[MPlayer-dev-eng] Lots of stuff for NUT

Oded Shimon ods15 at ods15.dyndns.org
Wed Jan 11 06:44:23 CET 2006


On Fri, Dec 30, 2005 at 09:10:02PM +0200, Oded Shimon wrote:
> [...]

Good news. Rich and I were discussing on IRC, and Rich found out we DON'T 
need to store per stream pts. We need to store pts for every stream with 
non zero decode_delay, and a SINGLE pts for all the decode_delay=0 streams. 
(which is all but one stream most likely). These streams are strictly 
interleaved, so a single pts for them works!

The common pts for the interleaved streams:
max(min(key_pts in back_ptr region) across all streams)

This is NOT more complicated than per stream pts for those worrying about 
simplicity. back_ptr is still NOT shared, so knowing min key_pts in 
back_ptr region is trivial, and you needed it anyway for per stream pts. 
(every time a keyframe changes the back_ptr of a stream, store the 
keyframe's pts.)

Now, we need the perfect spec.. I'm hoping for something which spends LESS 
than a byte per stream, and if all streams share a common back_ptr, it 
should take least amount of space possible, so additional streams take 
little or zero additional space.

The syncpoint needs to store - back_ptr for every stream, a single pts for 
all decode_delay=0 streams, and a pts for every non zero decode_delay 
stream.

If we can make overhead less than linear to amount of streams, that would 
kick ass. Any brilliant ideas anyone?...

- ods15




More information about the MPlayer-dev-eng mailing list