phiol

The patch below demonstrates a few things that are unclear to me.

2. I would like to be able to combing (add) two gradient matrices [jit.matrix 4 float32 640 1] + [jit.matrix 4 float32 1 480]

to have the same result as a single [jit.matrix 4 float32 640 480] .

1.
The patch below demonstrates a few things that are unclear to me.
All questions are inside.

2. I would like to be able to combing (add) two gradient matrices [jit.matrix 4 float32 640 1] + [jit.matrix 4 float32 1 480]
to have the same result as a single  [jit.matrix 4 float32 640 480] . 

adding-two-seperate-gradient-matrices-into-a-single-matrix

1. Keep in mind that different objects handle plane data differently. Some are argb while others are rgba.

1.    Keep in mind that different objects handle plane data differently.  Some are argb while others are rgba.  

you're right about the rgba and argb order. I 'm so used to working with gpu that it sometimes slips my mind. :-)

Any clue on how I would do it on the gpu using gl.pix .

I mean , as you've seen in the example, I could simply do this

but my surprise is more in the composite stage.

- How come that adding 2 gl.slab outputs does not give me the same results ? It does in Touch designer.

- how come , jit.gl.texture and jit.gl.slab give me 2 different results.

Hey thanks a lot for tips metamax,
you're right about the rgba and argb order. I 'm so used to working with gpu that it sometimes slips my mind. :-) 

You code is great.
Any clue on how I would do it on the gpu using gl.pix . 

I mean , as you've seen in the example, I could simply do this 

but my surprise is more in the composite stage.
- How come that adding 2 gl.slab outputs does not give me the same results ? It does in Touch designer.
- how come , jit.gl.texture and jit.gl.slab give me 2 different results. 

If you want to combine two smaller matrices into a larger matrix using [jit.gl.pix] you need to make the left input the same size as the desired output.

Instead of adding two matrices together, you add three with the 1st defining the output size. An empty left input matrix can be added to secondary inputs without changing cell values. However, secondary inputs can differ in planecount making it unclear how the object should map secondary planes to the left input matrix. Hence, you need to specify which planes/channels of the two smaller matrices are added to the main input to generate each of the output planes. I enclosed a patch with few different ways to do this.

Not sure about the slab and texture objects. I'm only now learning Open GL. Just make sure each object is getting what it expects. [*.pix] objects are better at remapping input alpha planes than other jitter objects. In your first patch, [jit.gl.slab] might be expecting ARGB and it's getting RGBA from [jit.pix].

If you want to combine two smaller matrices into a larger matrix using [jit.gl.pix] you need to make the left input the same size as the desired output.  

Instead of adding two matrices together, you add three with the 1st defining the output size.   An empty left input matrix can be added to secondary inputs without changing cell values.   However,  secondary inputs can differ in planecount making it unclear how the object should map secondary planes to the left input matrix.  Hence, you need to specify which planes/channels of the two smaller matrices are added to the main input to generate each of the output planes.   I enclosed a patch with few different ways to do this.   

Not sure about the slab and texture objects.  I'm only now learning Open GL.   Just make sure each object is getting what it expects.  [*.pix] objects are better at remapping input alpha planes than other jitter objects.   In your first patch, [jit.gl.slab] might be expecting ARGB and it's getting RGBA from [jit.pix].

I also get f**ked with and often forget about this 

But then again, over 3 years since gen came out and still no help file and example patches for each object.

Thanks again for the patches, they will help me learn codebox. Although my mission here will be to

I also get f**ked with and often forget about this 
But then again, over 3 years since gen came out and still no help file and example patches for each object.
not helpfiles for codebox.  

Thanks again for the patches, they will help me learn codebox. Although my mission here will be to
convert them back  into regular objects

Hey.. I agree. Some say that Max is well-documented. That may be... but good luck finding some of it. The search string "rgba" doesn't even exist in the [jit.gl.pix] help file nor the reference file. (!) Of course, then someone comes along and points to another doc - in this case, "Gen Overview" - which explains more.. but it still doesn't address the problem with the help and ref docs.

That said, the docs "Gen Overview", "Gen Jitter Operators" and "Gen Common Operators" are essential.

Also, some of my explanation in my previous post is just plain wrong (as far as why things worked/didn't work) but the patch should still work.

Hey.. I agree.    Some say that Max is well-documented.  That may be... but good luck finding some of it.    The search string "rgba" doesn't even exist in the [jit.gl.pix] help file nor the reference file.   (!)   Of course, then someone comes along and points to another doc - in this case,  "Gen Overview" - which explains more.. but it still doesn't address the problem with the help and ref docs.  

Also, some of my explanation in my previous post is just plain wrong (as far as why things worked/didn't work) but the patch should  still work.


>>Also, some of my explanation in my previous post is just plain wrong

(as far as why things worked/didn’t work) but the patch should still work.

>>Also, some of my explanation in my previous post is just plain wrong
(as far as why things worked/didn’t work) but the patch should still work. 

adding two seperate gradient matrices into a single matrix