Re: [xineliboutput-devel] [PATCH] new algorithm for live mode sync of xineliboutput

[Petri H. <phi...@us...>]

Paul Menzel wrote: 
> Thomas Hilber wrote a new algorithm [1 and attached]. Could you take a
> look at it, please, and give feedback or suggestions?

Smoother SCR adjustments and filtering the buffer fill level are
definetely good thing.

"static int pts_delay" is not used for anything. If it is needed, it
should be stored in some dynamically allocated struct (maybe in the scr
data in adjustable_scr.[hc] ?).
Is the (commented-out) shm code meant to interact with video driver ?

The old code (#if 0 ... #endif) can be removed.

(Unrelated) logging changes to SCR functions can be left out or in
separate patch.

In the future (with mpeg-ts) one could try to use PCR to sync clocks
after initial buffering perioid.

> It looks like it does not apply cleanly to HEAD and the patch is against
> 1.0.4. If you want me to I can clean that up.

That would be great. I don't know if I have time to maintain 1.0.x
branch anymore.


- Petri