[MPlayer-dev-eng] NUT (informal) proposal, based on discussions

[Michael Niedermayer]

Hi

On Fri, Jan 20, 2006 at 06:48:02PM +0200, Oded Shimon wrote:
> On Fri, Jan 20, 2006 at 06:12:07PM +0200, Oded Shimon wrote:
> > Index results:
> > 
> > high quality 700mb  0 subtitle streams: 81031
> > high quality 700mb 10 subtitle streams: 102475
> > low  quality 620mb  0 subtitle streams: 65790
> > low  quality 620mb 10 subtitle streams: 109409
> > 
> > BTW, my repeat stuff is apperantely quite unhelpful, it can be removed at 
> > the cost of 3kb or so for the sake of simplicity.
> 
> high quality 700mb  0 subtitle streams: 83609
> high quality 700mb 10 subtitle streams: 104683
> low  quality 620mb  0 subtitle streams: 65225
> low  quality 620mb 10 subtitle streams: 142253
> 
> The high change is thanks to the subtitle streams being very "idealistic", 
> giving exactly 1 frame every 100 frames, with pts 100 more than last 
> frame... So I still think it is a good idea to remove the repeat stuff.
> 
> > I'm sending a new patch soon and asking for immediatte aproval for 
> > commit...
> 
> Again:
> 
> step-4:
>   add EOR
>   add coded stream flags
>   just add sflags to header and give intial values to tmp_*
> 
> step-5:
>   change goals slightly..
>   max_index_distance removed
>   index changed to combination of syncpoint index and pts for keyframes,
>   using ideas by michael
>   syncpoint still single back_ptr and pts, however back_ptr is changed -
>   points to most correct keyframe, not most recent.
>   syncpoint timestamp is max(last_dts)
> 
> cosmetic:
>   move things around to IMO more logical positions.
> 
> whats left:
>   info streams, how do they deal with EOR, back_ptr, etc.
>   index repetition (atleast the option for it)
>   I'd like to define function like convert_ts: "compare_ts", for comparing 
>   timestamps of different timebases, mostly because I don't know how to do 
>   it myself. :) Function should return less than, equal to, or greater than 
>   zero as needed...
> 
> 
> Well, review please, I'd like to commit these things as long as the date 
> change in the patches hold. :)

patches ok, maybe i have a comment or 2 but these are minor and IMHO
better to disscuss later, this repeated reviewing is annoying

[...]

-- 
Michael