On Mon, Oct 31, 2005 at 11:15:34PM +0100, Stefan Lucke wrote: > On Sonntag, 30. Oktober 2005 03:21, Ville Syrjälä wrote: > > On Sat, Oct 29, 2005 at 11:45:57PM +0200, Stefan Lucke wrote: > > > On Samstag, 29. Oktober 2005 18:59, Ville Syrjälä wrote: > > > > On Sat, Oct 29, 2005 at 05:45:08PM +0200, Stefan Lucke wrote: > > > > > > > > > > Oh, I forgot to mention that my test were done with my G550, but I can > > > > > confirm that your fix is a fix for my card too. > > > > > > > > Great. Saves me from doing another card swap :) > > > > > > > > > The example streams I got from > > > > > http://www.divx.com/hd/ > > > > > > > > I have those too. There are quite a few HD H.264 samples at > > > > http://www.apple.com/quicktime/guide/hd/. Good luck trying to watch > > > > them > > > > if you don't have a fast CPU. My Duron 1300Mhz can handle 720p but only > > > > if > > > > I skip the loop filter step and 1080p is simply too much for it to > > > > handle. > > > > > > I can't go at such high resolutions. With 1920x1080 even a > > > TestConfiguration() segfaults. > > > > I forgot to mention that the BES max height is 1024 so you need to crop or > > pre-scale the video a bit. Though if you're going to scale you might as > > well use StretchBlit() and not use the BES at all. > > Shouldn't each field has up to 1024 lines (BESV1SRCLST + BESTV2SRCLST) ?
Yes, but only if you enable DLOP_DEINTERLACING. The field 2 registers are only used with deinterlacing. The hardware can only display one field at a time. Without deinterlacing field 1 == frame. Actually even that is not supported in the current code. It just limits the max height to 1024. I was about to change that but I couldn't reach directfb.org. Seems to be back now so I can actually commit :) There's one other thing I have to add for G450/G550. When deinterlacing the pitch is doubled but the hardware can't handle a pitch of 4096. So I need to limit max width to 2048 - 128 in that case. That 128 value used because the 128 bytepitch alignment equals 128 pixels in 4:2:0 modes. > 1. So even with a fresh checkout TestConfiguration() results in a thrown > excepiton. This counters the idea of doing a TestConfiguration() before > SetConfiguration(). > > 2. As this is a layer specific limit, like supported surface formats, > it should be possible to qurey such a area limitation. That's what TestConfiguration() is for. It's a blit clumsy though since it might be nice to be able to query just the min and max limits. Unfortunately that would not work very well in some cases since the limits can change according to pixelformat and other configuration changes. -- Ville Syrjälä [EMAIL PROTECTED] http://www.sci.fi/~syrjala/ _______________________________________________ directfb-dev mailing list [email protected] http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-dev
