How to make textures like these?


After filling my bin with failure this week I looked at Fallout 3 clothing textures to figure out what I was doing wrong.

In this texture (which I don’t own etc etc etc) you can see noise, creases, diverse variation of colour (there are even pink and green bits if you look close!) in some areas the thread texture of the cloth is visible, in others it isn’t and it all blends together so well that its impossible to find the original base colour (If there ever was one). There’s also dirt, though I’m more interested in learning how to make clean clothing textures that look as genuine as this one does. How did the artist achieve all this? I can zoom in but there’s a level of randomness that I don’t understand, I only know that it looks good, not what causes it. I’d love to see a photoshop version. I really want to understand the techniques, the theory, and the process the artist used to get this result so I can do something similar.


You should post your failed attempts so we can help you troubleshoot where you went wrong and how to improve.

As for textures, it’s basically combining photo textures with painted textures to create the desired look. If you have basic proficiency in handling digital brushes and photo manipulation, this kind of texture work isn’t that hard.


I’ll see if I can get something to post but the trouble is that once I gone wrong with a texture I tend to abort it, I don’t know if that’s bad practice or not, just that it isn’t finished and it just ends up dead. I wish I had kept some of the fails.
I’m trying to unentangle the problems I’m having, define them each seperately, and I think the main problem is clothing creases; I tried to paint them on in photoshop and it looked like a sea-gull had shat all over it, and I tried it mudbox but it looked like a lower resolution sea-gull had shat all over it, and no matter how I faded or overlayed it the affect was bad.


Without seeing failed examples from you, it’s impossible to tell what you’re doing wrong. It could simply be that you haven’t developed the necessary skills for digital painting, but do have the required foundation knowledge/training of visual art, or it could be the reverse. It’s also possible that you are missing both.


Alright, ugh… I’m afraid my failed attempts won’t really help, trying to recreate this, I drop in the base colour and get to the point where I have to add the crease detail and I get held up by questions I don’t have the answer to.
So here’s my failed attempt:

Did he use sculpting or photoshop to make those marks? Tech or hand-drawn? That’s what I really want to know.


If you are a beginner, there can be a lot of questions overwhelming you to the point of paralysis. That’s totally normal, because everyone’s that confused in the beginning. It just take time and patience and dedication to get over that initial hump of feeling lost and frustrated (of course, proper guidance helps a ton).

The example texture you first posted–it’s really quite simple and I’ll demystifying it for you a bit.

Basically, take base photo textures that are shot with flat lighting so you get just the texture and no forms.

Then you just paint in by hand some subtle lighting information for subtle forms that you want to be visible on the 3D model–forms that won’t be modeled due to poly count limitations. You can do this by simply pointing in the forms on a separate layer (highlights and shadows, on separate layers each, so it’s easy to change one without affecting the other).

Or, some people might also utilize the dodge/burn tools directly on the photo layer (or a copy of the photo layer, in case the artist changes his mind and want to erase back to the original photo in some areas later). But beware that dodge/burn are tricky because they don’t just change the value–the also change the hue and saturation too, so if you don’t want unpredictable behavior, don’t use it. A better method I use is to duplicate the photo layer and use the levels or curves tool and create one copy that’s lighter in in average value, and one that’s darker in average value. These lighter/darker copies don’t deviate in hue/saturation nearly as bad as dodge/burn, so you just place them under the original photo layer and erase the photo layer to expose the lighter or darker layers underneath to get your highlights and shadows.

All of this is done by hand with simple Photoshop brushes, since it’s just very basic digital painting technique. There’s no need for any other type of tools at all–any decent digital artist can do this. If you can’t do it yet, then it’s just a matter of practicing your painting techniques until you can.

BTW, I used to be a full-time game texture artist for a few years early in my CG career, so what I wrote above is based on my professional workflow. :slight_smile:


Thankyou so much for this Lunatique, that’s some valuable instruction. The technique of having a lighter and darker layer beneath and then erasing through the base photo is inspired. I was trying to shade white and black onto transparent layer to make the folds but it looked like something a sea-gull had done. About the base photo texture, is that a tiling texture like these:

or cuts taken from a larger sample such as this

Thanks for this, really getting the info I need now :slight_smile:


If you were simply painting on the texture with white and black color, then of course you’re going to create an undesirable look. You need to put the highlights and shadows on another layer and set them to other layer blend modes, so that the original texture won’t become obscured. Often, highlights are set to screen mode, while shadows are set to multiple mode.

You also can’t just use white and black on every color, because the local color’s color temperature needs to be considered. If you use black on a yellow local color, it’ll look dirty and ugly. You’ll have to use a brown instead if you want the yellow to remain vibrant. You have to consider each color family differently, and use the right color to paint the shadows for them. Same with highlights too (though to a lesser extent).

As for what kind of photo texture to use, it depends on the scale you’re trying to depict. In the texture example you first posted, since it’s an entire suit at a fairly small resolution, you can’t use a zoomed in close-up texture as the base, since the scale would be totally wrong. You need to match the scale of the photo texture to the scale of the finished texture. If a patch of high-resolution texture is only capturing about 5x5 inch patch in real-life scale, then you have to make sure you scale that patch of texture so it matches the texture you’re trying to create. You can’t just use that 5x5 inch patch on the entire suit at once, since the texture will be huge in scale.

Think of it like a polka-dot pattern or a checkered pattern. If you apply a high-resolution patch of that pattern to an entire suit, you’ll end up with humongous dots or checkers. You’ll have to resize the pattern down to match the scale of the suit.


I may be way off and misunderstanding the original question, but isn’t much of that texture baked into the image ( lighting, wrinkles, AO etc.) via the 3D program?


Not necessarily. It depends on the style. High-end productions aiming for realism would use as much bump mapping in the models as they can get away with, or bake details from high-res model into the texture and then use a lower-res model. But for more stylized/cartoony, or casual games, so much is still just hand-painted, because that’s they look the projects are going for.