[MPlayer-dev-eng] [PATCH] Realplayer codec support

[Florian Schneider]

> Hi,
> 
> > but after about 50 times re-reading this, i think i found the purpose of
> > this hack. this is just a table (pair of dwords) of sub-blocks (offset +
> size),
> > and this routine is called to fill up bit-buffers (or its pointers).
> this
> > explains those 8* multiplications before comparing start/end
> (bit)position
> 
> yep, i got it!!!!! mplayer plays rv30 video!

Cool! You're great!

> 
> let's see the real stuff:
> 
> transin1[pointer_to_transin1]: {subpackets-1/totallength} x0/x1 x2/x3
> x4/x5 ...
> (printed only first 2*subpackets of x values)
> 
> 
> so, the first value of the long pairs is always 0x1, teh second one is the
> offset of sub-packet in the assembled packet (always 0 for the first)
> 
> and it DOES WORK with mplayer! ;)
> 
> i won't commit yet, it needs "some" cleanup...

Yes, there is a lot to do. I am doing it at home, but currently
I am visiting my brother, so I can do it maybe tomorrow.

> and i want to move that sub-packet "demuxer" to the codec, where it
> belongs to

Do you mean to put the sub packet demuxer into vd_real.c?
I don't know if this will be good (i.e. without problems),
because often, the input data for one frame spans over multiple
packets - and I only get one packet per frame inside vd_real.c.

Regards

Flo

-- 
GMX - Die Kommunikationsplattform im Internet.
http://www.gmx.net