quote:Original post by Rah
Yeah I thought about that, but just thought it seemed a little inefficient (the final flip seems redundant as it could of been blitted directly from the sysmem surface to the primary surface.
I basically just wanted to know if a blit from an offscreen surface (equal size) will result in ''flicker'' effects? Does DDraw automatically wait for the vsync before a blit to the primary, or would waiting for the vertical blank be advised?
Thanks,
Russ
You cannot write to a surface and display it at the same time without flicker, so you cannot just blt all gfx to the primary buffer.
BTW, the surface->Flip() command is faster than anything you can imagine because all it does is switch pointers. In the final parameter for Flip() you can specify a flag (DDFLIP_WAITVSYNC I think) to make sure that the surfaces are only switched when it''s synced so you do not get tear.
You were right in your initial concern for blting directly to the primary, and that it would give you flicker.
As a general rule, if you are in fullscreen mode (just do an if test after rendering all to a back buffer) call Flip(), but if you aren''t, then just lpddsprimary->Blt(backsurface).
Hope this helps.
What is the best way to handle this? Should I still create a primary/backbuffer (both in vid ram), draw everything to an offscreen (sysmem) surface every frame, blit this to the back buffer and THEN flip? Seems slightly wasteful.
I realise that just a straight blit from an offscreen (sysmem) surface to the primary (vidmem) surface seems more straight forward, but I''m worried about the flicker posiblities.
Any clues on the best way to achieve this? I''m also planning on supporting windowed mode which makes this a little more complex 
