[mythtv] packed image format and new deinterlacers

Lincoln Dale ltd at interlink.com.au
Sun Sep 17 23:57:55 UTC 2006


> > > but with delaying the first frames i should go into AV-Sync troubles
> > > which should be noticable by the user.
> > > if i solved the above i would start with a first "copy-needed-frames"
> > > version.
> >
> > i'm sure folks wouldn't be opposed to an enhancement to AudioOutputBase
> > class (libs/libmyth/audiooutputbase.cpp) such that a deinterlacer could
> tell
> > the audio sync to be delayed to whatever delay an interlacer needs.
> >
> > i'm sure if we could take deinterlacers from DScaler with minimal
> effort, it
> > would be well worth it.
> 
> whats wrong with the regular ffmpeg deinterlacers?

if we had support for DScaler's deinterlacers that simply gives us a larger
pool of filters for people to play/experiment with.

i have no clue as to whether DScaler's filters are superior to ffmpeg or not
-- TV here is a 68cm 10 year old CRT - there isn't a lot of postprocessing
necessary on that.  perhaps when i pick up my 50" plasma i'll be able to
experiment first hand.

some folks have said good things about post-processing using the GreedyH
DScaler filter, some think they get superior image quality out of xine with
various DScaler filters.

i think its most likely that there simply isn't "one filter" that produces
best results all the time.  having a larger pool of filters to use may mean
that the most optimal setup in future is to have not only different filters
triggered based on source of material but perhaps also on genre of content
(e.g. some filters aren't suitable for sporting programmes, some more
suitable for cartoons/anime etc).
i see one of myth's main strengths versus commercial PVR/TiVo's etc is that
it is so flexible - and we can significantly improve on many commercial
appliances offer.  i think we already do this but i don't think we've
reached the summit there yet. :)



cheers,

lincoln.




More information about the mythtv-dev mailing list