Recherche avancée

Sur d’autres sites (6332)

• FFMpeg C++ encoder fails with hardware acceleration quick sync intel

  28 août 2022, par teals

  I am trying to debug an issue I have with a simple C++ encoder using FFMmpeg. The following code below works correctly on these other hardware acceleration systems/platforms :

  


  

  • Mac/VideoToolbox,

  


  • Cuda/NVEnc

  


  • Raspberry pi/h264_v4l2m2m.

  


  


  It fails on intel quick sync at avcodec_open2. I'm using Ubuntu 22.04. I built FFMpeg from source with the Intel Media SDK installed and configured. I verified that the QuickSync install works. I also have a corresponding decoder implementation that works correctly using intel quick sync. So I know everything properly installed. I am also running this on an 11th Gen Intel CPU.

  


  [hevc_qsv @ 0x564b0d34f0c0] Low power mode is unsupported (This sometimes shows up)
[hevc_qsv @ 0x563691234000] some encoding parameters are not supported by the QSV runtime. Please double check the input parameters.
Error initializing output stream 0:0 -- Error while opening encoder for output stream #0:0 - maybe incorrect parameters such as bit_rate, rate, width or height


  


  This is the only error that I get and I haven't found any helpful resources. I tried playing around with different parameters but nothing seems to work. Any help or insight would be extremely helpful.

  


  bool VideoEncoder::create() {&#xA;    char const* outfile = filePath.c_str();&#xA;&#xA;     // open output format context&#xA;    int ret = avformat_alloc_output_context2(&amp;ofctx, nullptr, nullptr, outfile);&#xA;    if (ret &lt; 0) {&#xA;        std::cerr &lt;&lt; "fail to avformat_alloc_output_context2(" &lt;&lt; outfile &lt;&lt; "): ret=" &lt;&lt; ret;&#xA;        return false;&#xA;    }&#xA;&#xA;    // create new video stream&#xA;    const AVCodec* codec = avcodec_find_encoder_by_name(encoderName.c_str());&#xA;    vstrm = avformat_new_stream(ofctx, codec);&#xA;    if (!vstrm) {&#xA;        std::cerr &lt;&lt; "fail to avformat_new_stream";&#xA;        return false;&#xA;    }&#xA;&#xA;    // open video encoder&#xA;    cctx = avcodec_alloc_context3(codec);&#xA;    if (!vstrm) {&#xA;        std::cerr &lt;&lt; "fail to avcodec_alloc_context3";&#xA;        return false;&#xA;    }&#xA;    const AVRational dst_fps = {fps, 1};&#xA;    cctx->width = width;&#xA;    cctx->height = height;&#xA;    if(pixel_format != AV_PIX_FMT_NONE) {&#xA;        cctx->pix_fmt = pixel_format;&#xA;    }&#xA;    else {&#xA;        cctx->pix_fmt = codec->pix_fmts[0];&#xA;    }&#xA;    cctx->time_base = av_inv_q(dst_fps);&#xA;    cctx->framerate = dst_fps;&#xA;    cctx->bit_rate = bitrate * 1000000;&#xA;&#xA;&#xA;    AVDictionary* options = nullptr;&#xA;    if(gpu_id >= 0) {&#xA;        av_dict_set_int(&amp;options, "gpu", gpu_id, 0);&#xA;    }&#xA;&#xA;    &#xA;    if (ofctx->oformat->flags &amp; AVFMT_GLOBALHEADER)&#xA;        cctx->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;&#xA;    ret = avcodec_open2(cctx, codec, &amp;options);&#xA;    if (ret &lt; 0) {&#xA;        std::cerr &lt;&lt; "fail to avcodec_open2: ret=" &lt;&lt; ret;&#xA;        return false;&#xA;    }&#xA;    avcodec_parameters_from_context(vstrm->codecpar, cctx);&#xA;&#xA;    //For mac/quicktime we need to add hvc1 tag.&#xA;    if (encoderName.find("hevc") != std::string::npos) {&#xA;        vstrm->codecpar->codec_tag = MKTAG(&#x27;h&#x27;, &#x27;v&#x27;, &#x27;c&#x27;, &#x27;1&#x27;);&#xA;    }&#xA;&#xA;    /*&#xA;    std::cout&#xA;        &lt;&lt; "outfile: " &lt;&lt; outfile &lt;&lt; "\n"&#xA;        &lt;&lt; "format:  " &lt;&lt; ofctx->oformat->name &lt;&lt; "\n"&#xA;        &lt;&lt; "vcodec:  " &lt;&lt; codec->name &lt;&lt; "\n"&#xA;        &lt;&lt; "size:    " &lt;&lt; width &lt;&lt; &#x27;x&#x27; &lt;&lt; height &lt;&lt; "\n"&#xA;        &lt;&lt; "fps:     " &lt;&lt; av_q2d(cctx->framerate) &lt;&lt; "\n"&#xA;        &lt;&lt; "pixfmt:  " &lt;&lt; av_get_pix_fmt_name(cctx->pix_fmt) &lt;&lt; "\n"&#xA;        &lt;&lt; std::flush;*/&#xA;&#xA;    // initialize sample scaler&#xA;    swsCtx = sws_getContext(&#xA;        width, height, AV_PIX_FMT_BGR24,&#xA;        width, height, cctx->pix_fmt,&#xA;        SWS_BILINEAR, nullptr, nullptr, nullptr);&#xA;    if (!swsCtx) {&#xA;        std::cerr &lt;&lt; "fail to sws_getContext";&#xA;        return false;&#xA;    }&#xA;&#xA;    // allocate frame buffer for encoding&#xA;    frame = av_frame_alloc();&#xA;    frame->width = width;&#xA;    frame->height = height;&#xA;    frame->format = static_cast<int>(cctx->pix_fmt);&#xA;    ret = av_frame_get_buffer(frame, 32);&#xA;    if (ret &lt; 0) {&#xA;        std::cerr &lt;&lt; "fail to av_frame_get_buffer: ret=" &lt;&lt; ret;&#xA;        return false;&#xA;    }&#xA;&#xA;    // allocate packet to retrive encoded frame&#xA;    pkt = av_packet_alloc();&#xA;    // open output IO context&#xA;    ret = avio_open2(&amp;ofctx->pb, outfile, AVIO_FLAG_WRITE, nullptr, nullptr);&#xA;    if (ret &lt; 0) {&#xA;        std::cerr &lt;&lt; "fail to avio_open2: ret=" &lt;&lt; ret;&#xA;        return false;&#xA;    }&#xA;&#xA;    // write media container header (if any)&#xA;    ret = avformat_write_header(ofctx, nullptr);&#xA;    if (ret &lt; 0) {&#xA;        std::cerr &lt;&lt; "fail to avformat_write_header: ret=" &lt;&lt; ret;&#xA;        return false;&#xA;    }&#xA;&#xA;    return true;&#xA;}&#xA;</int>

