I was creating the textures for the iPhone Game that we are making. Here are some thoughts about the work flow.
One complex part of the workflow was the “Texture Slicing” part.
You see all the graphics of the iPhone game that we are creating, were created using the amazing Blender 3D. For the 3D objects that will be used as textures inside the game, no big deal. I just render the 3D objects in a i.e. 512×512 pixels image with alpha map. There is no problem with that. We can import them in the game Engine and use them.
The complexity comes through the need to create some objects that should have two forms. The one should be “as it is” and the second one should be “sliced in pieces”. The Game Engine should do the animation of the sliced pieces, so I had only to create the slices as “textures”.
The first thing someone can think is “ok, so what is the big deal?”.
Practically things are a little more difficult than it seems with the first look.
Lets examine the situation.
The result of some renderings where 512×512 pixels images. Also there is a situation that some 3D objects need to be sliced into pieces and then be rendered. Then those rendering with the object pieces should be used to create the textures. I had to take each piece and fit it (without resizing it) inside a “Power of Two” dimension constrained image.
How should I slice the pieces of the 3D Object? What should be the workflow to create correct slices that could fit exactly in a various range of image dimensions that are constrained by the power of two?
First of all I created some basic rules that I found it would be useful and I should follow for the 3D Object slicing:
- Each texture piece size should have a dimension constrained to the “power of two” numbers (i.e. 512×512, 256×256, 128×128, 64×64, 32×32, 256×128, etc).
- Each texture piece should not leave much “empty space” inside the texture. It should “fit correctly” inside the texture dimensions.
- The texture pieces (the slices that consist the 512×512 rendered 3D Object) should not exceed in total the size of the original 512×512 pixels image (for “optimization” purposes). (this is a logical rule but there are some situations that I finally had to make sacrifices and disobey that rule).
- The 3D Object “cutting” is supposed to be somehow “artistic” and not just a square lasso cut.
- The 3D Object pieces should have different sizes (but they should follow the power of two rule at the same time).
So here is the big question. How should I easily cut the 3D Object in pieces and at the same time follow the above rules?
I tried different approaches on this “problem”. I was trying to find a way – a work flow - to do the 3D Object Slicing as quick and as easy as possible.
After a few tryouts, I “synthesized” a 512×512 Texture Slicing Guide Map image from Pixel groups with different colors. It was a guide to help me do the slicing more easily. Each color represent a “Power of Two” dimension from 2×2 pixels to 512×512 pixels (or even bigger).
For example the gray color is one big square 512×512 pixels image and in front of it, it have one big 256×256 blue square at the top left corner and one big 256×256 blue square in the bottom right corner. The viewer can see only the two 256×256 parts (upper right and lower left) of the big gray 512×512 square because the other parts are in front of it.
Here is the image I created:
With this image I created I could easily identify “power of two chunks” of the 3D Object to do the slicing and at the same time I could take the sliced pieces and test them inside the “Guide Map” and see if they were cut correctly and that they will fit well inside the “power of two” textures. I simply put this image as a background image inside blender camera and begin cutting the object.
At the same time this image is very helpful inside 2D Image Editing programs (like Gimp, Photoshop, etc) because it can help you easily create guide lines that follow the “power of two” rule.
Please let me know your opinion about my approach. What do you think? Is there a better solution for slicing a 3D Object in “power of two” constrained parts (following “the rules” that I described earlier)?
What is your workflow?
You can download the “Texture Slicing Guide Map” that I created if you want. 
