Why is Voronoi so sloooow?

[pokoy]

Since the map was introduced I find myself wanting to use it but find that it massively slows down rendering when used for clouds for example (right now from 20 seconds to 3 minutes for the pic I am working on, this is preview only).
Is the algorithm really that much slower than perlin noises? Is there any magic trick to speed it up? Like staying above/below a certain scale or something?

[Oshyan]

It's just a slower algorithm as far as I understand. Fewer octaves should help.

- Oshyan

[pokoy]

Thanks Oshyan. I'll have to play with the settings a bit more.
I hope there's a way to speed up the algorithm in general, though.

FWIW, there's a 3rd party noise map for 3dsmax (BerconMaps, look it up on Google if you're interested) that brings several flavors of noises, and its Voronoi implementation has some more settings and controls to choose from than TG's Voronoi, and it seems to be faster, too.
I have no knowledge of coding whatsoever, but from what I've seen the Voronoi in TG3 is hardcoded to produce a specific look. I can post some screenshots of the settings and noises available with the map I'm talking about, if anyone's interested (after my holidays, that is).

To be clear, this is not a complaint, it's just that I'd like to have a better understanding why these two Voronoi implementations are so different, and why Planetside decided to go for one specific flavor of the Voronoi noise.

[Oshyan]

Yes, we've seen Berconmaps before, and there is another library now that produces some novel and varied results too: http://www.codeproject.com/Articles/785084/A-generic-lattice-noise-algorithm-an-evolution-of
Both of those are things that I think people could consider implementing with the SDK since they are open source. We do want to improve the number of available basis functions and controls that are native to Terragen, but like everything it's a matter of time and priority. It's an important area, but there are many such things to consider.

- Oshyan