simone conforti

I would like to know if anybody has already tried to replicate the functionalities of jit.dimop inside jit.gen via the code box.

What I’m basically trying to do is to do the sum of some matrix cells, while I’m at trying to use jit.fft in order to subdivide the fft window in a bark scale, and therefore adding all the cells within each bark band.

What is seems is that while trying to do the calculation with += is that the result belongs always to the value of the same cell and that no accumulation happens.

So in order to simplify the task, is anyone able to do the sum of a plane using a += in codebox?

I would like to know if anybody has already tried to replicate the functionalities of jit.dimop inside jit.gen via the code box. 

What I’m basically trying to do is to do the sum of some matrix cells, while I’m at trying to use  jit.fft in order to subdivide the fft window in a bark scale, and therefore adding all the cells within each bark band.

Hello Maxers,
I would like to know if anybody has already tried to replicate the functionalities of jit.dimop inside jit.gen via the code bo

replicate-jit·dimop-in-jit·gen-via-codebox

I am dealing with a similar problem: how can accumulate values in the jit.gen codebox (or outside)?

I'm afraid I don't think that accumulating values is possible with jit.gen operations. If I understand correctly, each operation happens (or should be assumed to happen) simultaneously on each cell, so there's no way to keep a 'running total' within the process. Fantastic for speed, less suitable for iterative processes.

There is a way to do it, but it's a bit tricky. The main thing is that the output dim/type/planecount of jit.gen is determined by the input -- the function defined by the jit.gen is run for each output cell. But if you use @adapt 0, the jit.gen can process over other sized matrices in additional inlets. That way you can achieve processing over ranges of data with for loops etc.:

There is a way to do it, but it's a bit tricky.  The main thing is that the output dim/type/planecount of jit.gen is determined by the input -- the function defined by the jit.gen is run for each output cell. But if you use @adapt 0, the jit.gen can process over other sized matrices in additional inlets. That way you can achieve processing over ranges of data with for loops etc.:

for having shared your attempts and thoughts, and many thanks Graham for this nice solution!!!

I will post the code of my post as soon as solved.

How to accumulate pixel values in Jitter/Jit.Gen world ?

I would like to learn the best approach to accumulate/sum pixel values in jitter/jit.gen world.

Imagine the following hypothetical matrices:

M1= [0 0 1 1 2 2 2 2] this is coordinates;

From the positions of the numbers in M1 I wish to sum all the values in M2 that share the same position.

In this example the result would be [2 2 4].

Before you reply use ZL objects, yes, I know how to do it in the Max landscape, but I wish to know how to do it in the jitter and jit.gen landscape, if it is possible.

I saw Graham example and I think that's very close to achieve what I'm looking for in this example, but I haven't managed successfully to translate those concepts into this particular example.

jitter jit.gen accumulator.maxpat

How to accumulate pixel values in Jitter/Jit.Gen world ?
I would like to learn the best approach to accumulate/sum pixel values in jitter/jit.gen world.
Imagine the following hypothetical matrices:
M1= [0 0 1 1 2 2 2 2]   this is coordinates;
M2= [1 1 1 1 1 1 1 1]   this is values;
From the positions of the numbers in M1 I wish to sum all the values in M2 that share the same position.
In this example the result would be [2 2 4].
Before you reply use ZL objects, yes, I know how to do it in the Max landscape, but I wish to know how to do it in the jitter and jit.gen landscape, if it is possible.
Perhaps with codebox?

I would do it with JS, but if you want to do it with gen that's one option: 

replicate jit.dimop in jit.gen via codebox