Just saw another thread about blitting and remembered I never got a clear answer on this.
So which is faster when dealing with a lot of objects in a medium to large stage?
On one side there is copyPixels() to a canvas, pretty fast and reduces overhead to a minimum. Clearing the canvas is the place where it takes a hit, the bigger the canvas the bigger the hit. We can’t forget about handling rotation and scaling either, achievable with matrices but it’s a bit of a pain and it kinda kills the speed cause you need to use draw() instead of copyPixels (Unless every rotation is in cache, but that approach has clear limits).
Then we got beginbitmapfill(), I’m guessing that in raw speed it is slower than copyPixels() but you get to keep the ease of use of the display list and can reduce a considerable amount of the overhead that a Movieclip generates by using Shape as the drawing canvas for each object. You also get free rotation and scaling from the Shape.
Has anyone benchmarked these two? Or know of somewhere I can check those results?
