You could go to OpenGL and look at using glCopyTexSubImage2D. It does not overwrite an existing texture - you set up a dummy texture, it allocate the memory, and then use that call to grab an area from the backbuffer to an area within that texture (doesn't have to be the whole texture size). It is apparently faster than a regular grab image. Still has some slowdown though.
It would be nice if Blitz Research would use that and add some Max2D commands to take advantage of it - ie GrabToImage or something.
It would be nice if Blitz Research would use that and add some Max2D commands to take advantage of it - ie GrabToImage or something.