2 screens wide you mean?
I have seen an algorithm somewhere, not on Blitz I don't think, that described how to do a basic land-drop which involved some binary operations such as AND's and OR's, which would work on several pixels at once, but that was in the days of accessing the video memory directly with the CPU.
Like someone else said, it might be more efficient to grab verticle strips of land into images and then drop them. Bearing in mind that each image has to be minimally 64x64 pixels.
You could do better in Direct OpenGL, with a single texture for the area that is falling, and changing the texture coordinates for several vertical lines. That would be a lot faster.
I have seen an algorithm somewhere, not on Blitz I don't think, that described how to do a basic land-drop which involved some binary operations such as AND's and OR's, which would work on several pixels at once, but that was in the days of accessing the video memory directly with the CPU.
Like someone else said, it might be more efficient to grab verticle strips of land into images and then drop them. Bearing in mind that each image has to be minimally 64x64 pixels.
You could do better in Direct OpenGL, with a single texture for the area that is falling, and changing the texture coordinates for several vertical lines. That would be a lot faster.