Recherche avancée

Sur d’autres sites (7973)

• Chrome times out on streaming FFMPEG output from ASP.NET Web Api

  3 août 2014, par Hayden McAfee

  I’ve got a unique problem here !

  UPDATE 2 So it turns out the development below is FALSE, the inconsistency of the bug made it seem like not closing the stream made it work... but in fact the same issue persists !

  UPDATE Interesting development ; if I comment out ffmpegBufferedIn.Close(); below, the entire stream always goes through fine... the request just never ends. What could be going on here ?

  I’m writing a web service that stores audio files in Azure Blob Storage, and converts them to MP3 live when requested through my ASP.NET Web API endpoint. I accomplish this by using ’DownloadToStream’ via the Azure Storage API, feeding that stream through the STDIN of an FFMPEG process, and sending the STDOUT stream as the request response.

  The block of code that does this looks like this :

  public HttpResponseMessage Get(Guid songid)
  
  {
  
      // This could take awhile.
  
      HttpContext.Current.Server.ScriptTimeout = 600;
  
  
  
      Process ffmpeg = new Process();
  
      ProcessStartInfo startinfo = new ProcessStartInfo(HostingEnvironment.MapPath("~/App_Data/executables/ffmpeg.exe"), "-i - -vn -ar 44100 -ac 2 -ab 192k -f mp3 - ");
  
      startinfo.RedirectStandardError = true;
  
      startinfo.RedirectStandardOutput = true;
  
      startinfo.RedirectStandardInput = true;
  
      startinfo.UseShellExecute = false;
  
      startinfo.CreateNoWindow = true;
  
      ffmpeg.StartInfo = startinfo;
  
      ffmpeg.ErrorDataReceived += ffmpeg_ErrorDataReceived;
  
  
  
      // Our response is a stream
  
      var response = Request.CreateResponse();
  
      response.StatusCode = HttpStatusCode.OK;
  
  
  
      // Retrieve storage account from connection string.
  
      CloudStorageAccount storageAccount = CloudStorageAccount.Parse(
  
          CloudConfigurationManager.GetSetting("StorageConnectionString"));
  
  
  
      // Create the blob client.
  
      CloudBlobClient blobClient = storageAccount.CreateCloudBlobClient();
  
  
  
      // Retrieve reference to a previously created container.
  
      CloudBlobContainer container = blobClient.GetContainerReference("songs");
  
  
  
      // Retrieve reference to a blob
  
      CloudBlockBlob blockBlob = container.GetBlockBlobReference(songid.ToString());
  
  
  
      ffmpeg.Start();
  
      ffmpeg.BeginErrorReadLine();
  
  
  
      // Buffer the streams
  
      var ffmpegBufferedIn = new BufferedStream(ffmpeg.StandardInput.BaseStream);
  
      var ffmpegBufferedOut = new BufferedStream(ffmpeg.StandardOutput.BaseStream);
  
  
  
      blockBlob.DownloadToStreamAsync(ffmpegBufferedIn).ContinueWith((t) => {
  
          ffmpegBufferedIn.Flush();
  
          ffmpegBufferedIn.Close();
  
      });
  
  
  
      response.Content = new StreamContent(ffmpegBufferedOut);
  
      response.Content.Headers.ContentType = new MediaTypeHeaderValue("audio/mpeg");
  
  
  
      System.Diagnostics.Debug.WriteLine("Returned response.");
  
      return response;
  
  }

  This works quite well in all browsers - all except for Chrome, which has an interesting way of buffering audio streams. Chrome will buffer the first 2 megabytes of a stream, then keep the connection open and wait until the user gets closer to playing the next segment of a file before consuming the rest of the stream. This should be fine - and for some songs it is. For others, I get this :

  

  At first I thought this was due to some kind of timeout - But it happens at a different time and size for each file. It is consistent within about 15 seconds on the same songs, however. The output on the server side is normal - no exceptions thrown, and FFMpeg finishes encoding the song successfully.

  Here’s the server-side output of the above request :

  ffmpeg version N-64919-ga613257 Copyright (c) 2000-2014 the FFmpeg developers
  
    built on Jul 23 2014 00:27:32 with gcc 4.8.3 (GCC)
  
    configuration: --enable-gpl --enable-version3 --disable-w32threads --enable-avisynth --enable-bzlib --enable-fontconfig --enable-frei0r --enable-gnutls --enable-iconv --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libfreetype --enable-libgme --enable-libgsm --enable-libilbc --enable-libmodplug --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenjpeg --enable-libopus --enable-librtmp --enable-libschroedinger --enable-libsoxr --enable-libspeex --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvo-aacenc --enable-libvo-amrwbenc --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxavs --enable-libxvid --enable-decklink --enable-zlib
  
    libavutil      52. 92.101 / 52. 92.101
  
    libavcodec     55. 69.100 / 55. 69.100
  
    libavformat    55. 48.101 / 55. 48.101
  
    libavdevice    55. 13.102 / 55. 13.102
  
    libavfilter     4. 11.102 /  4. 11.102
  
    libswscale      2.  6.100 /  2.  6.100
  
    libswresample   0. 19.100 /  0. 19.100
  
    libpostproc    52.  3.100 / 52.  3.100
  
  Input #0, mp3, from 'pipe:':
  
