Advanced Texturing and Lighting with Andrew Klein

Week 4: Defining Textures, Pt. 2:

Texture File Size Issues:
RULE 1:
When working with our textures, we will want to keep our files as small as possible since the larger the texture file we reference into our material, the longer our scene with that material present will take to render.
RULE 2a:
Our UV layout SHOULD maintain a square format (with a few exceptions which will not be discussed just yet).
RULE 2b:
Furthermore, the size of this square texture file SHOULD be built in powers of 2 (2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, 2048, 4096, etc...). This is because the computer's architecture is designed to read files of these pixel sizes the fastest.
Now, according to rule 2, we have only a few Quantum sizes available to us, and (from rule 1), we will want to choose the The smallest file size possible that still gives us the texture complexity that we need.

SMALL FILE SIZES
32x32 Textures: These sized textures should only be used for backround objects that take up less than 10% of the screen, are barely seen due to speed, or are in an online gaming enviornment as a prop or bg object.
64x64 Textures: These sized textures, while twice as big, are still really only to be used for objects with very little screen or render prominence. These are also suitable for props of distant background objects, but are also very useful for small foreground objects in a real-time enviornment.
MEDIUM FILE SIZES
128x128 Textures: We now enter file sizes suitable for displaying more prominently on the screen. 128s should still not be full character textures, but may be used for objects that take up 20-25% of the screen, need more visible detail, or are more complicated props for a real-time environment.
256x256 Textures: This size is suitable for backround characters or parts of characters that have multiple UV regions for a real-time enviornment, either online or local. This size is also suitable for props and bg objects that get rendered, but may be too big a file size now for most gaming enviornments.
512x512 Textures: This size is perfect for most objects or props that are getting rendered, or a Main Character in an online game that is featured most of the time on roughly 40% of the screen. This size is FAR TOO BIG now for every single background object in a Last-Gen console or current online game.
LARGE FILE SIZES
1024x1024 Textures: This is the first of the larger sized texures. 1024x1024 textures should only be used in Last-Gen console or current online games if they are featured on the Main Character or for a whole enviornment, with some possible exceptions. A 1024 texture might also be useful for a Head Texture that is to be rendered.
2048x2048 Textures: This file size is mostly too large for real-time enviornments of previous generations, but is still suitable for PS3 and XBOX 360 development for primary characters. Usefull for more detailed single hand or head textures in a rendered animation.
4096x4096 Textures: This should be the largest file size we would work at in the current itteration of Maya. After 4096, it really is to the benefit of the user to break our character up into UV regions and apply multiple smaller maps. This size is capable of excellent detail for full characters, or extreme detail (down to pores) if used for just a human head.

Texture Terminology:

Value- Value is a measure of where a particular color lies along the lightness–darkness axis
Saturation- colorfulness is the perceived difference between the color of some stimulus and gray. A highly colorful stimulus is vivid and intense, while a less colorful stimulus appears more muted, closer to gray. With no colorfulness at all, a color is a “neutral” gray

Saturation can play a huge part in determining areas of prominence in an image/texture as well. Look at these two images below and examine how the focus changes when the same image is desaturated:

Hue- Hue is that aspect of a color described with names such as "red", "yellow", etc.
The Importance of Tonal Range- Click here for an important discussion about using the historgram and curves.

In general, I approach the texturing/sculpting process in 3 distinct steps that allow me to individually focus on the needs on my project. You can see a great example of this process in action in my RUG TUTORIAL here and listed below as well. In these 3 steps, I will examine different overlapping parts of the same area. A good texture (YOUR TEXTURES FOR THIS PROJECT) should show evidence of a cohesive combination of all 3 of these levels of detail. For more on creating the "lived-in" look described here in Level 3, take a look at the animated image/description at the bottom of the page in the Technical Data Section.

Also going into this assignment is the fact that each of your maps may look look remarkably different from one another. For instance, lets take a look at the image below.

1- The first difference that we can notice is between the diffuse color and the specular color. In the color map we see that because of the wearing on top of the diamond plate, the surface has become darker in terms of it's visibile diffuse color. Yet, due to that very same process and because of the constant wearing down of the top surface making it more polished and smooth (and thus shiny), in the specular map the tops of each diamond are brighter than the surrounding areas.

2- The second difference we might notice is between the color map and the bump map (which becomes the normal map through CrazyBump). Here, the tops of every diamond need to be brighter to make the surface raise up higher in the bump map, where in the color map, this needn't be so. Also you will notice that all the dirty and grime really doesn't have enough of an effect on the height to be considered, leaving the bump map with just two tones, the mid gray (no-change in height) and the white (up).

This is an example of the differences you will have to encounter and should be able to create on this project.

