FX-Map / Quadrant puzzle

shuar

Hi, I'm playing about with FX-Maps and Functions in SD5, and I've can't figure out how to do something which I'm pretty sure should be possible.



Basically (please ignore the dodgy artwork, I'm only trying things out here!), I'm taking a plant leaf, and using an FX-Map to iterate it with a random rotation function. I'm then using another FX-Map to iterate each "plant" with a random size, position and rotation.

This is OK, but the leaf positions of each plant will be the same. I was kind of hoping that I could set up the first FX-Map like this:



...hoping it would generate a new "plant" with random positions which I could then position randomly with the second Quadrant. Of course it doesn't work like that because the input to the quadrant are just a half-sized output from the next node in the chain.

So I guess my question is, how can I generate a new random "plant" each time I place it? Or is this beyond SD's capabilities? My alternative is to duplicate the first FX-Map and keep adding it as inputs to the second one, but that feels messy to me.

lukors

I think there is no way of getting random seeds for each time the iterator runs. I have solved it in one of my materials by first creating a substance with the single item. Then another substance uses that substance eight times as inputs to an FX map with different seeds and picks one of them at random.

So it's basically exactly what you're doing, except that it's reused a substance, so you only need to do it once and then you're done.

NicolasW

Hi,

Generating a new plant for each drawn pattern in the fx-map is not possible.
What you can do is:
- generate multiple variation of the plant (duplicate, change the random seed)
- plug the multiple versions to the fxmap
- in the Quadrant: create a function in the "Input Image Index" to randomly choose between the multiple input

Nicolas

shuar

Yeah that's the way I've done it for now, cheers guys

lukors

NicolasW, is this something that you can implement or plan on implementing? It would surely be an interesting feature to have!

NicolasW

I'm afraid this will never be possible.
Basically it requires a loop function in the compositing graph, and this is not the easiest thing to implement regarding the current architecture..

But heh, never say never..

shuar

NicolasW wrote: »

I'm afraid this will never be possible.
Basically it requires a loop function in the compositing graph, and this is not the easiest thing to implement regarding the current architecture..

But heh, never say never..

That's ok, I don't think there will ever be a situation where I'll need so many variations that I can't just plug in new variations with a different seed, bearing in mind if I'm randomising the scale I can always scale half of them negatively to double up as well. Just wanted to make sure that I couldn't automate it somehow if possible!