Texturing Theory, Methods & Layer Sets:
 
  Texturing Theory
 
  With the use of a Shader you can assign a color material to an object. But objects in the real world are not just one color. Painting a wall white doesnít make it appear as a white wall. How the wall is lighted plays a major factor in this, not only the color of the light source but direction the light travels to hit the object and how it reflects back to you. Plus if you have a green armchair near the wall, the wall may have a greenish cast where the chair reflects off the wall. As an artist you must be able to duplicate what you see and imitate it. We are not looking to paint a wall just plainly white, but to make the illusion that thereís a white wall behind the armchair.

  Now when you paint a 3D texture itís just like painting on a traditional canvas or at least the experience is relative the same. You apply color and shape and use a lot of the guidelines in traditional painting in texturing. Creating cohesion, depth, composition, and a conception of what the object is. Oh, and another rule that I go by, Never, Ever unless you absolutely have no choice: use 100% pure black. Black will flatten your object. You can still achieve coloring something black by mixing cool dark colors together like blue, green, and red or you can use warm and cool grays.

  With color comes other elements that need to be added. You need to show the history or age of the wall. Maybe the paint is peeling off or drips ran where the painter applied too much paint so the paint is not uniformly even. Also dust could have collected on it or itís dirty from things scuffing up against it. Speaking of that perhaps the paint has been stripped or has gouges in it from furniture thatís no longer there in its surface. Maybe the wall is not perfectly flat or is porous like stucco. All this historical elements adds age to the wall and makes it interesting to the viewer. You want people to notice your art not get over burden by it or bored with it.

 
  Lighting plays a key factor
 
  When creating your color map for a game, artists must keep in mind the intensity of the light value. Most games have poor lighting with single spotlights, area (omni) lights, or illumination. This could mean your model will not have the greatest lighting setup for it at all times in the level. Sometimes the intensity of the light is set incorrectly. But with the first two you can get at least some change in light values. However the major factor that will dictate on how you will paint lighting on will be the illumination setting.

  At 0% illumination this allows dynamic shadowing and lighting created by the Shader. This would be ideal as the lighting can change based on direction the polygon face is facing. But most game art doesnít get this ideal lighting because it takes calculation time, and the model might fall in to complete darkness if not enough dynamic lighting is set up (target spot lights or area lights). Rather most game models will receive about 60% to 100% illumination. The higher the self-illumination the flatter it will appear as the color map is being over saturated with none-dynamic light. So painting in or baking in the shadow information is one of the things that game texture artist must do or at least know how to do.

 
figure 6-01 figure 6-01 figure 6-01
    0% illumination     70% illumination     100% illumination
 
  So when painting your texture set the self-illumination to 100 to see how it will look as you paint in the shadow and light information. I have never seen a model with a texture that can standup to 100% illumination and fail under less than 70% illumination. So as a texture artist you try to estimate or predict how lighting will cast on the 3d model, faking a light source with our brush and brain.

  With faking lighting you are also trying to create geometry that really isnít there. Like a edge of a belt strap that you didnít model to save on polygons or fabric folds, or the fibers in clothe. This would all be really tedious to model but very easily done in a color texture map with a little skill. When doing this keep in mind that the surface gets brighter the closer it gets to the imaginary light source. When the surface gets further away it gets darker. In general most lighting comes from above since we are used to that orientation of light because of the sun. So anything that faces the ground would be darker, so knowing the orientation your model will be in game is helpful. Now by doing this baking in the shadow and light information your model will fail if the lighting or orientation changes compared to what you original painted it. That is the risk you take, as you know you will not have dynamic lighting all the time in game. You can get away with it if you donít go to heavy with shadowing and light on a object that light orientation changes or create two different textures with completely different light orientations.

  Another trick that texture artist will do is fake how light will react with certain materials. Steel has a very high fall off specular but can reflect the environment its in, while dyed wool absorbs light with very little fall off specular. Skin is even harder to capture as it absorbs, refracts, and reflects light. But skin is composed of several colors not just pinkish beige. If a wall is painted white its just not white, there are other tones in it. There is blues, greens, reds depending on the rest of the room and color of light. This is where a traditional painting artist can excel in, since they are better trained to recognize these things.

 

 
  Texturing Methods
 
  There are Four Methods of Texturing an object you created in 3D. Each method has its own advantages and disadvantage. They are: Procedural, Hand Painting, what I call Photo Painting, and Photo Stealing. The two painting methods run on similar theories, its just one is quicker than the other but might not produce the best result for what your using it for. Either method requires layers set up as base, elements, light/shadow, grunge, and history as its building blocks.
 
