Ok, I asked a few questions on here before about alpha blending and I now have a rough (non optimised) version going at about 18fps in 16bit on a bitmap (no transparency) of 320x240. As reading from video memory is so slow, I am doing it like this: I create an offscreen surface (the same size as my screen res.) and blit everything to this (from surfaces also in system memory). That way I can perform some pixel manipulations without the overhead of reading from video memory. When everything's done, I blit this surface up to my 'back buffer' which I then flip to the primary surface, also in video memory. Basically, what I want to know is can this system be improved? Can any steps be taken out? Thanks for your time, -- Russ Edited by - Russ[x2] on 1/8/00 6:56:50 AM Edited by - Russ[x2] on 1/8/00 6:57:42 AM

I guess you could place you''re backbuffer in system memory and totally skip the offscreen surface, but I''m not sure.

Anyone: can it be done?