hello julien .
did you toy around with the initial patch?
I am using td.rota for it's zoom and anchor point functions. if you play around with the zoom value , you' ll see that it's the same image zoomed in but starting from different anchor points. These anchorpoints are positionned using a norm.
when zoomed in , this gives the illusion of cut out pieces of the same image that are gathered together to create a puzzle.
So, yes it is the same concept that is happening here in gl.pix. But I find the equation to change the anchor points . I am not versed enough with glsl and gl.pix to recreate the effect in gl. pix.
maybe someone could help out.
moreover, why can't you use my jit.multiple version??
if you use jit.gl.gridshape @dim 2 2 . then you you can create quite a lot of instances before it become heavy on the gpu. But you wouldn't want that (many slices)with the scissor concept anyway.
hope this clearafies things
phiol