  Procedural
figure 6-04
  Using a Procedural texture is using the algorithmic code of a shader to generate the texture. Used mostly by the movie industry to simulate the surface it can accurately react with the light setup inside your 3d application. This can simulate a surface correctly but you need to build it up by using several sub materials stack on each other. This method relies heavily on real time lighting (which is lighting and shadowing created at the moment the scene is rendered). Another issue is that the surface is restricted to Material IDs, meaning a polygon face will be assigned to a certain Material in the Material editor, while a polygon face adjoining that one can be assigned to another Material and still be joined in the same shape. Now this doesnít sound so bad, but when it comes to games less polygons improves frame rate. So you canít have a row of polygons on a model because you want to have a gold edge. You also need additional geometry if your character you modeled is to have a simple different colored strip in their white shirt.

    Accurate simulation.
    Takes forever to render one frame depending on the complexity of the scene.
    No unwrapping required.
    Requires really great 3d lighting environment.
    Takes the lest amount of time to setup artistically (no unwarping).
    Requires a high polygon model to define not just edges but material areas.
    Not ideal for video game art.

 
  Hand Painting
figure 6-05
  Hand Painting is placing in the texture pixel by pixel and is the slowest process to do. But you have great control about pixel placement and definition. This is the preferred method of most video game texture artist, especially when doing a character texture. This is pure artistic skill driven and is related more to the techniques employed by traditional artists. You place in your colors and build up the surface to get the look that you want. You paint in the surface like as if youíre painting an oil painting but on the UVW unwrap image you created. An example is blue jeans get a blue hue throughout the whole area the blue jeans are in the UVW, you lighten areas of wear and parts that face the imaginary light source. Then add in folds in the jeans by coloring in some shadows under where the folds stick out at and lighten the tops to emphasis the fold in the fabric. You can even add in the yellow stitching where the seams are on the sides, pockets and crotch area. Now since youíre adding in shadows to create elements to the texture that was not defined by the original geometry such as the folds in the blue jeans. The area will have a flat appeal to it in the right lighting and camera positioning. This is unavoidable, the shadowing trick can not push out geometry, if your doing the cuff area of the pants and they create deep folds around the shoe or feet area, its best to add some geometry in to that area to make you folds. If this pushes you over your target polygon limit, then you might reconsider the extent of the folds. But coloring in folds around the crease in the area where the thigh and hip meet to indicate stretching of the material on tight pants is fine as the folds will not stick out as far, just be like tiny ripples. Subtly is best. But known how to paint a material is important or having an extensive photo library of images to study and to mimic.

  Another thing to watch out for is lighting changes. If you move from a dark brown interior to a moonlight snowy area you might not capture the lighting for one of the area correctly on the same texture, this is even more so with metal. This is fixed by creating texture sets, a group of textures that are assigned to models based on environments. This will increase the space needed and the amount of textures for a game but can be loaded and unloaded when needed. The most difficult lighting challenge is when the light changes not only in tone but its main direction. Like instead of majority of light coming from above, it rotates or flashes to coming from below and then back to or from the top again. This quick change can expose your shadowing and lighting you painted in to the texture is coming and going in the wrong direction.
Most video game characters use this method of texturing as it allows the most control.

    Accurate control of detail
    Artistic skill driven
    The Texture artist needs to know how to paint various type of material or have sample photo library to figure out how to paint that material.
    Longest process to do but the end result is worth it, mostly done for video games.
    Requires an unwrap UVW
    Can/must fake lighting/shadowing conditions on the texture.
    Will lock lighting condition on texture, may require texture sets for different environments. Which means another texture on the same UVW.
    May look flat in certain lighting conditions and camera placement.
    Great for characters and organic shapes

 
  Photo Painting
figure 6-06
  Photo Painting is so under rated. This is often times grouped with Photo-steal being called photo-sourced and doesnít get the credit it deserves. But this is a quick method to create a texture that sometimes works just as well as if you had Hand Painted it. This method is using the Alex Lindsay approach by using layers to apply your texture. You use base colors or maps and add on grunge maps, which are black and white patterned photoís using different blending options. You also use filters, masks and a little bit of hand-painting to build up a texture. The main goal is to use photos as a base and your adding too it. You can achieve almost the same result if you spent Hand-Painting for hours in just minutes. However you lose a certain type of control and if you donít have a large photo library of the right quality images you might not get a great quality texture. And the end result may look pixilated or blocky and blurry. Itís very effective to use this technique when in a crunch time period and you need to turn out lots of static or environment textures. But this also inherits many of the same issue as the hand-painting method like the lighting and faking elements problems.

    Texture might end up blocky or pixilated looking.
    Requires some artistic skill.
    Requires less time to create than Hand-Painting.
    Requires an unwrap UVW.
    Requires an extensive large photo library.
    Requires photo grunge maps.
    Can/must fake lighting/shadowing conditions on the texture.
    Will lock lighting condition on texture, may require sets of texture for different environments.
    May look flat in certain lighting conditions and camera placement.
    Great for producing lots of static or environment textures in crunch time.

