Problem with persistent scaling/shifting in RADEONDisplayVideo()

Thomas Hilber xorg-driver-ati at toh.cx
Wed Jul 30 18:07:02 PDT 2008


On Wed, Jul 30, 2008 at 11:12:05PM +0200, Roland Scheidegger wrote:
> That's what fast cpus are for :-). Though I agree in particular the

really:-) ?
VDR was designed with the goal in mind to also support very slow CPUs
like 600Mhz Pentiums and slower. So the use of so called full featured
cards (cards with hardware/firmware decoder and proper stream
synchronization) was mandatory until now.

But I don't like these expensive cards running undocumented firmware.
Furthermore they are full sized and don't allow small casings.

So I started this project with Radeons. I'm really amazed what's possible 
with these cheap and small cards. 

> sharpness), it looks much better to me due to this. I'm not sure native
> deinterlacing is much better, after all you still deinterlace - in your
> brain :-).

I experimented a lot with these things. The best method (at least for me)
really is to trick your brain:-)

> Hmm ok that's only about 5000 fields, I'd have expected a bit better.
> Though for progressive doesn't sound too bad - depending on the scene
> you will never notice this happening every 2 minutes.

that's not quite true. Watching tennis or football where you have a 
steadily moving camera it's very likely you hit almost every field loss.
You can't build a good quality htpc with that.

I forgot to mention another important effect:

DVB streams and VGA video timings of course don't come along with
mathematical exact frequencies. That means because of hysteresis effects
you e.g. don't loose 1 frame in intervals of 100s. 

Without proper synchronisation you rather loose a bunch of frames every 
100s. What is clearly unbearable.

> Well, vdr serves as a media (and dvd player) too, so unless you restrict
> it to use it only with dvb cards you'd still have that problem.

right. But the original VDR is designed for full featured DVB cards 
only. These cards I now get rid of without any quality loss.
 
- Thomas


More information about the xorg-driver-ati mailing list