[PD] Gem fullscreen bug.

vade doktorp at mac.com
Tue Mar 1 21:37:35 CET 2005

Actually, out of habit I used pd_filmYUV+pix_texture. If pix_movie is 
indeed faster (this is news to me, I thought the YUV gl path was 
fastest..), I will re-try it on the machine and post results.

What codec did you use for the 1080i footage? Where are you? If youre 
around NYC id be more than happy to give you some time on a G5 machine 
to do some optimizations :)

On Mar 1, 2005, at 2:31 PM, chris clepper wrote:

> On Mar 1, 2005, at 11:27 AM, vade wrote:
>> With fullscreen set to 1 and the display at its native 1900x1200 
>> aspect ratio, GEM would incorrectly size geometry in the gemwin 
>> context. Only if I manually set the dimen param to gemwin would it 
>> correct itself and properly display a square as a square, and not as 
>> a stretched out cubiod.
> I wonder how this would happen.  The 'fullscreen $1' method queries 
> the OS for the screen dimensions and sets the window to that.  Maybe 
> the dimensions are not returned properly?
>> Just a heads up, nothing major.
>> BTW, we had 2 1080i DVCPro HD Streams running to 2 rects with barely 
>> a frame dropped, with 3D transforms and all. Very.. very impressive..
> Are you using pix_movie or pix_film + pix_texture?  pix_,movie has an 
> additional optimization for getting pixels to the card.  It's 
> basically CoreVideo as a GEM object.  With frame sizes that big it can 
> make a lot of difference.  I was able to get one 1080 stream going 
> 60fps on a 2Ghz G5 textured to hundreds of Geos.  There's room for 
> more code enhancements for the PPC 970, but I don't even have decent 
> access to one for testing!
> Does that codec use both CPUs for decompression?  I would think it 
> would, but haven't really confirmed it.
>> Keep up the good work!
> Thanks, you too.  Post some documentation of the project online if you 
> can.
> cgc
> _______________________________________________
> PD-list at iem.at mailing list
> UNSUBSCRIBE and account-management -> 
> http://iem.at/cgi-bin/mailman/listinfo/pd-list

More information about the Pd-list mailing list