[MPlayer-dev-eng] real demuxer

[Balatoni Denes]

On 2003. február 16. 19:49, Arpi wrote:
> Hi,
>
> > I hope real demuxer fixes will not be forgetten and included in 0.90.
>
> in 0_90? dunno. i still don't like your messy if()
actually, as I said helix is not using object version 0 for data packets 
afaict, so there is no need to worry that the sync will break with newer rv40 
streams. 


> > I looked at Pen3.rm and Noir.rm and dt6... .rm - they all have wrong
> > dimensions recognized. The problem is that the dimensions embedded in the
> > rm in the usual place doesn't match that of the actula dimensions of the
> > encoded
>
> yes, we know - what do you think why are they in real-bug/bad_WxH/  ? :)
yes, dt9871324.rm was there indeed, others iirc not 
if I knew it was known, I wouldn't have duplicated your work I can assure you 
about that ;)

> yes
> and realplayer passes the correct WxH to the codec (according to trace
> logs) so it must be stored somewhere in the headers. or - afair Alex
> sugegsted - there may be a table (in the player) of the standard
> resolutions, and they store only the index. at least the h263 codec (a.k.a.
> RV10 :)) is such thing, there is 3 bits in the header, value 0..6 are
> standard resolutions, 7 is user defined (7 is followed by W and H).
yes good idea, if I have motivation I might attempt to find it. only if I saw 
the point in doing it ;] why would sane people use rm... and for them there 
is realplayer

> A'rpi / Astral & ESP-team

bye
Denes