Quote: --- Original message by: waffles
As far as i know, there isn't any more modern/recent information on CE's shader system (excluding the little snippets of things from cmt and me piping in every once in a while). There may be some old stuff buried in the gearbox forums.

As for the tutorial i started long ago, I just didn't have a clear idea on how to structure a tutorial based around the ENTIRE shader system, it kinda overwhelmed me in the end. A wiki is actually a good idea; one i never thought about. It would be a good project to compile accurate up-to-date information on pretty much everything relating to ce's tools. And I for one would be fine contributing to the shader entries, after the long hiatus of time I spent learning more about the blam engines.

I learned from multiple sources; stock shaders (campaign/multiplayer shaders), experimentation, math, terminology. Halo's shaders are still just math at its core, from blend modes to tinting and reflection. It also helped having experience with different engines and even 3ds max.

It just seems that the many resources we may have, are out of date or just scattered far and in between, if we could localize all tutorials/information in one place and update frequently it would eliminate the amount of searching in between. The halomaps database IS large, but not everybody can relate to the information. Theres also the fact of many similarly named files that contain varying information...

If theres anything in particular that you don't understand, I can try and clear it up for ya the best I can for the moment.

Quote: --- Original message by: waffles
Well, I don't have my original ripped xbox maps anymore after the transfer to my new desktop. And xboxdb doesn't seem to have the bitmaps.map thats required to grab those bitmaps in the first place. I don't have the resources i had with my old desktop to re-rip the disk again. I could go ahead and do something similar, but with a newer game where the resources are much easier to get. Kinda wanted to bake out the h3 energy sword shader and do a frame function setup.

Quote: --- Original message by: waffles
For the xbox shaders i did, it was done in photoshop with blend modes. And (as painful as it was) I manually moved the noise bitmaps in increments scaled by how large the bitmaps were to make sure it would loop properly. Then it was just alot of copy and pasting to prep the frames into one bitmap. Most of the math/work was done in guerilla with a function that snapped to each frame over half a second to a full second (depending on how many frames i bothered to do)

15 frames @ 0.5 seconds per frame = 30fps
30 frames @ 1 second per frame = 30fps

it all depends on how much filesize your wiling to spend on a large bitmap, or how fast or slow the shader would go. If I could ever just figure out what CE uses to code its shaders, and how to use Opensauce to the advantage of adding new functions/math to the framebuffer I would be all over making/fixing the plasma systems.

Quote: --- Original message by: waffles
The blend modes were all pretty much difference in photoshop, which is an inverse function; white+white=black
black+black=white

The noise maps are blended together through the difference mode in photoshop, the result of those two are inversely blended with the offset mask (difference). Basically what an offset mask is/was for halo1 was a bounding area of sorts, it limited how far the blended plasma could go.

ex: the shield generator had an offset to keep the animation in a specific area, and not all over the geometry.

layer order in photoshop

noisemap a -difference
noisemap b -difference
offsetmask - normal

the frame bitmap is blended in the transparent shader to get color values, first the actual mask, then a solid color, using the multiply blend modes.

The main reason I never wanted to make a full tutorial on such a thing is the downsides of doing this method in the first place.

Tagspace due to bitmap size; 1 128x3840 bitmap (30 frame bitmap) compressed can reach up to 2 or more mb's. Using no compression would fail to compile due to being near 10-15 mb or so (if it did compile, that would be way too much for any single bitmap anyway).

Functions always run until they are told to turn off by w/e stimuli you set for it. This means the animation will always play even outside the bounds of bsp.

It would be much more worthwhile if the OS team manages to restore the old transparent generic shader somehow, or adds inverse blending to an existing shader

Quote: --- Original message by: waffles

First off I made a mistake on the bitmap size for 30 frames, I forgot that wasn't a power of two, if your using 30 frames, just round down/up to the nearest power of two size (4096 for 128x128).

But, you need to do a bit of math to get the value needed for the bitmap scaling. For the bitmap scaling, you just divide the amount of frames you have by the bitmap width.

ex:30/4096=0.0073

The result is your new U axis scale, its important to scale the frame bitmap because otherwise CE will attempt to fit the large bitmap to uv space, which won't get you what you want.

In the function set the step count to +1 the frame count, so for 30 frames you set step count to 31.

PS: its been a few years since i did this, so my math may not be correct, its either division by the bitmap width, or frame width.
Edited by waffles on Jul 22, 2013 at 08:02 PM