On 10/18/07, Tom Metro <[EMAIL PROTECTED]> wrote: > OK, so the mvpmc code uses a software codec to decode the audio and then
specifically mvpmc uses liba52 http://liba52.sourceforge.net/ for analog. You seem to want to ignore the digital output which is where true 5:1 is available. > it writes it to some buffer to be processed by the MVP hardware. The output is to one of several hardware devices /dev/adec* > But isn't the MP3 audio also handled by a software codec per your > statement above? Why does it not have the same sync issues? No there is a mpeg audio device that does hardware decoding. Without too technical for the user forum I suggest you look at the doxygen output for libav. You should note that most of the functions that are used for sync'ing are based on the mpeg audio device > It's the mvpmc code that demuxes the stream and feeds the audio and > video buffers separately - no? Right and the new JIT code does this more precisely. > Wouldn't one way to accomplish this kind of sync be to drop the > necessary number of video frames as you're feeding them into the video > buffer? The code does this already and it works with AC3, kind of, but only when the video is ahead, and as I said in my last message when the audio is ahead there are problems. The JIT code can in theory do this more precisely, but the AC3 frames are also crc'd so dropping "frames" is not as easy.as with mpeg data. Martin ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ Mvpmc-users mailing list Mvpmc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mvpmc-users mvpmc wiki: http://mvpmc.wikispaces.com/
