VRayEnmesh modifier

[nicklas_byriel]

Vizioen Yea, probably, it's definitely a minor thing, and could just be fixed by subdividing - just wasn't sure if its supposed to respect smoothing groups or not and get facetted like in Anthonyyg 's attached example

[racoonart]

So I tried the new modifier but I'm getting all sorts of weird results. I'm not sure if these are real issues or if I'm doing something fundamentally wrong here. Is there something obvious I miss here or should I open a thread in 'Issues' ? Never used geopattern, so I might just misunderstand the idea.

[Edit] Welp, nevermind. I noticed that the gizmo was not enclosing the mesh entirely. Not sure how this happened.

[kosso_olli]

I'd love to know how this works internally. To me it's all magic.

[matanov]

So I tried the new modifier but I'm getting all sorts of weird results. I'm not sure if these are real issues or if I'm doing something fundamentally wrong here. Is there something obvious I miss here or should I open a thread in 'Issues' ? Never used geopattern, so I might just misunderstand the idea.

[Edit] Welp, nevermind. I noticed that the gizmo was not enclosing the mesh entirely. Not sure how this happened.

Yes, proper gizmo position and size are quite important for predictable and render-correct results.

[racoonart]

Would it maybe make sense if we had an "auto-gizmo" checkbox? So any changes in the target object would trigger the gizmo bounding box to re-fit to the mesh. This sounds like an issue that is going to happen all the time.

[racoonart]

Ok, I've made some tests but it's all flying apart everywhere not sure if these artefacts are expected or not.
Also, UV's beyond 0-1 seem not to be working, which is a bit of a problem when dealing with textures that are dependent on the scale of the object (like roof tiles in this case). I'd love it if it could work with real-world-scale uvs. The height percentage also seems to be absolute, not relative, so the lower my tiling settings, the flatter it looks and vice-versa.

[Edit]: also, some way of randomizing elements within the pattern would be very handy, like color uv offsets etc.

[sirio76]

[Edit]: also, some way of randomizing elements within the pattern would be very handy, like color uv offsets etc.

+1
That would be very useful

Don't know if is possible but ideally with color/UV randomization and an offset/rotation system it will be perfect for stuff like this roof tile

[racoonart]

The problem with roof tiles however is that currently you can't really tile them with this modifier unless you cut them somewhere in the middle - which then of course makes for ugly randomization. We would need tiling that can slightly overlap to make it actually useful. This is another thing I'd like to request.

[Vizioen]

Just use Railclone for these types of things.

[racoonart]

Railclone is fine for a few buildings but not dozens or hundreds of them (which is what I want to use it for and where it would be brilliant). Also, overlapping tiles would be useful for other things as well, like grass patches which would otherwise be cut off at the seams. Short grass? fine, longer grass nope.

[nicklas_byriel]

+1
That would be very useful

Don't know if is possible but ideally with color/UV randomization and an offset/rotation system it will be perfect for stuff like this roof tile

Maybe if it would be possible to use the VrayMultiSubTex's randomization options for Enmesh ? i.e. like using "Randomize by Element" if the MultiSubTex is somehow able to distinguish between the different "objects" Enmesh produces.

[tzaar]

Would/will it be possible for EnMesh to randomly pick from the multiple objects? I thought that was to whole point of adding a list, but it seems it creates a common bounding box for the and distributes them as a group. It will be extremly powerful if it can pick from a list based on color, like Forest Pack can do with it's distribution/cluster settings. I saw someone creating really nice crossnitting with pixelart driving the distribution in a competing renderer that has a similar function

This feature is so powerful and sought after that I'm considering moving back from Corona, if the displacement workflow and quality in v-ray could be more similar to Corona. Speaking of which, could it be worth considering implementing EnMesh on a material basis instead of as a modifier?

[vlado]

a simple sphere with low segments will be faceted instead of smooth for example.

I have a solution for this; it wasn't ready in time for the beta but I hope it will go into builds soon. Still, it would be helpful to post a scene or an image, just to make sure we are talking about the same thing.

Best regards,
Vlado

[nicklas_byriel]

vlado

I have a solution for this; it wasn't ready in time for the beta but I hope it will go into builds soon. Still, it would be helpful to post a scene or an image, just to make sure we are talking about the same thing.

Best regards,
Vlado

Ah my bad! - here is a quick example of a Sphere with 20 segments, with smoothing groups on, and enmesh modifier, pointing at a torus... The geometry from the enmesh sits flat on each polygon resolving in a faceted sphere of tori (torus's?) Obviously this example is a bit extreme with only 20 segments, but is sometimes still visible on more dense meshes

[datalost]

Hi! i just wanted to share my experience with the enmesh feature. As long as the unwrap is decent it can handle complex objects just as well as any map. But the fact that you are dealing with actual geometry gives it another dimmension, literally. It would be nice to have some kind of preview about the distribution even if it were with boxes like with the scatter feature.
Also, corpping and rotating the Sample mesh is doing ok. But, at least in my case, if I wanted to give the gizmo a bigger size than the sample, it hangs MAX whenever I want to render. The distortions in the paws and in the "hairy" part of the face are beacuse a lame mapping. I couldn't find a way to guess the right scale rapidly. I hope to have more time to make some more tests but overall it's wonderfull to be able to distribute a sample in a model without having a diplacement map and without having to do it with a particle system. Cheers.