Texture Creation Tutorials (for your consideration):

Understanding Specular Maps:

Color Maps for a Character: Zbrush and Photoshop (ADVANCED):

Creating a Rug using Maya's Transfer Maps and Photoshop Filters (MORE BASIC OF TECHNIQUE ABOVE):
(for more on this, visit the Week 5 class notes as well)

Wood, Stone, and Rock using Photoshop Filters:

How to make a tiling blue-jeans texture in Photoshop (see Example 4 below):

Creating Patterns in Photoshop:

Creating Hair, the Epic Studios Way with Maya's Visor:

Texturing Old Paper with Photoshop:

Texturing Metal with Photoshop:

Skin Texturing tutorial (click on image below):

Skin Texturing tutorial (click on image below):

Example, the Desk Handle.
In this example, we use mesh-bake techniques and photoshop filters to create the different textures

Here is my reference photo:

From a combination of Zbrush and Photoshop work, here is my color texture-

By altering the levels and color saturation/hues/values, here is my specular color map:

From a Zbrush sculpt and document export using the NormalRGB material, here is my normal map:

And a Render of all he maps together, lit, on a flat plane:

Example, the leaf:
This technique looks at using Photosource to break out different texture files

In order, here is my Alpha, Bump, Spec, Down color, and Up color maps. My leaf is rendered using the flipped normals command for shaders:

Example: HOW TO CREATE AN ALPHA CHANNEL:

Technical Data for Creating Better Textures:

MipMap files:
Create Smaller file sizes by using .DDS files with NVida Mipster Plugin for Photoshop-

The point of MipMapping or Pyramid Mapping as it is sometimes called, is to create a single image with several "layers" of resolution so that when rendering, your scene can be optimized to the smallest size possible. For instance, if you have a 1024x1024 image, then saving a mipmap'd file with 3 states will save a 512 and 256 image as well for ultimate scalability should you zoom out. We will look at some examples of this in class.

This plugin can be downloaded at: http://developer.nvidia.com/object/photoshop_dds_plugins.html

Once it is installed (this is the same plugin used to generate normal maps using the normal map Filter as well), you can take you texture file, and choose Save As...
When your dialogue box opens, choose .dds as the file format. The point of a

Choose either DXT1 if you do not have an embedded alpha, or choose DXT5 if your image contains an alpha channel. I usually set the number of levels to 3, instead of leaving it at All. Hit "Save", and your .dds file will be ready!

CREATING TILEABLE TEXTURES IN PHOTOSHOP:

How to Tile a texture Without Seams-
CLICK HERE FOR THE LINK. Go from this:
to this:

How to Define a custom brush-
CLICK HERE FOR THE LINK. Go from this:
to this:

Creating "lived-in" Textures-
The Animated Gif below goes through several states which we will examine in class. The entire image is handpainted in Photoshop (does not use photo source) and is the texture for a wooden table.

	1. The base layer is created by using both the Motion Blur Filter and the fibers brush.
	2.Using a Multiply layer and a complimentary tone of blue, a soft pass was put down at a very low opacity to scuff up the table surface. Using quick noodling strokes and then erasing back in gives the best look.
	3.Using another multiply layer, the creased corner were darkened to simulate so of the effects that Ambient Occlusion would have on the surface. Alternately, bake your AO texture at this step.
	4.Using a Darken or Multiply Layer, now oyu can add further age on the surface and bake in shadows based on objects that might be sitting on the table, further faking AO.
	5.Next I want to scuff up the corners an give the look of exposed, non-treated wood. This is done by using a very small brush size in ligher beige tones, and stippling the surface. I use a Photoshop Bevel and Emboss Layer style with a color just slightly darker than the surounding main wood tone to offset this lighter section.
	6.With a custom Brush Shape, I add coffee stains in the upper right corner of the top of the desk. 7. Next, using another motion-blur based layer, I add further scuff from some lighter tones near the edges which might be sanded down over time. 8. Finally, I create a new layer, set it's blend mode to Overlay, and add some lighter tones, first in broad flat strokes, then with smaller noodling strokes to produce a dusty look.

ANTI-ALIASING and FILTERING for MENTAL RAY:
For your renders to display properly and not blury, in many cases we will have to alter the anti-aliasing settings as well as the pixel filtering. Shown below are two images with their Render Settings displayed as well. The only difference between the two is their settings. The first image has a lower quality filter (Box) and less samples for anti-aliasing. Notice the slight blur. The second image has increased it's filter to a Lanczos, and has more samples for anti-aliasing. While it takes longer to render, the result is superior. Pay attention to the pattern on the inside of the gun handle, as well as to the highlight on the barrel area, thats where the action is.

Here is another example with an animated Gif to show you the transition:

Project 3: You will be asked to take close-up Photographs of 5 "things" which are no bigger than 2 feet in size. ANY Photographs/Textures pulled from the Internet will result in a 0 for the project. For each photo, you will be asked to create texture maps for the object as needed. In most cases this will require at least a Color, Specular and Normal Map, but may also include a Displacement Map, Transparency Map, a Glow Map, or a Cavity Map. Each texture should be created for a flat polygon plane and at least 3 of the map types must be included in the 15+ total textures that you will be producing. You will then render out your textures on flat planes to compare them to the original photo. Due on week 6: 5 photos, 15 textures, 5 renders.

EACH TEXTURE SHOULD NOT BE LARGER THAN 1024X1024

FOR SOME EXAMPLES OF STUDENT WORK FROM THIS ASSIGNMENT, See Week 6

BACK TO TEACHING HOMEPAGE BACK TO COURSE HOMEPAGE

MAA 3312: Advanced Texturing and Lighting In this course the students will learn to apply traditional paint concepts, tools, and techniques for use in computer animation. They will develop critical ideas for surface treatment, texture, and lighting and demonstrate the layering of light in space to create mood, emotion and theme. Furthermore, they will demonstrate an understanding of global illumination, final gather, Radiosity, and HDRI. Finally, upon completion they will be able to critically understand how light affects a surfaces’ color based on material: Diffusion, Gloss, Specularity, Reflectivity, Translucence, and Ambience.