Hello, and apologies for what feels like newbie questions – I’m newish to jitter but not to computer video / vision (good in c and java, not objects and wires!) (oh, veteran max/msp user, otoh)
I’d like to load one movie file into ram, set 6 (or 8, or whatever) sets of loop points, and output six tiles of video, in a grid, in one window. Think Brady Bunch, but one character looping at different points in the file.
The output would look something like (excuse the non-monospace font):
| | | |
| | | |
(stupid font, grumble)
I’m looking for a container analogous to buffer~ in MSP; that is, something that is one block of memory that can be read randomly from different objects. I’m getting the feeling that this doesn’t exist in jitter. Will I really have to load the same file 6 times?
As far as tiling it, I’ve gathered that the popular way to do this is not with matrix operations but by texturing planes in OpenGL. This seems very strange to me as it is expensive (bandwidth-wise) to push all the texture information across to the video card. Or, internally, does jitter just use glColor commands and lots of little quads, even though we’re using the texture message? (This is how I’ve done video -> OpenGL in the past, in c++ and java)
Or – how would one do it with just matrix ops? It seems like most of them are for splitting (ie jit.split, jit.scissors, etc), not combining. jit.concat only takes 2 matrices.