Personally, I'm not sure if I like such strange multiplicators just to satisfy one decoder (which isn't even a good one). From the source code comments it seems that he did that to satisfy the Sonic audio decoder. Maybe we should contact orbitlee and ask him about that. I'm not sure where the "magic" multipliers are coming from. The main change I've done is just a special handling for 2012 bytes DTS frames. Some of the changes you're talking about were originally done by orbitlee.
Are you sure your patch doesn't break parsing of regular AC3/DTS stuff? Because framesize and bytesperframe values are much bigger now (multiplied by magic numbers). I have looked at the code changes (compared to guliverkli SVN). I would like to include this patch in my patched MPC build. We have to mathematically handle the 2012 frames as if they were 2013 bytes long to keep audio sync while seeking. But audio sync slowly drifted away from the video. DTSAC3Source handled this mathematically correctly. But DTS tracks taken from HD DVD and Blu-Ray strangely only have 2012 bytes per frame. The problem with (3) is this: Usually 1536kbit/s DTS tracks have 2013 bytes per frame. (3) Fix for problems with audio sync with Blu-Ray/HD DVD DTS files. (2) m_AvgTimePerFrame calculation was incorrect for DTS.
(1) m_AvgTimePerFrame calculation was incorrect for AC3. Can anybody with SVN access please upload a little patch for me? I've done the changes to Casimir's SVN, but I don't know SVN well enough to "upload" something.Ĭontained is a new version of DTSAC3Source.cpp (together with a new compiled dtsac3source.ax).