I'm recording/displaying from six 4-input bttv cards in yuv420p(need this for recording). The problem is that I have only have one yuv overlay (www.libsdl.org), so I'm stuck memcpy the bttv yuv planar data into this one big yuv overlay. This takes ~20ms on a P3 866 (10ms overhead caused by the memcpy's), just as much as it takes me to do motion detection on them or recording! Is there a better way to do this? Keep in mind that the video from all 4-inputs of each bttv card gets put onto their own screen area/offset.
