Recherche avancée

Sur d’autres sites (10027)

• x86 : XOP pixel_sad_{x3, x4} high bit-depth

  9 avril 2014, par James Almer

  x86 : XOP pixel_sad_x3, x4 high bit-depth
  

  • [DH] common/pixel.c
  • [DH] common/x86/pixel.h
  • [DH] common/x86/sad16-a.asm

• High latency when reading latest frame from H264 stream with FFmpeg

  22 juin 2022, par massivemoisture

  I have an app in c# that receives a H264 video stream of Android device's screen from scrcpy-server (https://github.com/Genymobile/scrcpy).

  


  I want to continously get the latest image from the stream and send it to a web app for streaming.

  


  I created a Process for ffmpeg and gave it these arguments :

  


  -fflags -nobuffer -flags low_delay -probesize 32 -re -i pipe: -g 1 -f rawvideo -pix_fmt bgr24 -an -sn pipe:

  


  I have two threads. One thread to write to stdin of the FFmpeg process the bytes that I get from scrcpy-server. The other thread to read the stdout.

  


  But the image stream that I get has very high latency, especially if there is not much movement on the device's screen. And the latency just keeps mounting up, from a few seconds behind to a minute and more.

  


  How can I fix this issue ?

  


• Playing sound from a video using FFmpeg and SDL_QueueAudio results in high pitched audio

  26 septembre 2021, par plieblang

  I'm trying to play audio from an mp4 file using SDL2 and FFmpeg, and using SDL_QueueAudio seems much easier than setting up a callback.

  



  All solutions I've found, whether here or in the dranger tutorials, are deprecated or use callbacks. I tried browsing all questions with both the ffmpeg and sdl tags (there aren't many), to no avail. I tried converting the dranger tutorial to use non-deprecated calls but ran into the same problem. I'm using C, FFmpeg 4.1 and SDL 2.0.9.

  



  This is the setup for AVCodecContext and AVCodec :

  



      int audioStream = -1;
    for (i = 0; i < formatContext->nb_streams; i++) {
        if (audioStream < 0 && formatContext->streams[i]->codecpar->codec_type == AVMEDIA_TYPE_AUDIO) {
            audioStream = i;
        }
    }

    AVCodecParameters *audioParams = formatContext->streams[audioStream]->codecpar;

    AVCodec *audioCodec = avcodec_find_decoder(audioParams->codec_id);

    AVCodecContext *audioCodecCtx = avcodec_alloc_context3(NULL);
    avcodec_open2(audioCodecCtx, audioCodec, NULL);

    SDL_Init(SDL_INIT_AUDIO)

    SDL_AudioSpec desired, obtained;
    SDL_zero(desired);
    SDL_zero(obtained);
    desired.freq = audioCodecCtx->sample_rate;
    desired.format = AUDIO_F32SYS;
    desired.channels = audioCodecCtx->channels;
    desired.silence = 0;
    desired.samples = AUDIO_BUFFER_SIZE;

    SDL_AudioDeviceID audioDevice = SDL_OpenAudioDevice(NULL, 0, &desired, &obtained, SDL_AUDIO_ALLOW_ANY_CHANGE);


  



  This is the main packet decoding loop :

  



      while (av_read_frame(formatContext, &packet) >= 0) {
        if (packet.stream_index == audioStream) {
            if (!avcodec_send_packet(audioCodecCtx, &packet)) {
                avcodec_receive_frame(audioCodecCtx, audioFrame);
                SDL_QueueAudio(audioDevice, audioFrame->data[0], audioFrame->linesize[0]);
            }
        }
    }


  



  The audio plays at the correct speed but at a much higher pitch than what it actually is. I would like it to sound the same as in any media player.
  
Edit : I just realized the test video has stereo audio but I'm only queueing audioFrame.data[0], which I assume means I'm only playing one channel. I tried queueing audioFrame.data[1] which has data as well but it did not solve the problem. Am I correct and if so, how do I play both channels ?