Cloning ICE instances?

[Pooby]

Hello.. Before I ask the question. I'd like to say I've extensively tried many option and I'm convinced that there is something I'm fundementally doing wrong but I can't work out what; so I'd request politely that any reply be really idiot proof, comprehensive and coming from experience of having achieved the result, not just point me in the direction of a compound that might work, because it probably wont, following my current approach.

What I want to do is firstly to take a mesh with UVS and use a pointcloud to distribute ICE instances of it. ( this bit is easy and I dont need help with this part)


Where I am struggling is then to make new Geo that effectively merges all the instances into a new mesh. Writing their UV's to a set that is on the new mesh. ( not just an ICE attribute)

Its the UV part that I cannot work out. I managed to do it once, but could not reproduce what I had done.

[rray]

I did this successfully a while ago. Forgot about a lot of the details though, so I'll repeat it and write down each step.

What you'll definitely need is "Convert Instances to Mesh(rcfix).2.1.xsicompound" from here:
http://www.si-community.com/community/v ... 8&start=20

brb

[rray]

Darned .. forgot that the way I did it was in fact using the whole ICEMaterial workflow and an UVW attribute instead of actual UVs.

You can get merged uvs however by copying the attribute over an existing UVW projection. The projection has to be created manually because ICE can't create projections by itself. The copying part has to be done outside the modeling stack.

what ICE can't do is create those clusters and cluster materials for you that you'll need to pick the right materials out of the ones you used in your instance masters/cluster materials. I don't think there's an existing solution for that. Although this could be scripted.

If you have only 1 instance master then it could be done manually.

Pardon the messy replies.. I'm editing these post as I discover things

[Pooby]

Thanks.

The bit that sounds like I overlooked (or wasn't aware of) is writing the UVs outside the modelling part of the stack.

I did at one point make a second ice tree to write the UVs and it worked. Maybe I fluked it in animate mode that time, and on subsequent tests didn't.

I'll try tomorrow.

Much appreciated. I'll let you know how I get on.

[Pooby]

Yes, that works. Thanks very much.

I can't quite see the logic of not being able to write the UV's in modelling mode, but as long as I know the rules, I can work with it.

[rray]

Glad to hear. I think all UV projections are applied directly after the modeling stack because they need to make sure the topology won't change anymore.
They seem to be part of a hidden "after modeling" area of the operator stack. So they woud overwrite anything that's written in the modeling stack (I'm not even sure if they exist there at all)

[jonmoore]

Nice one. I was attempting something similar myself recently and hit a brick wall of my own ineptitude/inexperience. Bookmarking this thread for future reference.