Requires audio output. Here's the basic idea now: (someone please verify)
No time stamps or nothin! - we do have timestamps, be they're not used at the output right now.
No audio output lets the video display freely... ie, really fast.
Wait for updated Mpeg2Dec support to have a queue of frames to display. (???)
Have time stamps on output frames. Display thread (or whatever) waits till correct time to output frame.
Audio similar. BUT we want to give priority to audio! Don't let audio have breaks if at all possible. Much worse on the senses than some dropped video frames.
frame dropping? pre or post decode? would be better to not decode the frame if possible. allowing both would be best.
partial frame decoding?
multiple buffers could also be implemented, i.e. a buffer for the point of css decoding, then the buffer after mpeg2dec before it's actually rendered, even rendering should be double-buffered for smoothness, but, options would be there for card support, I have a great card for DVD, supports all known features besides hardware decoding (ATI Rage 128 Mobility M3) and I have oms working fine, but, it's choppy video with good sound... who knows. Current State
... and on too slow computers we instruct the video decoder to drop some frames
(on demand) Future
Buffering, buffering, buffering, buffering.... needless to say, buffering is the key, buffer those decoded screens like, 10-15 seconds if possible, check memory, and then render, like streaming video on most systems, buffering is good for you, even sound can be buffered. config = 0.000
total = 0.590
imports = 0.450
send_page = 0.130