[Gnash-commit] [bug #33760] gst fails to play loaded dynamic sounds.

[Bastiaan Jacques]

Update of bug #33760 (project gnash):

                Category:               sound-gst => sound                  

    _______________________________________________________

Follow-up Comment #3:

MediaParser base class is in charge of obtaining new buffers from the derived
Media Parser class, which I will call "DMP". MediaParser asks DMP to parse
frames until a buffer (set to 1 minute) is full. Then it waits for Sound_as to
consume the buffers and parsing continues. The trouble is that there is a
chicken-and-egg situation between MediaParser and Sound_as:
Sound_as waits for parsing to be completed before it consumes audio and
MediaParser is waiting for Sound_as to consume some data before it parses the
rest of the file.

So, you ask, why does it work with Ffmpeg at all? Well, there is a bug in the
ffmpeg MediaParser, that somehow detects a single video frame in the mp3, and
the buffer length is calculated to be the minimum of either the audio or video
time, and since the minimum video time will never reach the one minute
barrier, Ffmpeg is permitted to parse the whole file; and then Sound_as can
begin consuming.

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?33760>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/