Multitouch data "pixelate" into a matrix

    Oct 04 2013 | 4:54 pm
    Hi, sorry for the confusing title.
    For example I have a 16x16 2D matrix. and I've pressed on a multitouch pad around (4,4) with a certain force, then the data in the (4,4) slot will be 1. And the slots around (4,4) - 8 of them - will have a number a bit lower (like 0.8). and the slots a step further away - 16 of them - will be lower again (e.g. 0.5).
    And in this process, I can use my force to control how much the date would be in the slots.
    Here in attachment I've made a simple drawing shows what I mean, there are 2 presses, the top-left is harder than the lower-right.
    Edit: the top attachment is the right one, how do I delete posted attachments?

    • Oct 04 2013 | 6:08 pm
      So uhm... what's the question? :)
    • Oct 04 2013 | 6:45 pm
      you could do that with jit.matrix - jit.expr a 7x7 single plane float matrix with something like expr sin(norm[0]*PI)*sin(norm[1]*PI) or expr 1.-hypot(snorm[0]\,snorm[1]) - multiply this matrix (jit.op @op *) with a float that is the scaled velocity (f.e. scale 0 127 0. 1.) - overwrite your main 16x16 matrix at the right offset with that matrix - jit.spill
    • Oct 05 2013 | 9:32 am
      @tobiasros Thank you very much! I've nailed this down.
      But my macbook pro gets so slow and lagging when I use this patch. Is this normal? I am totally beginner to jitter so I wonder if there's some setting or optimization I can do to the computer or the patch. Thanks.
      Here is the patch:
    • Oct 05 2013 | 1:08 pm
      here some quick improvements:
      call this abstraction placethematrix:
      this is main patcher:
      next you could f.e. put the abstractions in a poly...
    • Oct 06 2013 | 1:04 am
      i gave it another go: this one now is pretty fast on my 4 year old mbp
    • Oct 06 2013 | 4:47 pm
      Thank you @tobiasros! This is super awesome and fast.
      Btw, I found that if I have too many UI objects on my retina macbook, it can get slow very quick. I found out I can lug the system pretty efficiently just by dragging the jit.pwindow bigger... seems like a bug between max and retina macbook.