Must. Resist. Urge. To turn. Duckling. Into. Dinosaur.

Demo Reel

I haven't put anything up for a while because nothing looks all that much different yet, but here's my demo reel, for anyone who needs something to look at:

http://vimeo.com/10783307

Breakdown here.

Also, my shamefully outdated website.

Lost my USB where all of this work was stored (exhibit A for why you should back up insanely important data), so unfortunately I cannot fiddle with anything on there now, but I'm hoping that the work I do this summer will be quality enough to toss in and hopefully distract from the many, many existing problems.

Also P.S. Joe just expressed his approval for my work on the side--YES.  No more frantically minimizing my raytracer VS project whenever I hear footsteps behind me.  I was beginning to feel like a man hiding his mistress.

Refraction

I think I got it wrong, but here it is anyway.

Note: You probably noticed something missing--caustics. This will be implemented when I get to the photon mapping (if I can manage to actually finish this raytracer).

Phong shader

Badly working Phong implemented. LOTS of kinks to work out on that one.

Notice the angry highlight. Eek.

Tomato collapse with mold

Tomato collapse with texture changes

There are still about 30 frames left to render, but I thought I would put up what I have so far. No fur-mold yet; look for that later today.


Additional note: I will probably be putting the other stuff on hold because this project (which I happen to be getting paid for) takes first priority for the next couple of months, and at least for another week staying at work for ridiculously long hours is not possible.

Fur test

Played around with the surprisingly realistic Maya Fur. Still don't know how to do anything except light it; this is using the Duckling fur preset slightly modified. Maya Software render, because mentalray hates fur and hates me.

Basic Ray tracing

I am deeply unsettled by the fact that my fudgings have been piling on--I really need to learn not to cheat so much; it's awful. One day I will go through this code and clean everything up, but for now I at least have reflections and shadows implemented.



The shadows should really be black, but it looked ugly so I made them a more delicate gray. Yet another instance of cheating.

Ray casting corrected

I'm still pretty sure I'm fudging a lot of things I'm not supposed to fudge, but at least it looks the way it's supposed to now.


And because I'm not too fond of that look, the wrong version and the right version comped together:


Theoretically that would be calculating the color values based on both angle from light source and distance from light source, but something tells me that's not the whole story and that this is the wrong way to do it. Oh well.

Ray casting

VERY basic ray casting implemented. Only 3 types of primitives, major cheating and hard-coding going on, but hey--it works. (Somewhat).



For comparison, here is a Maya software render of a similar scene:



Now that I look at those two, I realize the "ray-casting" algorithm I pulled out of my ass is clearly wrong. Damn. Looks like I will actually have to reference my old CIS460 notes. Groooaan.

UPDATE: Figured out what my problem was. I was calculating color values based on distance from the light, not the angle between the point on the surface and the light. Laziness: 1, Actually knowing what I'm doing: 0

Tomato video 1

No collapsing yet.

Drool

This is my inspiration for the kind of images my raytracer should be able to produce by the end:



Done by Marc ten Bosch of Brown University, I believe.

Main resource: Realistic Image Synthesis Using Photon Mapping by Henrik Wann Jensen.  Really great book and an easy read (with a full C++ implementation in the back!  I will try my best not to cheat.)

Planned work

• Tomato


• Fixed raytracer with photon mapping


• Fast caustics (might give up on this because I'll hopefully have caustics implemented with photon mapping)


• Some sort of vaguely-defined aging surfaces plugin


• Carvaggio-like still life (with cheese, grapes, wine, tablecloth--subsurface scattering, layered textures, Maya caustics, cloth, and LIGHTING)


• A duckling (Maya fur)


• Grace's drawing

Tomato part 2

I considered waiting until this actually looked passable before putting it up, but I decided that there is no point in having a progress blog if I don't bother sharing the sadder portions of my progress.







Still need to figure how to make the mold look like mold.  Maya Fur is the closest thing I've come up with so far (the same thing we used for the Senior Project render) but it looks like... fur.  And crap.

Notes:  I'm using mentalray this time because I really don't know how to use the Renderman shaders effectively and I'm short on time.  If the opportunity presents itself, I might make more of an effort to learn Renderman, but at this point I don't really see a reason to.  mentalray is fine for now.

Julie Bell

I love, love, love Julie Bell's attention to detail in terms of musculature.  Most of her artwork is of the sci-fi/fantasy/bordering-on-porn variety, and I love it.  THAT is how you should draw naked people.  Forms that are so accurate to life and beautifully rendered that it doesn't matter that she is the science fiction equivalent of whoever paints Fabio.

I particularly like this piece because I like comic book art.  It's so distinctly American, and there is something about the way they portray action that I love.  In a lot of other styles (I'm thinking of anime and manga artists right now) you'll get a real sense of movement and speed, but you don't see that here.  They're just suspended in air, muscles mid-ripple.  It's like time just stopped--you can almost hear the silence and the stillness, and you wouldn't expect that from an action shot.  It's fantastic.

Oh my god, it’s back!

Yes, the tomato is back.  It looks exactly the same, except that now the mold is growing in a somewhat biologically-based fashion and it doesn't collapse yet.  Here is a fresh one:

And here is an orange with some sort of dent in it (for the hell of it):



Less appetizing versions to come.

NOTE: Tomato/stem model credit goes to Michelle Chen!  And spoiler: the orange is a sphere.  There is a reason I am not a modeler.

Hello there

I wanted to come up with a clever name but then I thought, "This isn't my clever blog anyway."  This is my project blog.  It's mostly for me to keep track of the many projects I am convinced I will eventually start during the remainder of the summer, because my brain does a bad job of keeping track of anything.  It is public and available for you to peruse because A) it would be nice to have conclusive evidence that most of what I do does not, in fact, involve explosions or spaceships or dinosaurs and B) it forces me to be accountable for the crap that has my name plastered on it.  Perfectionism has never posed much of a problem for me, but public embarrassment has, so I'm hoping that this blog will discourage me from producing work I wouldn't be proud to show my friends.

Also, I kind of just want an audience.