  


• Powerful Video Analytics and Audio Analytics for Piwik

  10 novembre 2016, par InnoCraft — Plugins, Press Releases

  Over the years, one of the most frequently requested feature by users was to be able to measure how videos and audios are watched and engaged with on your website. We are finally able to announce that it is here ! We are very excited to launch Media Analytics, which will help you understand and grow your audience.

  This article is a showcase of the new powerful video and audio analytics product built for Piwik.

  Why media analytics ?

  We all love media content such as videos as they can make our experiences on websites and apps so much more interesting. A growing number of websites now utilize media files in one form or another : a video presentation of a product or service, a video tutorial teaching you how to do something or interviews with key speakers. Also many creators and distributors are publishing audio files such as podcasts or music songs, and even broadcasting live video events such as music concerts or an entire conference online.

  Whenever you publish videos or audio media on your websites or applications, Media Analytics provides you with clear insights on how your audience interacts with your content. It helps you see what content works and why – so you can better understand and further grow your business !

  

  Valuable insights in Real time

  

  See where your audience comes from.

  How will Media Analytics help me grow ?

  • Better understand your audience : who are the users playing videos and for how long, how often, and where have they dropped off.
  • Gain quick insights into how interaction with your media changes over time with easy to use graphs and report overviews.
  • Get closer to your users by seeing every action of your visitors before and after they utilized your media.
  • View valuable insights in Real time : ‘most popular content right now’, your real time audience map, and more.
  • See where your audience comes from. Drill down right from continents to specifics such as cities.
  • Share and export media analytics reports with your colleagues by creating custom email reports.
  • Video and audio players are supported either automatically (for Youtube, Vimeo, HTML5…) or via a simple custom player integration.
  • No data limit and 100% privacy and data ownership.

  Best of all, it is easy to use and understand, and integrates perfectly with Piwik. Media Analytics complements other reports to give you a 360 degree view of how your users engage with your content.

  Learn more on the official website : www.media-analytics.net

  How do I get Media Analytics ?

  All premium plugins come with our 14 day money back guarantee and 1-click installation & updates. Customers get all product updates for free.

  Learn more about Media Analytics Get it now on the Marketplace

  Media Analytics is available for purchase and download on the Marketplace.

  If you are not using Piwik yet, you can also signup for a free trial of Piwik Cloud (including Media Analytics !).

