No subject
Mon Jul 25 00:22:01 PDT 2011
/* FIXME :
*
* The usage of parsers in encoding/muxing scenarios is
* just too undefined to just use as-is.
*
* Take the use-case where you want to re-mux a stream of type
* "my/media". You create a StreamEncodingProfile with that type
* as the target (as-is). And you use decodebin2/uridecodebin
* upstream.
*
* * demuxer exposes "my/media"
* * a parser is available for "my/media" which has a source pad
* caps of "my/media,parsed=True"
* * decodebin2/uridecodebin exposes a new pad with the parsed caps
* * You request a new stream from encodebin, which will match the
* streamprofile and creates a group (i.e. going through this method)
* There is a matching parser (the same used in the decoder) whose
* source pad caps intersects with the stream profile caps, you
* therefore use it...
* * ... but that parser has a "my/media,parsed=False" sink pad caps
* * ... and you can't link your decodebin pad to encodebin.
*
* In the end, it comes down to parsers only taking into account the
* decoding use-cases.
*
* One way to solve that might be to :
* * Make parsers sink pad caps be "framed={False,True}" and the
* source pad caps be "framed=True"
* * Modify decodebin2 accordingly to avoid looping and chaining
* an infinite number of parsers
*
* Another way would be to have "well-known" caps properties to specify
* whether a stream has been parsed or not.
* * currently we fail. aacparse uses 'framed' and mp3parse uses 'parsed'
*/
/* FIXME : Re-enable once parser situation is un-$#*@(%$#ed */
--
Configure bugmail: https://bugzilla.gnome.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
You are the assignee for the bug.
More information about the gstreamer-bugs
mailing list