New to UV-mapping, Cycles, and going around the bend

Just starting to dive into using Cycles materials and textures, which is the point of using Blender rather than my usual software: it’s texture mapping is very limited, especially with cylinders and other curved objects. This really is some noob stuff I’m asking, as I know some of the technical basics but not the necessary detail, and not how to apply it in Blender.

Attached is a sample blend for testing (this is from 2.76; I’m kinda-sorta holding off on upgrading until I see a stable 2.8). Basic contents are some dummy pipes with a flange set in the middle; all of that’s imported from a CAD model. What I’m looking to do is both line up the checker pattern nicely along the pipe and bend it nicely around at the elbows. I’m hoping that this can be done in a repeatable, structured fashion as the actual models will have a lot, possibly too many to practically work on one-at-a-time.

Recommendations on procedures? Tutorials that will help?

Attachments

Textures.blend (1.16 MB)

There’s no automatic way to align the UV’s to whatever the geometry is in each object. That I know of at least. Might be with some clever vector math but haven’t come across one yet.

One way that gets close is to produce a geometry for just the UV’s, and then use data transfer modifier to get UV’s from that on each object. The geometry doesn’t match so need to use some projection or interpolation method to do it.


There’s still a problem of bunched up coordinates that seems to follow the UV seam. Not completely sure why, or how to fix it with the least amount of effort.

Attachments

Textures_ja12.blend (545 KB)

Hmmm. Not entirely what I was after, but did help get me started - thank you.

I’ve narrowed down at least some of the problem, the end caps are composed of long, narrow triangles which are buggering up the UV generation (“it doesn’t flow” seems to be the best description). When I generate a seam around the edges of the ends it starts to look better, but there’s still a longitudinal seam required for best looks. And there’s a slight “twist” to the ring of faces at each end which seems to be shifting the end result a bit.

I don’t want to be that guy but Houdini would do that easy (probably)

Well, Houdini would most definitely do that easily same as Blender would. There are no problems with UV mapping clean geometry modeled keeping UV mapping in mind. The issue here is that the geometry is not clean as it comes from other CAD software. If I had to do it, I would start remodeling straight away. :confused: I know this sucks! But it’s the fastest way if clean UV maps are needed. It’s generally faster than trying to fix bad geometry. If you work with curves and use the existing meshes as reference it might not be that bad.

I face the same issues at my work all the time - I can get 3d models of furniture for interior visualizations from higher end furniture manufacturers, but they are never good - I always have to remodel them just to get UVs even in the rare cases if the models look OK. And I have tried fixing bad geometry so many times just to end up with bad results. It always takes way longer than to redo it completely. I have done it a lot of times - fixed bad geometry, did not like the result and remodeled anyway. It’s not Blender’s limitation it would be the same with other software - I have faced the same issues in Maya as well. If the geometry comes bad and unfit for quick UV mapping from other CAD tools there is not much that can be done. Remodeling can have other benefits as well like for example it’s often easier to add bevel to objects with clean geometry and that adds a lot to renders because it gives the highlights at the edges that our eyes are very used to seeing.

With clean topology it’s just selecting one loop across each pipe, marking it as seem and unwrapping. Add-on UV Squares can be used to straighten them in one click.

There is a chance the problem could be attacked at some other point of production - maybe it is possible to get results with procedural textures for example.

I was really hoping to avoid having to redo geometry that much. The end result is not intended for an open-ended schedule so time getting from “here” to “there” is a very real consideration. I’ve still got a few non-Blender alternatives to look at but Blender provided the better starting point.

Sounds like the best I can hope for is building and expanding my own library of parts needing to be replaced. Not everything is going to require this level of work, most will require relatively simple materials which looks to be workable at this point. At least I’m dealing with something that has well-established design standards. I don’t even want to think about furniture. :wink: