Last week I talked about how to build a texture library with your own photographs. This week’s post is about how to manipulate them into something you can use in either a design project or 3D modeling. Once you develop skills with Photoshop or its clones, it becomes faster and easier to turn a half-decent texture into a great one yourself rather than scouring the internet for the perfect image. Plus the one you make yourself will be unique to your project, saving yourself the embarrassment of finding out you’ve used a stock photo or texture equivalent to the Wilhelm Scream.
Websites like tuts+ have tutorials in the bag, so I won’t go into heavy detail here. Instead I’m going to compile a list of tips I have to remind myself of over and over again:
The most important part of the texture is the texture.
Colors can be fudged. Pay attention to the pattern instead. I often use multiple images, one for color matching and the other for the texture. Photoshop’s color match tool is great for making sure a set of textures blend with each other, color wise. I also use layers to make color changes to the image underneath – more often than I change the color of the actual texture layer. Every photo editor comes with tons of color manipulation tools, so you don’t have to worry about finding an image the right shade of green. You can take a washed out, even black and white image, and make it the right shade of green.
Textures, on the other hand, are a lot of work to mock. So if you find yourself in a bind between the right pattern vs. the perfect color in tons of subtle gradients, desaturate the patten you want and overlay the other image for its color effects. I have tons of images that are useless on their own, but keep them around for layer effects, which brings me to my next point.
Layers are your friend.
Whenever you add a new element to your image, put it on a new layer. Also, don’t merge layers until you have a specific purpose for doing so. If you’re unsure, make a new copy for the merged layer and save a backup of them unmerged. I’ve done drawings that use 40+ layers, and that’s nothing for some illustrators. If you find them starting to bog you down, get into the habit of grouping and naming them properly.
Layers are also great for replicating the real layers on an object. If you have a bed filled with straw, you can lay a picture of fabric over a picture of straw, which you can then erase holes in to make the fabric look worn through. To have straw poking out of it, add another layer to draw wisps of straw over the fabric texture. Mimicking paint on a canvas is similar. You have the canvas on one layer, and paint on one above it, then blend the paint layer into the canvas to make it look more realistic. Too many people forget to blend the two, making the painting look fake like a decal.
I learned how to work with layers from the Masters (Classical and Neo-Classical), who layered pigments of varying translucency to mimic the subtle glow of human skin. In oil painting particularly, they didn’t just paint with one or two colors and call it a day. They started with a warm undercoat and added layers of color on top. Since I transposed that idea into using digital layers, my digital paintings became a little more lifelike.
Make your own tiling textures.
There are a lot of tutorials out there and multiple techniques, but anyone not using the offset filter doesn’t know what the hell they’re doing. In Photoshop it’s found under Filter > Other > Offset. Most Photoshop clones should have it, or at least have an added plugin you can download somewhere. To use the offset filter, figure out how big your image is and split it exactly in half, down and across. This will offset the layer so its edges meet in the middle in a big cross. Then you use the clone tool to blend them. Don’t forget to remove any obvious repeating details. Standing back or zooming out will help you catch them, because it’s easy to get tunnel vision when you’re working with it up close.
Start as high-res as you can and keep it that way until the last minute.
Also keep your high-res versions. Downsampling is always easier than up-sampling. It is possible to turn a low res image into a higher one, but it’s a lot of work. I’ve resurrected old textures by enlarging the image (most people stop there, when all they’ve done is turned a pixelated mess into the same pixelated mess but double the file size) and then overlaying a higher res texture overtop. It’s a useful skill to have if you’re modding an old game and want to keep its original aesthetic intact, but for anything else it’s easier to dump the old textures and start from a higher resolution image in the first place.
Next week I’m going to talk about the bigger picture: aesthetics and design.
by 
Thanks for the information. I’m waiting for my brain to be available so I can see what Pixelmator has in its Effects that might give me some interesting surface textures.
I’m already reasonably good with the Color effects – I know which controls to tweak.
Keeping things in layers is key – if I want them merged, I make a copy, merge that, and sometimes start a new file with that merged image. Don’t ever give up layers if you don’t have to.
I can’t even imagine working without layers anymore. I remember a conversation with someone who was reluctant to modify a texture he already made, because he put so much work into it. When I told him to make the changes on a new layer, he said, “Oh, I don’t use layers.” My reaction was pretty much, “Are you nuts?” Of course he was reluctant to make changes if he handpainted the whole thing on one layer with no saved iterations or backups. The idea still makes my blood pressure spike. Gah!