Multiple UV maps possible or necessary?

[RTSchramm]

I'm just starting to get into texturing using UV maps, so I'm a beginner at this. So I have three texture maps, one for diffuse, bump, and specular, do I need to use separate UV maps for each texture, or can I use the same one.

I have a complex shape that I want to texture. Am I correct in that I can only have one UV map per layer?

If there any good tutorials on this.

Thanks,

Rich

[Sensei]

I'm just starting to get into texturing using UV maps, so I'm a beginner at this. So I have three texture maps, one for diffuse, bump, and specular, do I need to use separate UV maps for each texture, or can I use the same one.

If they are the same, you should use just one..


I have a complex shape that I want to texture. Am I correct in that I can only have one UV map per layer?

No, you're not correct. There is no limit of vmaps..

[jin choung]

you SHOULD use the same one.

the only thing a uv map does is simply joins up with a texture and answers the question, "how does this 3d geometry line up with this 2d texture?"

here are the scenarios:

- a single texture can use a single uv map
- a single texture can use multiple uv maps
- many textures can use a single uv map
- many textures can each have separate uv maps

but whatever of the scenarios apply, you should have a reason to use more than a single uv map (or a single texture).
--------------------------------------------------------------------------
you can have as many uv maps per layer as you need, want or decided to create for some reason.

a uv map is a KIND of vertex map. so just like you can have many endomorphs, you can also have many uv maps.

but OCCAM'S RAZOR applies here. you don't multiply entities without reason. in other words, ECONOMY is always a virtue.

so if you have more than one vertex map, you have to ask yourself WHY?
--------------------------------------------------------------------------

THE DIFFICULT EXAMPLE:

let's say you're doing a whole character and you've decided that you need one whole UV page to unwrap and lay out the head.

but you've also decided that you need a whole texture page to layout the HAND because it gets really close to camera and there's some dramatic significance.

you can create a uv map named "UVMAP01" and unwrap the head to completely fill that uvmap from 0,0 to 1,1 and assigned to head geometry is a texture color map that is 1024x1024 that is all used to paint just the head.

and you can then unwrap the hand onto that EXACT SAME "UVMAP01"... such that if you had the head and hand visible at the same time, you would get an AWFUL OVERLAPPED MESS in the uv view.

but you can do this IF the head and hand don't share the same surface.

if you assigned the hand a separate surface and a separate texture (1024x1024 bitmap that just paints the hand), when the texture sees that it's assigned to the geometry of the hand, it will look for a uv map to see how the hand geometry is laid down onto a flat 2d space.

the surface is only looking for the geometry it is assigned to. so it doesn't care that there is also HEAD GEOMETRY there.

--------------------------------------------------------------------------

anyhoo, this subject has a lot of fine points to it that probably reaches the pinnacle of hodge podge for game developers (because they have to be SOOOO economical) but it's good to understand it all and to be as economical as possible.

jin

[RTSchramm]

Thanks for the info.

Rich

[toonafish]

You can use a single UVmap and a single surface if you add a Gradient Weightmap Alpha between the texture image layers to block out the rest of the surface you don't want affected by one of the image maps.

I personally prefer this method, because this way I don't have to change all my surfaces twice or more often when I'm tweaking the surface settings.

[jin choung]

right but that's just using the weightmap gradient as an ALPHA CHANNEL to blend between the two textures right?

so you have FACE_BEAUTY.tiff and FACE_BURNEDANDSCARRED.tiff and you're using the weightmap as an alpha channel to show block out one image to let through another?

you could also just use a bitmap for the alpha too but much easier to do it your way if you intend to stay in lw throughout (cuz otherwise, weightmaps don't translate over on conversions and other apps don't use them the same way as lw).

jin

[toonafish]

Actualy, an image would not work because the UV's of the hand and the head are overlapping in your example.
If you could use and image as an alpha, you might as well do it the easy way and use a single image for the head and the hand texture. :D

[jin choung]

ok, i can definitely see how what i said didn't make any sense but....

?

i'm having trouble visualizing.

ok, so if (on modeler layer 1) i unwrapped a head onto uv map "A" and the head totally fills up the uv map.

and then, (on modeler layer 2) i also unwrap a hand onto uv map "A" and the hand totally fills up the uv map.

and if i select layer 1 and layer 2, the uvs of head and hand overlap.

in YOUR technique, what does the single texture map look like?

jin

[toonafish]

in YOUR technique, what does the single texture map look like?

jin


A single texture map with overlapping UV's would look like crap.

So I wouldn't use a single texture map. Just like you, I'd use 2 different ones, one for the head and one for the hand.

The only difference with your technique is that I use a single surface and a weightmap as an alpha because a weightmap doesn't need UV's.

[jin choung]

OOOOOOOOOOOOOOOOOOOooooooooooooooooooooooooooooooo ooooohhhhh!

wow...

add ONE MORE PERMUTATION to the list of possibilities.

yeah, head and hand wouldn't work at all with an image alpha. the pristineface/burnedface would but the head hand is different when i thought it was the same.

thanks much toona! this never occured to me!

jin

[toonafish]

NP


took me 15 years to wrap my head around this :D

[othornton]

toonafish & jin: Don't forget that any gradient or weight map you put on a UV-mapped object can be baked out if needed in another program. I've used this several times to generate specific alphas for ZB that I want/need to re-use.

RTSchramm: One reason to use one UV-map is that you can use one image to cover two or more objects/surfaces. This can save a lot of memory at render time, especially if you're designing a game. Consequently a lot of game characters will have their UVs for their entire body laid out in one map even if the polys are separate pieces or a different material (skin, metal, fur, etc). Two or more UV Maps come in handy when you want to lay down a base texture over everything (say skin for example) then go in and remap a more limited region for a detail overlay (a tattoo that wraps around the arm, eyebrows, lips, etc) using a 2nd UV map.

With regards to limitations, the general rule is that a poly can only appear once in each UV-Map. So you could have a million layers all sharing one map, a million layers with their own individual maps, whatever. But in each map whichever points or polys are entered in can only be entered once in that map. This is similar to a weight map in that no points can carry two weights in a single map, but any point can carry one, two, or more maps.

-Oliver

[JeffrySG]

@jin and toona: I know you both just threw out a ton of great info, is there any chance you could post a couple of workflow pics to which you are referring in your posts? Kind of like a mini-tutorial? I know it would help out many people!

Thanks!
Jeff

[jin choung]

toonafish & jin: Don't forget that any gradient or weight map you put on a UV-mapped object can be baked out if needed in another program. I've used this several times to generate specific alphas for ZB that I want/need to re-use.

regarding toona's technique of using weightmaps to blend between two separate image maps on a SINGLE MATERIAL...

yah, you COULD bake out the weightmaps but it wouldn't help you to recreate the effect that you pull off in lw in any other app (that i know of).

jin

[nemac4]

Here is a scene showing the basic application.