    Metadata:
  
      TSRC            : AUUM71001516
  
      title           : Sunlight
  
      track           : 2
  
      artist          : Bag Raiders
  
      copyright       : 2010 Modular Recordings
  
      genre           : Electronic
  
      album           : Bag Raiders
  
      album_artist    : Bag Raiders
  
      disc            : 1/1
  
      publisher       : Modular Recordings
  
      composer        : Chris Stracey/Jack Glass/Dan Black
  
      date            : 2010
  
    Duration: N/A, start: 0.000000, bitrate: 320 kb/s
  
      Stream #0:0: Audio: mp3, 44100 Hz, stereo, s16p, 320 kb/s
  
      Stream #0:1: Video: mjpeg, yuvj420p(pc, bt470bg), 600x600 [SAR 300:300 DAR 1:1], 90k tbr, 90k tbn, 90k tbc
  
      Metadata:
  
        title           : 
  
        comment         : Other
  
  Output #0, mp3, to 'pipe:':
  
    Metadata:
  
      TSRC            : AUUM71001516
  
      TIT2            : Sunlight
  
      TRCK            : 2
  
      TPE1            : Bag Raiders
  
      TCOP            : 2010 Modular Recordings
  
      TCON            : Electronic
  
      TALB            : Bag Raiders
  
      TPE2            : Bag Raiders
  
      TPOS            : 1/1
  
      TPUB            : Modular Recordings
  
      TCOM            : Chris Stracey/Jack Glass/Dan Black
  
      TDRL            : 2010
  
      TSSE            : Lavf55.48.101
  
      Stream #0:0: Audio: mp3 (libmp3lame), 44100 Hz, stereo, s16p, 192 kb/s
  
      Metadata:
  
        encoder         : Lavc55.69.100 libmp3lame
  
  Stream mapping:
  
    Stream #0:0 -> #0:0 (mp3 (native) -> mp3 (libmp3lame))
  
