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

[Arpi]

Hi,

> > 
> > i have the feeling that they are somehow prefixed pointers and other data.
> > the 0x8?????? seems to be memory pointers, while the others looks like
> > special codes or counters. maybe this is a huffman table or other kind of
> > codebook?
> > 
> 
> The 0x8xxxxxx are pointers to the global heap, the 0x4xxxxxxx are 
> pointers to the codec and other dynamically loaded dll's. I think that 
> somewhere in a previous transformation (or somewhere else) these data 

i hope it's constructed by the transform stuff.
custommessage can't do, as i can ignor ethem in wrapper and realplayer still
work. hivemsg is called only once.

but i have the feeling that it's come from rv30.so, their sub-demuxer.
maybe this is somewhere between the packets, an other stream (note that your
sample file has 3 stream, audio, video and an unknown type) or other source
(built-in tables?)

> will be constructed - at least I hope so. Or a thorough look at the code 
>   might help a bit.
hmm. maybe that transin* is just a pointer to some externally allocated
buffer, used by the codec to store inter-frame data (like motion vector
predictions etc, they seem to missing causing the strange decoding bugs)

but your current decoder interface just set it to a temp array of 0,1,0,0,0
so stuff stored by the codec lost between frames.
maybe this is their codec-specific 'context' field... i'll try this at home

> If you are interested, I can send the disasm of the codecs to you.
yes, uploa dto incoming


A'rpi / Astral & ESP-team

--
Developer of MPlayer, the Movie Player for Linux - http://www.MPlayerHQ.hu