 
  Photo-stealing
figure 6-07
  What I call Photo-stealing is the negative aspect of photo-sourcing. This is when the artist doesnít do any dressing up of the texture and just maps the UVWís to the image or cuts the image to the UVWís. This is by far the laziest and quickest method, but requires the best sourced image the artist can find. The artist will also have difficulties matching the UVW up to the texture. I have seen someone make a book texture by the use of taking pictures of the six sides of a book and it being applied to the model. Even worse each picture was its own texture. Imagine a six side wooden crate with six separate images at 512x 512. The Photo Source sample to the right is the rocks found in Grand Theft Auto San Andreas. These rocks look nice but are out of place compared to the rest of the textures in the game. Further with closer interaction the surface is flat and was probably done during crunch time or when the artist was forced to reduce geometry.
 
figure 6-08
  Now a good use of this technique is to use it on a Sky Box. A Sky Box if youíre unfamiliar with the term is a inverted box the surrounds the camera of the gamer, placing the camera dead center of the box. This box moves with the camera no mater where the camera goes and is the outside boundary that the gamer sees. Often clouds or stars are depicted on this box, where the seams are blended in so well they canít be discerned from the camera maintained distance. Most 3d games use a Sky Box or a low count Sky Geosphere to reduce rendering a intensive polygon hog object that you can never get close to. The Sky Box sample to the left cycles through how it would appear in game at 100% Self-illumination, where the seams are hard to notice. To where I just dropped in a Omni light set at 1.25 illumination to emphasis the edges of the box.

  This texturing method is easy to spot when done and looks poor, unnatural, or out of place compared to the rest of the environment. Often times its very pixilated and blurry.

    Texture will look absolutely flat.
    Lighting direction will be wrong for in the game or movie. Sometimes will even have conflicting light sources because the way it was cut.
    Laziest and quickest method for models, requires no artistic skill and it shows.
    Great if done right for sky boxes.

 

 
  Layer Theories
 
  The use of layers will help you separate your texture for easy fix after the texture was created. By breaking the object to its elements and then layers you have greater control of editing the texture once you started. Like for instance you want to increase the intensity of your shadowing but not sure how it will look. Well with shadow work on its own layer you can change the layer blend mode to Multiply or Darken to see how it will look if you went darker. Plus you also want to use the Folders (layer sets) to help group areas. Also if you introduce a color your boss or customer doesnít like, you can quickly change it by going to its layer and changing it or switching it off. If they didn't like it and want to go back you can just switch the layer back on no loss of work or time wasted.

  The use of Mask can also help keep areas clean and when using filter effects will have extra pixels to draw on. Masks are really effective on controlling the base color. That is they control the bleed edge that is outside the UVW open edge wires.

  I set the folder layer set in the following organization as the example.

      Alpha ( off )
      Folder Set Ė Selection ( off )
            Shoe only select2
            Shoe only select1
            Gloves only select1
            Shoe select2
            Direct shoe Set1
            UVWís edges (opacity > 60%)
      [end of set]
            Hair + Mask
      [Set Ė Bandana & Legs]
            Bandana Shadow Legs + Mask
            Bandana fold Shadowing
            Bandana & legs + Mask
      Set Ė Bandana imported from Illustrator ( off )
            Folding section4
            Folding section3
            Folding section2
            Folding section angles
            Folding section1
            image
      [Set Ė Shoe & gloves]
            Shoe & gloves Grime & dirt
            Shoe & gloves detail & decals
            Shoe & gloves with shadowing + Mask
      Layer Black

 

 
figure 6-09
  Now Iíll point out my mistakes in the example Iím showing. I didnít set the Shoe and Gloves in a folder set, and I did a bad thing and put the shadow information on the base color. It was too late before I noticed I screwed up. This means that it will be very hard to change the color of the shoes while still maintaining the shadowing information. I created the bandana base texture and design using a vector base program Adobe Illustrator to save time. Some game artist may see this as an issue as I then cut it up instead of hand painting the whole thing pixel by pixel. I didnít set the Bandana and Legs in a Folder set. Using Folder sets will keep the Layers window organized and help you know what youíre dealing with.

  Now what I did do right. I renamed everything from the default Layer ##. I separated the elements like shoe with similar color gloves, some people would even separated this more but I wanted the gloves and shoes to be close to the same value. I put shadowing and lighting information on a separate layer except for the Shoe and Gloves. I used masks when they helped the most on the base color. Dirt and grime was placed on a separate layer except there is not dirt on the bandana. I put my UVW Unwrap above the texture layers and set it the translucently to less than 60%. I also created sold shapes to easily select and notice color like red to create design elements like the lace holes and where the leather edges of the shoe in order to make the creases.

 
 
 
    Back to the Main Tutorial Page
 
 
3D Gallery   2D Gallery
Tutorials   Rťsumť   Contact
 
Copyright © 2002-2004 by Jeff Jonaitis. All Rights Reserved.
Best viewed in 1024 x768+ @16m colors; With IE 5+ Ns 7+