Markus Baumknecht

I have a somewhat strange issue with jit.cellblock.

in the attached patch, I've tried to write a system in which data can be input into a jit.matrix for storage by means of a jit.cellblock object.

the mechanism itself should be quite simple: one can change the planes via the 'plane $1' message and then for each plane puts in the data in the rows and columns of the jit.cellblock.

everything else in the patch is built around this mechanism; when data is entered into the cellblock it gets passed on and stored in the matrix. that's essentially it.

I noticed that a strange thing happens though when I change the planes: the stored numbers change after a few times of switching back and forth between the planes (see screenshots).

I was thinking if it's an issue with the float numbers but even when I have it at 64 bit it's still the same.

can anyone help me on this? it's driving me crazy ...

plane02.png

plane02_changed.png

matching_logic.maxpat

the mechanism itself should be quite simple: one can change the planes via the 'plane $1' message and then for each plane puts in the data in the rows and columns  of the jit.cellblock. 

hi everyone!
I have a somewhat strange issue with jit.cellblock.
in the attached patch, I've tried to write a system in which data can be in

strange-data-issue-with-jit-matrix-jit-cellblock

Hi Markus, I can also find this strange behaviour with your patch. Without having made a deeper analysis I would recommend the following, as I am thinking the strange behaviour is due to the feedback loop from your matrix to your cellblock and some timing issues with different objects updating and sending messages.

a) Use the trigger object e.g. like [t b l] instead of [del 1]. 

b) Have a look if you can directly reference to the matrix by using the jit_matrix message. As far as I understand it you won't need the complex set cell mechanism.

a) Use the trigger object e.g. like [t b l] instead of [del 1]. 
b) Have a look if you can directly reference to the matrix by using the jit_matrix message. As far as I understand it you won't need the complex set cell mechanism.

hi Martin, thanks for getting back so quickly!

I can't believe I didn't see the reference to the matrix via the jit_matrix message, this is so much easier.

Strange data issue with jit.matrix + jit.cellblock