I have an initial 3d structure, I distort it (adding noise to its vertices for instance, not continuously, but sporadically firing a new noises values)) and I’d like to go from previous state to the next one smoothly.
I found a way by using SOP to CHOP. doing all operations in CHOPs and then go back to SOP.
The blend family of operators is likely going to be helpful for you here. There’s both a blend CHOP and blend SOP. Adjusting the weights of your blends will let you move between inputs. I’m pretty sure there’s an op snippet for the blend SOP that should help you get started.
The idea is more:
I have a shape, I fire a new noise value (or values) and the shape is altered suddenly and distorted. (here with clay SOP but it could be more general question with other nodes)
I’d like my shape to be altered progressively from the inital state to the new one.
If I understand correctly, blend SOP provides a way to blend two SOPs which is VERY nice. But in my case, it would requirea kind of flip/flop structure that would push the initial shape to the first input, and the distorted next version one to the second, do the blend then the next next one to the first, blend etc.
Hello,
Coming from photography and cinema, I always prefer to,deal with images (so do my GPU), here are some noise Tops using switch to blend together and then Top to chop and chop to sop. noiseToSop.toe (4.2 KB)
But in this case, cpu is still involved for all CHOP operations, right? I understand that using your method, a big part of the process is on gpu even if you use sop to chop, then chop to sop to go back to SOPs territories.
In my early td head,
TOPs run on gpu
CHOPs run on cpu
SOPs run on cpu until further rendering
DATs run on cpu
MATs on gpu
(obvisouly, it is short-sighting everything but it can help to understand big picture. I understand that TOPs don’t run completely on gpu, I meant)
In Max, it is pretty clear for me to know which object runs on which part of the computer. Here, I’m too fresh and new.