Hi,

I have been trying to figure out whether it is possible - in some reasonably easy way - to do low-frequency fractal displacements on geometry, and then add further details (from the same fractal pattern) in the shader.

For example, I have a grid that I want to turn into a realistic terrain, so I divide it into e.g. 100 rows/cols, put a fractal noise via vopsop on it. Then, I would like to put a shader on top of it, which would use the same fractal pattern, so that in the viewport I can have a fast, lightweight geometry, but on a render maximum detail.

I can do it all displacing geometry, but I would have to make at least 1000+ rows/cols on that grid - and this slows the viewport down significantly.

I'd be grateful for all hints & suggestions on how to approach such issue.

Thanks,
Greg

Look at the Anti-Aliased noise node in VOPs. This is fractal brownian noise which create terrain like look.

Then you can use a VOPSOP in SOPs to preview the look but when rendering you render a flat grid with a shader that have the same settings for the noise. This will then give you the same shape but with more detail and faster speed.

Hi Erik,

Thanks for your reply.

I know I can preview shader displacements. But that's so far from being user-friendly, that I am looking for a solution which would give me a low-res preview in the 3d viewport, and high-res results in render.

E.g. look now this is done in Vue - you can see a low-res preview in the viewport, while you can mix fractals in the terrain editor.

I'm not looking towards using a single fractal, but rather mixing multiple fractal noises of different frequencies for higher realism.

Anyway, thanks very much for your suggestion!
Greg

The Anti-Aliased noise already mixes multiple fractals. It is just to specify the number of octaves.

For a more user friendly way than having a separate preview and render object I don't know what to do.

For a more user friendly way than having a separate preview and render object I don't know what to do.

Well, since I can mark different node for display than for render, it can indeed be all it takes - thanks for your help.

Greg

Hi there,
aiworks - have you found a different solution than having one geo for display and the other for render? It's been some time…

I'm completely new in Houdini and Matnra but in Vray/MentalRay it's possible to have a level of displacement for work in the viewport and a render-time subdivision which takest things further. I don't know how to translate this to Houdini.

I'm aware of Houdini's “Render Polygons as Subdivision” option but it just lacks control and is not the same. I've spent hours watching all GoProcedural tutorials, including Terrain Masterclasses, before asking here, but this problem is still not solved.

Thanks for any clues

Hi @loogas,

I'm afraid I haven't been following Houdini for some time now, so I'm unable to say whether there are some new ways of achieving this in Houdini 15.

Anyway, good luck learning Hou!