Quoting Ville Syrj?l? ([EMAIL PROTECTED]): > On Thu, Jan 09, 2003 at 01:08:49PM +0100, Thomas Wentzel wrote: > > Hi Ville, > > > > Ville Syrj�l� wrote: > > > > > On Wed, Jan 08, 2003 at 06:07:39PM +0100, Thomas Wentzel wrote: > > > > I was thinking about whether it would be possible to add support for > > > > the DXR3 chip? I mean - are there any hardware issues preventing such > > > > a thing? > > > > > > Hmm. I have no idea how these decoder boards work. What kind of data does > > > it accept and does it have video memory? > > > > It takes MPEG 1 and 2. And yes it has some (don't know how much) video > > memory, but it isn't general purpose - well actually it's simply a buffer. > > (as far as I'm informed, anyway) > > It would then need a new "pixel format" for MPEG data. As for the buffer > issue... Supposing that the buffer can't be mapped or if it's just > impossible to present it as a surface buffer I suppose it could use a > double buffered surface in system memory and feed the data to the hw on > Flip().
If the YUV data of a video layer cannot be mmapped, e.g. if it's live fed from an MPEG2 decoder, the layer simply has no surface. There's only a limited configuration support (DLCONF_OPTIONS allowed) and stuff like SetScreenLocation, SetOpacity is supported. Currently it's not really clear where the video comes from. There's no video provider that would control the video being displayed because video providers are surface and frame buffer based. Until there's an API for this end point of the architecture only software that is written for or adapted to a specific platform makes usage of this kind of layers. -- Best regards, Denis Oliver Kropp .------------------------------------------. | DirectFB - Hardware accelerated graphics | | http://www.directfb.org/ | "------------------------------------------" Convergence GmbH -- Info: To unsubscribe send a mail to [EMAIL PROTECTED] with "unsubscribe directfb-dev" as subject.
