h264 (avc1 / 0x31637661), yuv420p(tv, bt709), 1280x720, 12005 kb/s, SAR 1:1 DAR 16:9, 30 fps, 30 tbr, 90k tbn, 180k tbc (default)
  
      Metadata:
  
        rotate          : 90
  
        creation_time   : 2019-02-23T11:18:50.000000Z
  
        handler_name    : VideoHandle
  
      Side data:
  
        displaymatrix: rotation of -90.00 degrees
  
      Stream #0:1(eng): Audio: aac (mp4a / 0x6134706D), 48000 Hz, stereo, fltp, 256 kb/s (default)
  
      Metadata:
  
        creation_time   : 2019-02-23T11:18:50.000000Z
  
        handler_name    : SoundHandle

  Now, the problem is that the output video I get from ffmpeg is a 90 degree counter clockwise rotated video. The reason I believe is the following info being removed from output file,

  Side data:
displaymatrix: rotation of -90.00 degrees

  File info for rotated output file test.mkv

  Input #0, matroska,webm, from 'test.mkv':
  
    Metadata:
  
      MAJOR_BRAND     : mp42
  
      MINOR_VERSION   : 0
  
      COMPATIBLE_BRANDS: isommp42
  
      COM.ANDROID.VERSION: 8.0.0
  
      ENCODER         : Lavf58.12.100
  
    Duration: 00:00:25.87, start: 0.000000, bitrate: 12265 kb/s
  
      Stream #0:0(eng): Video: h264, yuv420p(tv, bt709, progressive), 1280x720, SAR 1:1 DAR 16:9, 30 fps, 30 tbr, 1k tbn, 2k tbc (default)
  
      Metadata:
  
        ROTATE          : 90
  
        HANDLER_NAME    : VideoHandle
  
        DURATION        : 00:00:25.866000000
  
      Stream #0:1(eng): Audio: aac, 48000 Hz, stereo, fltp (default)
  
      Metadata:
  
        HANDLER_NAME    : SoundHandle
  
        DURATION        : 00:00:25.813000000

  But when I convert this rotated output test.mkv to mp4 again, I get the portrait file. The displaymatrix side data appears again in the file info.

  Converting the same mp4 file to m4v by copying the steam works fine.

  So, this whole thing is very much confusing to me.

  Now my question is, how can I convert the mp4 file to mkv without re-encoding and avoid getting rotated output ?

  In this post they solved it for c++. I am working on android and using ffmpeg android wrapper to use the ffmpeg library. Is there any ffmpeg flag to handle this situation ?

• libavutil : introduce AVFilmGrainParams side data

  12 novembre 2020, par Lynne

  libavutil : introduce AVFilmGrainParams side data
  This patch introduces a new frame side data type AVFilmGrainParams for use
  
  with video codecs which support it.
  It can save a lot of memory used for duplicate processed reference frames and
  
  reduce copies when applying film grain during presentation.
  

  • [DH] doc/APIchanges
  • [DH] libavutil/Makefile
  • [DH] libavutil/film_grain_params.c
  • [DH] libavutil/film_grain_params.h
  • [DH] libavutil/frame.c
  • [DH] libavutil/frame.h
  • [DH] libavutil/version.h