  Have a question about this product ? Get in touch.

• What's wrong with my use of timestamps/timebases for frame seeking/reading using libav (ffmpeg) ?

  17 septembre 2013, par mtree

  So I want to grab a frame from a video at a specific time using libav for the use as a thumbnail.

  What I'm using is the following code. It compiles and works fine (in regards to retrieving a picture at all), yet I'm having a hard time getting it to retrieve the right picture.

  I simply can't get my head around the all but clear logic behind libav's apparent use of multiple time-bases per video. Specifically figuring out which functions expect/return which type of time-base.

  The docs were of basically no help whatsoever, unfortunately. SO to the rescue ?

  #define ABORT(x) do {fprintf(stderr, x); exit(1);} while(0)
  
  
  
  av_register_all();
  
  
  
  AVFormatContext *format_context = ...;
  
  AVCodec *codec = ...;
  
  AVStream *stream = ...;
  
  AVCodecContext *codec_context = ...;
  
  int stream_index = ...;
  
  
  
  // open codec_context, etc.
  
  
  
  AVRational stream_time_base = stream->time_base;
  
  AVRational codec_time_base = codec_context->time_base;
  
  
  
  printf("stream_time_base: %d / %d = %.5f\n", stream_time_base.num, stream_time_base.den, av_q2d(stream_time_base));
  
  printf("codec_time_base: %d / %d = %.5f\n\n", codec_time_base.num, codec_time_base.den, av_q2d(codec_time_base));
  
  
  
  AVFrame *frame = avcodec_alloc_frame();
  
  
  
  printf("duration: %lld @ %d/sec (%.2f sec)\n", format_context->duration, AV_TIME_BASE, (double)format_context->duration / AV_TIME_BASE);
  
  printf("duration: %lld @ %d/sec (stream time base)\n\n", format_context->duration / AV_TIME_BASE * stream_time_base.den, stream_time_base.den);
  
  printf("duration: %lld @ %d/sec (codec time base)\n", format_context->duration / AV_TIME_BASE * codec_time_base.den, codec_time_base.den);
  
  
  
  double request_time = 10.0; // 10 seconds. Video's total duration is ~20sec
  
  int64_t request_timestamp = request_time / av_q2d(stream_time_base);
  
  printf("requested: %.2f (sec)\t-> %2lld (pts)\n", request_time, request_timestamp);
  
  
  
  av_seek_frame(format_context, stream_index, request_timestamp, 0);
  
  
  
  AVPacket packet;
  
  int frame_finished;
  
  do {
  
      if (av_read_frame(format_context, &packet) < 0) {
  
          break;
  
      } else if (packet.stream_index != stream_index) {
  
          av_free_packet(&packet);
  
          continue;
  
      }
  
      avcodec_decode_video2(codec_context, frame, &frame_finished, &packet);
  
  } while (!frame_finished);
  
  
  
  // do something with frame
  
  
  
  int64_t received_timestamp = frame->pkt_pts;
  
  double received_time = received_timestamp * av_q2d(stream_time_base);
  
  printf("received:  %.2f (sec)\t-> %2lld (pts)\n\n", received_time, received_timestamp);

  Running this with a test movie file I get this output :

      stream_time_base: 1 / 30000 = 0.00003
  
      codec_time_base: 50 / 2997 = 0.01668
  
  
  
      duration: 20062041 @ 1000000/sec (20.06 sec)
  
      duration: 600000 @ 30000/sec (stream time base)
  
      duration: 59940 @ 2997/sec (codec time base)
  
  
  
      requested: 10.00 (sec)  -> 300000 (pts)
  
      received:  0.07 (sec)   -> 2002 (pts)

  The times don't match. What's going on here ? What am I doing wrong ?

  ---

  While searching for clues I stumbled upon this this statement from the libav-users mailing list…

  [...] packet PTS/DTS are in units of the format context's time_base,
  where the AVFrame->pts value is in units of the codec context's time_base.

  In other words, the container can have (and usually does) a different
  time_base than the codec. Most libav players don't bother using the
  codec's time_base or pts since not all codecs have one, but most
  containers do. (This is why the dranger tutorial says to ignore AVFrame->pts)

  …which confused me even more, given that I couldn't find any such mention in the official docs.

  Anyway, I replaced…

  double received_time = received_timestamp * av_q2d(stream_time_base);

  …with…

  double received_time = received_timestamp * av_q2d(codec_time_base);

  …and the output changed to this…

  ...
  
  
  
  requested: 10.00 (sec)  -> 300000 (pts)
  
  received:  33.40 (sec)  -> 2002 (pts)

  Still no match. What's wrong ?