Recherche avancée

Sur d’autres sites (6924)

• ffmpeg video unable to play properly in most video players

  7 septembre 2022, par Lars

  I'm trying to set up a node.js app that allows me to download files from a web interface.
I'm using yt-dl to get the video and audio stream and ffmpeg to pipe those streams into an mp4. This works and returns a mp4 file.

  


  Now the only issue I have is that when I play the file through most video players, the video player is unable to seek, or skip through the song. I found somewhere deep down on a forum that means that that means the mp4 headers are not working but that is all I could find.
Here is my code, almost unchanged from this response on another thread.

  


  Can anyone provide a solution for this issue.

  


  ytdl.getInfo(link, options).then(info => {
        audioStream = ytdl.downloadFromInfo(info, { ...options, quality: 'highestaudio' });
        videoStream = ytdl.downloadFromInfo(info, { ...options, quality: 'highestvideo' });
        // create the ffmpeg process for muxing
        ffmpegProcess = cp.spawn(ffmpegPath, [
            // supress non-crucial messages
            '-loglevel', '8', '-hide_banner',
            // input audio and video by pipe
            '-i', 'pipe:3', '-i', 'pipe:4',
            // map audio and video correspondingly
            '-map', '0:a', '-map', '1:v',
            // no need to change the codec
            '-c', 'copy',
            // output mp4 and pipe
            '-f', 'matroska', 'pipe:5'
        ], {
            // no popup window for Windows users
            windowsHide: true,
            stdio: [
                // silence stdin/out, forward stderr,
                'inherit', 'inherit', 'inherit',
                // and pipe audio, video, output
                'pipe', 'pipe', 'pipe'
            ]
        });
        audioStream.pipe(ffmpegProcess.stdio[3]);
        videoStream.pipe(ffmpegProcess.stdio[4]);
        ffmpegProcess.stdio[5].pipe(result);
    });


  


• Re-encode a video keeping the GOP structure of the original video [closed]

  5 avril, par Baltar27

  Is there a way to re-encode a video with ffmpeg keeping the GOP structure of the original file ? That is, if a frame is IDR, I, B or P in the input file, keep the same type in the output file, even if the GOP is variable and/or adaptive (it changes dynamically GOP type Open/Closed, the I period N or the P period M). Encoding in this way allows to maintain the maximum quality as the Open/Closed, I/B/P and GOP length decisions has been taken already by the original encoder.

  


• FFMPEG C++ API audio/video sync, video is longer [closed]

  11 mai 2023, par Gábor Gombor

  I create a video from frames in C++ with a given FPS and supply it to FFMPEG API. I record from an Unreal engine viewport and feed FFMPEG with the images. In this interval I have also an audio track in FLAC which I want sync with the video. When the music ends, I close the video and merge them, but the final video has sync problems, the video is a little bit longer than the audio, so I will have an increasing delay. For example I record 0:55 secs, the audio is ok=same length, but the video from frames will be 0:56 secs.

  


  I think the following code is problematic :

  


  bool MyVideoExporter::writeFrame(OutputStream* oStream, AVFrame* frame, int& framePTS)
{
    auto* packet = oStream->pkt->pkt;
    auto* codecContext = oStream->enc->codecContext;

    frame->pts = framePTS;
    frame->pkt_dts = frame->pts;

    auto ret = avcodec_send_frame(codecContext, frame);
    if (ret >= 0) {
        auto retVal = 0;
        while (retVal >= 0) {
            retVal = avcodec_receive_packet(codecContext, packet);
            if (retVal == AVERROR(EAGAIN) || retVal == AVERROR_EOF) break;
            else if (retVal < 0) {
                return false;
            }

            // rescale to audio, usually 1/44100
            av_packet_rescale_ts(packet, m_audiotimestamp, oStream->st->time_base);
            // rescale to FPS, usually 1/30 or 1/60
            av_packet_rescale_ts(packet, codecContext->time_base, oStream->st->time_base);

            packet->stream_index = oStream->st->index;

            retVal = av_interleaved_write_frame(m_avFormatContext.avFormatContext, packet);
            if (retVal < 0) {
                return false;
            }

            framePTS++;
        }

        return retVal == AVERROR_EOF;
    }

    return false;
}



  


  Any idea what is wrong ?

  


  I tried change the order of av_packet_rescale_ts lines or move the frame increase code to another places, but getting far worse results.