Moderator
Question for MostAwesome Dude.
Hey, got a question for ya.
I was reading back into some linked mailing list posts on gnash and came across this:
http://article.gmane.org/gmane.comp....top.xorg/40358
Is the plan of implementing vdpau on the opensource ATI drivers still planned or was it killed?
X.Org ATI Driver Developer
The best bet would probably be a VDPAU front end for gallium so any gallium driver can take advantage of it.
Senior Member
I would be much more comfortable using something if it wasn't nvidia that came up with it....
MostAwesomeDude says (in that link) that VDPAU is the only one worth implementing (compared to VAAPI, XvBA, and XvMC). What is it that makes it more worth implementing than VAAPI?
Moderator
No offense but that's pretty irrational since the API is open and free to implement.Originally Posted by droidhacker
Probably because it's the most complete API out there with the most advance features that is actively maintained with quite a bit of untapped potential yet.MostAwesomeDude says (in that link) that VDPAU is the only one worth implementing (compared to VAAPI, XvBA, and XvMC). What is it that makes it more worth implementing than VAAPI?
Senior Member
It's the current de-facto standard for video accel and most likely to be adopted widely by applications. That means it's the best choice for any driver to support. I think what OpenGL is for 3D graphics, VDPAU will be for video. So really, it's pretty much a must.
X.Org ATI Driver Developer
It doesn't really matter which API is used, the actual decoding implementation work will mostly be the same.
Senior Member
What is difference between OpenGL-based decoding and shareds-based one?OpenCL would be fine for video, but stupid and slow, just like
OpenGL-based video decoding.
(...) so we'll have to use 3D shaders instead.
X.Org ATI Driver Developer
Just different APIs; both generate shaders that run on the hw.
Posting Permissions
Reply With Quote