Does this board not have a general questions thread? I don't feel this is important enough to warrant it's own thread but I don't see anywhere else to ask it.
Anyway, I'm wondering if in blender there's a way to either straighten out the curved bit of the uvs in my screenshot here so that the whole thing is straight, or do some fuckery with modifiers to make an arrayed curve shrinkwrap to it without having to apply the curve modifier first.
I'm trying to convert a track from one racing game to another, but they use different methods to get a track in game. The one I'm converting from uses two meshes, one for the bumpy physical road that the physics engine interacts with, and another much simpler one that the game renders visually. The one I'm converting to only uses one mesh.
That physics mesh isn't uv mapped at all, so I want to transfer the UVs from the simpler visual mesh to the complex physics one, so that the physics one would look like the visible one in game. I don't think that's possible in blender since they have different amounts of vertices, so I figure I'll probably just have to make a new road from scratch and then shrinkwrap it to the physics mesh.
The only way I know how to texture roads easily is to make an array of planes, texture that and then warp that into a track shape with a curve modifier, which I've done already. The problem comes when I want to shrinkwrap the new road onto that physics mesh, it obviously messes up the UVs. If I didn't have to apply the curve modifier first in order to apply the shrinkwrap modifier I could just unwrap the (now shrinkwapped) mesh into a completely straight road shape, adjust the scaling, and then reapply the curve modifier and it would end up as an exact copy of the physics mesh but with uvs that allow me to texture it exactly like the visible mesh.
Is it even possible to get around this in blender? Sorry if this makes no sense, it's kinda hard to explain.
It's a laser scanned point cloud of the track the devs ran through some algorithm to make a mesh out of.
I'm trying to play around with softimage mod tool now since I read somewhere it can transfer uvs between objects of similar shape but different vertex count, but it doesn't seem to be working.
Does 3dsmax have a way to doing what I'm trying to do? I have a year or two left on my student license from college but I don't have it installed to check.
Forgot to mention this, but this is what the visible mesh for that same section of track looks like.
I want to take the uv mapping from this, and transfer it to that mesh in the OP pic. If I can't do that I want to remake the mesh you see in the OP using an array along a curve in blender so it can easily be textured.
You can see here on the left how the visual part is textured, the mask for the asphalt texture follows the curve. I somehow need to get that mess of vertices to do this.
It is projected from view in the OP pic, yeah.
I guess I can try that but I have a ton of these tracks to do and was hoping there would be a quick way to do it. I'll keep looking I guess.
do this op. just a reminder though, do U > Reset on everything before you do follow active quads.
it bases how it unwraps on how the active quad is currently shaped in the UV editor, so if the quad doesn't currently have square UVs, then the unwrap will come out skewed when you follow active quads.
They're not flat planes, but the difference in height between any two given points is pretty small so that might work.
I tried that a few different times and it never seems to do anything for faces more than 2 or 3 faces away from the active one, I guess the mesh is just so fucked up it doesn't follow it well.
It turns out there might be a way to implement the original physics mesh + visible mesh in the other game. That would be ideal since now that I've been reading up on the game I'm trying to convert these into the more I think it might not work well to have a 1:1 copy of the mesh since it uses the geometry to determine how rubber gets laid down on the track visually, and it assumes square/rectangular quads to do that (maybe). I'll try it tomorrow and see.
If not I think I'll probably just convert the centerline of the visible mesh to a curve, then curve an array of smallish (6"x6") squares that curve, shrinkwrap those to the laserscan data, and call it good. If you could apply modifiers out of order without it fucking things up it would be easy to get a 1:1 copy but as it is I kinda doubt this is even possible in blender.
I have a gen question:
In zbrush's uv master it has a "attract from abmient occlusion" button which just blams some ocntrol paitning into the dips and crvices of your mesh for UV purposes. What I wanna know is if there's such a button for regular poly painting. I dont want to go through the tedium of setting up surface materials in maya or using projection master in zbrush if i can just get a quick and dirty AO as a base to begin paintng. I am aware of the automatic masking for brushes that lets you paint only dips/bumps but is there like a paint bucket version?
Just go to the Masking section in ZBrush and do a "Mask by AO". There's all kinds of different auto-masking features there. Once it's masked, just reverse the mask and you can now only paint the ambient occluded areas. Just do a "Color>Fill Color" if you want to do a basic AO fill after masking.
I'll take this chance and make my stupid question here:
What program would one use to make a concept like this?
I use Blender but I'm pretty bad, would it be hard to make something like this in Blender? Any good tutorials for this?
I was gonna give up and just have two layers of polys since it is apparently possible, just a huge waste of resources, but then someone else beat me to the punch with a different track and did exactly what I'm trying to do and straightened out that clusterfuck of polys.
I'm wondering if what they did is only possible in 3dsmax or something but I'm going to try and contact them to find out. I'm still hoping for a blender-based solution if possible though. Thanks for the help so far senpai.
Sorry to necropost, but I finally got to talk with the guy and he said he used spline projection mapping to do it.
I've been looking for how to do that in blender and all I can find are other people complaining that they don't know how to do it in blender either. Is it possible or am I wasting my time looking for it?
How to I connect the two points in Blender so that the face is split in two?
You are one of a kind.
I'm kissing your eyes and caressing your neck.
how do i get it so my texture doesnt bend like this? i want a straight line across the side of the car but instead it bends the texture. im sorry if it's a stupid qusetion but i am a noob at this.
loop cuts. ctrl/cmd+r
read up about topology. almost everything should be a quad (4 edges)
In exactly 2 places.
1. When you're merging edge loops and you've found a place that's pretty flat and isn't going to be deformed by animation or subdivision algorithms.
2. When you seriously don't have any other choice and you're forced to add a tri here to cleanly merge loops, so you decide to put it in the place that deforms the least by animation or subdivision.
Incidentally I would also call UV map unwrapping "deformation", but as long as you limit the number of tris you use to the minimum, then you will be fine.
The logical reasoning as to why you need to do things this way is that mathematicians wrote the animation, subdivision, and other mesh deformation algorithms.
Mathematicians found that it would be a lot easier to write these algorithms if they assumed that every polygon was a box-shaped, flat quad.