On 2/29/20 5:39 PM, Leroy wrote:
> I was making my own Mesh2 pulleys belts. Every thing work find or so I thought.
> I test the mesh2 using solid colors and color maps with Normals Finishs in a
> complicated texture.
> Then I found out that part of the mesh worked back on it self. When I fix that
> and tried color map or image maps it froze. That is :: if you try to translate
> the texture then put it on the mesh The IMAGEs don't move! But the Normals do!
> I've never seen anything like it.
> 
...
> 
> I've set the uv vectors from <0,0> to <1,1>
> 
> Anyone seen this before?
> 
Hi Leroy,

Not here. Perhaps someone using meshes regularly understands why you see 
what you see. I'll offer guesses/rambling in the meantime - while having 
a feeling you are using the mesh uv mapping mechanism in a somewhat 
unusual way.

- In my experience the mesh image maps would point into the image, 
texture or material with uv vectors in some sort of fixed / one time 
way. Perhaps this behavior is hard coded in a way that bypasses a 
texture's modifiers for image maps? The normals you specified would be 
outside any fixed uv image mechanism - if such a thing exists. Hmm, what 
happens if you use something other than an image_map with BTex[3] - some 
gradient x pattern with vertical stripes maybe? Does that then move with 
the normals?

- It looks like the BTexX translate is in x only. Have you tried an off 
axis translation to be sure you are not hitting some sort of harmonic 
with the image movement where it happens to be the same at every B_TexLoc?

- When you say you've set the uv vectors in the mesh, I assume you mean 
you are pointing mesh vertexes at the appropriate locations in the BTexX 
texture. I suppose you're aligning those with the gradient y in some 
fashion(1).

I've played with meshes in mapping other formats into POV-Ray and doing 
simple stuff color to vertex stuff, but, my understanding of how the 
POV-Ray uv mesh mechanism 'really works' is shallow(2).

Bill P.

(1) - Find myself wondering whether uv coordinates outside the <0,0> to 
<1,1> range would work... I can't now see any reason they couldn't 
function. Suppose, I'm wondering too, if texture movement could be 
accomplished by shifting the uv vectors for each frame. Today, we reload 
everything in the scene for each frame - so should work I think, and 
even if constrained in the 0-1 range. Hmm, Could make for some 
interesting animations shifting materials/textures around on an incoming 
mesh - blurring a mesh model's materials by merging frames. Anyway...

(2) - With mesh image mapping, wondered how the mesh mapping pros handle 
an image's interpolation seams. Are appropriate extra pixels carefully 
added to the images internal edges/cut-outs to be sure any 'image' 
interpolation results in a clean texture on the mesh at those seams? 
Does such a thing happen by side effect of mesh painting - is there some 
overspray area? I'm curious if anyone knows - otherwise, someday maybe 
I'll investigate. You've mapped without POV-Ray image interpolation / 
smoothing, so my question matters not in your scene.

Post a reply to this message