  size=       6kB time=00:00:00.21 bitrate= 227.6kbits/s    
  
  size=     102kB time=00:00:04.31 bitrate= 193.7kbits/s    
  
  size=     202kB time=00:00:08.56 bitrate= 192.9kbits/s    
  
  size=     341kB time=00:00:14.49 bitrate= 192.5kbits/s    
  
  size=     489kB time=00:00:20.82 bitrate= 192.4kbits/s    
  
  size=     642kB time=00:00:27.35 bitrate= 192.3kbits/s    
  
  size=     792kB time=00:00:33.75 bitrate= 192.2kbits/s    
  
  size=     950kB time=00:00:40.49 bitrate= 192.2kbits/s    
  
  size=    1106kB time=00:00:47.15 bitrate= 192.2kbits/s    
  
  size=    1258kB time=00:00:53.63 bitrate= 192.1kbits/s    
  
  size=    1415kB time=00:01:00.31 bitrate= 192.1kbits/s    
  
  size=    1563kB time=00:01:06.66 bitrate= 192.1kbits/s    
  
  size=    1710kB time=00:01:12.90 bitrate= 192.1kbits/s    
  
  size=    1857kB time=00:01:19.17 bitrate= 192.1kbits/s    
  
  size=    2008kB time=00:01:25.63 bitrate= 192.1kbits/s    
  
  size=    2162kB time=00:01:32.21 bitrate= 192.1kbits/s    
  
  size=    2299kB time=00:01:38.03 bitrate= 192.1kbits/s    
  
  size=    2457kB time=00:01:44.80 bitrate= 192.1kbits/s    
  
  size=    2600kB time=00:01:50.89 bitrate= 192.1kbits/s    
  
  size=    2755kB time=00:01:57.52 bitrate= 192.1kbits/s    
  
  size=    2864kB time=00:02:02.17 bitrate= 192.1kbits/s    
  
  size=    3022kB time=00:02:08.88 bitrate= 192.1kbits/s    
  
  size=    3172kB time=00:02:15.31 bitrate= 192.1kbits/s    
  
  size=    3284kB time=00:02:20.06 bitrate= 192.1kbits/s    
  
  size=    3385kB time=00:02:24.40 bitrate= 192.1kbits/s    
  
  size=    3529kB time=00:02:30.51 bitrate= 192.0kbits/s    
  
  size=    3687kB time=00:02:37.25 bitrate= 192.0kbits/s    
  
  size=    3838kB time=00:02:43.71 bitrate= 192.0kbits/s    
  
  size=    3988kB time=00:02:50.11 bitrate= 192.0kbits/s    
  
  size=    4138kB time=00:02:56.53 bitrate= 192.0kbits/s    
  
  size=    4279kB time=00:03:02.54 bitrate= 192.0kbits/s    
  
  size=    4408kB time=00:03:08.03 bitrate= 192.0kbits/s    
  
  size=    4544kB time=00:03:13.85 bitrate= 192.0kbits/s    
  
  size=    4683kB time=00:03:19.78 bitrate= 192.0kbits/s    
  
  size=    4805kB time=00:03:24.95 bitrate= 192.0kbits/s    
  
  size=    4939kB time=00:03:30.67 bitrate= 192.0kbits/s    
  
  size=    5049kB time=00:03:35.38 bitrate= 192.0kbits/s    
  
  size=    5141kB time=00:03:39.32 bitrate= 192.0kbits/s    
  
  size=    5263kB time=00:03:44.49 bitrate= 192.0kbits/s    
  
  size=    5372kB time=00:03:49.17 bitrate= 192.0kbits/s    
  
  The thread 0xb24 has exited with code 259 (0x103).
  
  size=    5436kB time=00:03:51.91 bitrate= 192.0kbits/s    
  
  size=    5509kB time=00:03:55.02 bitrate= 192.0kbits/s    
  
  size=    5657kB time=00:04:01.32 bitrate= 192.0kbits/s    
  
  size=    5702kB time=00:04:03.22 bitrate= 192.0kbits/s
  
  
  
  video:0kB audio:5701kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.005738%

  Any ideas ? I’m grateful for suggestions - I’ve been chasing this for a week now !

• FFMPEG : Output tiles onto a black background [closed]

  14 juin 2013, par MollyRazor

  I have the following command, which converts a video and also generates a 10x9
  tileset. Currently it gets the longer side and either generate a 90px wide or
  45px tall tiles. My problem is that each tile has to be 90x45 large, so in
  essence what I need is a properly resized and centered tile inside a black
  bounding box. How can I achieve this ?

  The resulting tile

  

  What I would like to achieve

  

  Actual command line

  ffmpeg -i K/vertical.mp4 .... vertical.out.mp4
  
  ffmpeg -i K/vertical.mp4 -an -vsync vfr \
  
  -vf select="isnan(prev_selected_t)+gte(t-prev_selected_t\,2)",scale="'if(gt(iw,ih),90,-1)':'if(gt(iw,ih),-1,45)'",tile="10x9" \
  
  -qscale:v 3 './f/f%03d.jpg'

• lavfi/delogo : use weighted interpolation

  26 juin 2013, par Jean Delvare

  lavfi/delogo : use weighted interpolation
  The original delogo algorithm interpolates both horizontally and
  
  vertically and uses the average to compute the resulting sample. This
  
  works reasonably well when the logo area is almost square. However
  
  when the logo area is significantly larger than high or higher than
  
  large, the result is largely suboptimal.
  The issue can be clearly seen by testing the delogo filter with a fake
  
  logo area that is 200 pixels large and 2 pixels high. Vertical
  
  interpolation gives a very good result in that case, horizontal
  
  interpolation gives a very bad result, and the overall result is poor,
  
  because both are given the same weight.
  Even when the logo is roughly square, the current algorithm gives poor
  
  results on the borders of the logo area, because it always gives
  
  horizontal and vertical interpolations an equal weight, and this is
  
  suboptimal on borders. For example, in the middle of the left hand
  
  side border of the logo, you want to trust the left known point much
  
  more than the right known point (which the current algorithm already
  
  does) but also much more than the top and bottom known points (which
  
  the current algorithm doesn’t do.)
  By properly weighting each known point when computing the value of
  
  each interpolated pixel, the visual result is much better, especially
  
  on borders and/or for high or large logo areas.
  The algorithm I implemented guarantees that the weight of each of the
  
  4 known points directly depends on its distance to the interpolated
  
  point. It is largely inspired from the original algorithm, the key
  
  difference being that it computes the relative weights globally
  
  instead of separating the vertical and horizontal interpolations and
  
  combining them afterward.
  Signed-off-by : Jean Delvare <khali@linux-fr.org>
  
  Signed-off-by : Stefano Sabatini <stefasab@gmail.com>
  

  • [DH] libavfilter/vf_delogo.c
  • [DH] tests/ref/fate/filter-delogo