On Wed, Jul 26, 2006 at 10:48:40AM +0200, ext Frantisek Dufka wrote: > Daniel Stone wrote: > >On Wed, Jul 26, 2006 at 09:34:03AM +0200, ext Frantisek Dufka wrote: > >>I thought DSP only decodes video to shared framebuffer in main memory > >>and the actual transfer to epson chip is done by omapfb driver i.e. by > >>ARM CPU from gstreamer or directly from mediaserver code. > > > >Correct. > > Sorry, but is is still not clear to me which parts you quoted are > correct. So currently DSP code decodes mp4 into RGB and then lets the > framebuffer code blit it to the epson chip in RGB? And ioctls with > OMAPFB_COLOR_YUV420 is not used due to not solved problems with multiple > surfaces?
The paragraph I've left quoted is correct: the DSP decodes into a shared framebuffer, which is then DMA'ed across. > >>Am I wrong with this? If not, is it possible to add YUV surfaces to > >>xserver (via Xv or Nokia Xsp extensions) so it could be used from X11 > >>code in Mplayer or (nonpartched or patched) SDL ? > > > >Not at the moment: you need multiple planes, so you can also set up a > >separate YUV surface, you can't just blit in with a different format. > > And why this is a problem? Yes I understand you need extra memory > (surface) allocated for YUV operations. Why it is a problem to have > 800x600 in RGB and additional 320x240 in YUV both managed by xserver? > Isn't this generic thing already solved in xserver and used on other > plaforms? Sorry for being too ignorant about X internals :-) It means that the video hardware does not support RGB565 (everything on the device, basically) and YUV together. You can't have part of the screen being RGB and part being YUV; the problem is a couple of layers below the X server. Cheers, Daniel
signature.asc
Description: Digital signature
_______________________________________________ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers