I'm working on a 5 channel video installation, where each channel is a tall vertical strip (1/5th of an HD frame). The projectors are all mounted sideways.
I need to read in the source movie, split it up, turn the slices 90 degrees, and send out 2 channels on one output, and 3 on another (TripleHead2Go units on each).
I was going to do this the hard way and pre-render everything, but it occurs to me that a far better strategy is to read the video, and send it to 5 videoplanes, using the tex_offset_x attribute to pick out the slices, and the rotate message to turn the slices.
Does this make sense? Is using multiple videoplanes and tex_offset better than slicing the incoming video using jit.scissors?