AW: [MPlayer-users] ati radeon 8500

Arpi arpi at thot.banki.hu
Sat Mar 9 01:16:09 CET 2002


Hi,

> > In theory - yes. In practise, Nick made it knowing only the ATI hw and its
> > requirements, and as soon as we started to port mga_vid to vidix, found
> > that it's impossible without api changes...
> 
> So the mga vidix module is "incomplete"?  Should I stop using it?  It
> seems to have been working fine for me on the framebuffer.  I was
> goign to try to fix the HW EQ with it but if it's broke, there is no
> point.

I don't know.
Alex hacked it together from mga_vid kernel module, but it didn't worked (he
did it blindly - he has no matrox card :))
then i tried to fix it but found the vidix interface bad, at least for
things i need it was a way too redundant and complicated, while missing
important things. so i didn't spent more time on finishing it and
re-enabling lot of temporary disabled code...

anyway it was in the past, vidix interface changed a lot, and i like it a
bit better, so one day i will may help finish vidix version of mga_vid...
anyway the vidix if. still needs improvement, as matrox cards work a way
different than ati ones, including colorspaces, buffer changing, pitch
calculation etc...

i just wanted to say, and again: it's my opinion: current vidix interface is
designed for ati cards, and needs some improvement to fully support features
of other cards. maybe i'm wrong, i didn't spent too much time on vidix yet.

while i like the base idea of it, and accepted the way of accessing hw from
user space (i prefer kernel modules, but the vidix way is much more
portable) i still don't like the design and silly structures it uses.

mga_vid kernel module has much more features implemented than vidix one,
including 'equalizer' (it's Nick's terminology, i think this is silly name to
the function).


A'rpi / Astral & ESP-team

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




More information about the MPlayer-users mailing list