[FFmpeg-devel] [PATCH] HWAccel infrastructure (take 3)

[Stephen Warren]

Michael Niedermayer wrote:
> 
> On Thu, Feb 19, 2009 at 05:34:03AM +0100, Gwenole Beauchesne wrote:
> > I don't find it very useful, or
> > what/how the user should do any good with that?
> 
> skip spliting the frame in slices in software when its available as a whole
> frame and the hardware ca handle that directly.

FYI, I'd like to point out a *possible* (but by no means certain) future
extension to the VDPAU API. Right now, the entire picture's bitstream is
passed to VDPAU, and the HW parses out where the slices start. In order to
implement better error recovery, we *might* extend the VDPAU API to also
accept an array of byte offsets into the buffer, which list where the
individual slices start. These would be used as resync points in the case
of bitstream parsing errors. So, parsing out the slices may be useful.

Also, in general, I've worked on other manufacturer's HW where you pass
individual slices to the HW one at a time, rather than the whole bitstream,
so to keep the new ffmpeg data structures fully general, that'd also imply
parsing out the slice start positions.

Of course, perhaps this could be added later, if and when needed, rather
than right now, and for performance reasons could be optional based on some
accelerator-provided flag. Presumably though, since ffmpeg searches the
data for start codes to find the next picture, recording the slice start
positions would be almost zero extra work anyway.

-- 
nvpublic