Here is the final version of my thesis… a bit delayed I know, but here they are.

Mac version (2.2 Gamma)

PC version (1.8 Gamma)

Just a write-up of some of the things I have learned while doing this project:

Displacement:
Even after mastering displacement, its still pretty tricky to get it to do what you actually want it to do, especially when you throw in another variable like generating your displacement from a procedural network. Every little attribute can blow up your displacement if its pushed too far. And while Mental Rays fine approximation tools are capable of producing great looking displacement, it takes hours and hours of fiddling with the settings in the approximation editor to get the right mix of detail and reasonable render time. Still it was a great way to create these large terrains.

Fire:

My main way of controlling fire was by modeling objects that I could emit the fire from, and then controlling the progression of the fire through the texture fill rate via a ramp. I think this was a very smart way of handling fire instead of struggling with all sorts of different fields and goals to get my fire to go to the right place.

Managing the glow of the fire was also a little bit problematic with moving cameras. I had to go through and keyframe good glow settings throughout the shot. The glowing would also sometimes have a mind of its own, for instance in a few situations the fire glowed much brighter when it was in front of a certain piece of geometry, but ONLY reacted that way to that one piece of geometry. Very Strange.

Unfortunately I had to render my fire in Maya due to a certain limitation with Mental Ray. The reason this was unfortunate, is that my fire was often moving in and around displaced objects, which Maya rendered much differently than Mental Ray, which meant my layers wouldn’t line up correctly. I had to model various stand in objects in the shape of my displacement as a work-around.

Lighting:

I put in a great deal of work on my lighting, and I was actually able to render most scenes without the use of shadows. (They don’t add anything to dark rock, and lava certainly doesn’t have shadows on it.) In areas where I needed a shadow to block the light entering a cave, I instead made use of negative lights. Some of my favorite parts of the lighting process were using expressions on my RGB and intensity values to better simulate light from fire, as well as mixing the intense light from the lava and that of a dreary sky.

Here’s a preview of my fire effects blasting up the mountain inside. Its gonna be sweeeeeet.

Here is the video I presented at the industry review.

I received a lot of great feedback, the highlights of which include:

1) Adding more midtones into my rock color so that the detail isn’t completely lost.

I have already improved my rock texture exponentially thanks to this comment. My scenes look much better.

2) If possible, change the first person camera to react more like a person’s eyes would, leading the viewer and looking at different things.

I may not have time to implement this to the level that they were talking, but I am going to make some adjustments to move towards this idea.

Overall they seemed to enjoy my piece… which was a pleasant surprise to me as I was sort of expecting to get ripped apart.

A longer journal soon re: displacement, fire, and lighting.

I’ll be posting longer articles with actual content this week, but I wanted to post about a fun little trick I found. (Everyone else probably already uses it and I’m slow to the party.)

Anyway, I made a more complicated shader for my rock tonight to give it more mediums, instead of being 100% dark.

The annoying part of this would have been the fact that all of my rock scenery uses a different amount of UV repeats in the main fractal driving the color to get the look right. Well now I have three other nodes that multiply together with the original fractal to give me my final color result. The thing is I didn’t want to have to do the math every time I changed the UV repeats on the one fractal and figure out what the ratios of the other ones should be (they aren’t nice and simple… all using separate amounts of UV repeating.)

Then I realized… Maya does math. Doye.

So I grabbed a multiply divide node in the hypershade and connected the U and V Repeat values from my main fractal to the X and Y A inputs of the multiply node. Then I just stick in whatever number it needs to be multiplied by to keep the correct ratio in the X and Y B inputs of the multiply node.

Hook the result to their respective inputs in the 2D texture placement node for the other texture node, and walah! All I have to do is change the UV repeats on the main fractal and the others stay in sync.

Its so simple I feel dumb being excited about it. I really just love piping nodes through other nodes in Maya and Shake. Quite the nerd am I.

Here’s an updated version of one of the shot sequences.

You are seeing updated displacement, lighting, and fire effects. I did some cool tricks with the lighting intensity and color to make it look more like a fire flickering (thanks to Aditi for sharing the expression she was using.) I adapted it to vary both the intensity of the lights, and the individual RGB channels. I’ll post more about that later.

The 3 fire balls along the side are actually torches, but the torch base is missing in action for now. I’m actually thinking about switching it over to having the track light on fire, similar to my other shots, except this time the fire will kind of lead the car out of the cave. Doing that would have a more dramatic effect of gradually illuminating the whole cave. It would mean less time to look at my beautiful dancing light though… and that makes me sad.

I’ll also be posting a more detailed note about my latest dances with displacement, what I’ve learned about controlling fire, and a few other techniques I’ve been using on my piece.

I’m hoping these posts count as the “journals” we are supposed to be doing. I’m not sure what else I could write that would make this more journal-y.

Good news everybody!

I shortened my required render time to somewheres in the area of one week!

I was able to do this by shortening my piece to 45-50 seconds, as well as fine tuning my settings. This will give me more time to setup the final scene, one other, and increase the quality of the lighting on all other scenes.

I did a quick HD still of one scene. It will look a little bright on macs as I am lighting based on the PC and TV standard gamma now. I’ll make a separate version that will look better on macs.

Updated Animatic here with a new scene of rendered shots.

I’m trying to nail down a bug with h.264 codec. Sorenson 3 reproduces the colors of my animatic fairly well, but h.264 looks faded out on some computers but not others. You can see my research so far by going to my Quicktime h.264 bug page (also found in the menu to the right.)

If you have time, give a download to both and let me know how similar they look (along with what type of computer and video card you have.)

Sorenson 3 version

H.264 version

Here’s a quick update with an animatic, and some new stills from up and coming renders. The renders of this new shot should be done by tomorrow (hopefully.)

Animatic

Still 1

Still 2

Still 3

Its been awhile since I’ve updated what I was doing here, but I’ve made a huge amount of progress. Latest Animatic with renderings below.

Displacement:

I have pretty much mastered displacement inside and out. I had to mess around with it for a long time before I could control it to the point of getting exactly what I wanted and keeping render time down. In fact in perhaps my biggest scene, I fully displaced 2 giant mountains and a modeled mountain face, and the rendering with lights is still under 2 minutes. Motion blur will make it take a hit, but its still well below my render time limit per frame.

Fire:

I met with Vick for any advice on controlling fire with particles (previously I was planning on doing most of this with fluids.) In about 2 minutes he showed me how to completely tame fire with particles, and I will be using it for most of my scenes. A fireball here and there could still possibly be done with fluids, but we will see. The best part is that particles barely add any extra render time.

Here is my latest animatic with renderings for a bunch of the shots.

A few changes since last time:

Some shots are cut a bit quicker.

Added 1 shot that needs to be adjusted (both the camera move and the length).

Cut off the last 2 shots… I like the idea of the piece ending with the rollercoaster car rocketing into the air.