Internet Ray Tracing Competition Entries

Sep-Oct 2000 - The Laboratory

In The Flask


The image

A chemistry laboratory, with a rather unusual experiment in recursion taking place.

How this image was created

Conceptually this just grew out of random bits of lab equipment. Before I had an idea for a scene, I started modelling flasks and test tubes. Then I had the idea of putting the camera inside a flask - but a test render showed you couldn't tell the camera was inside a flask! So I hit on the idea of putting the lab inside the flask, and went from there.

The room and windows are vaguely reminiscent of my old chem lab classes at uni. All of the objects are constructed with CSG of primitive shapes, though the geometry got very hairy at times. Lots of trig calculations. I've used layered textures to make things appropriately grimy, but the glassware is all pretty clean, as it should be. The textures need more work, but time was too short.

The trees in the background were made using a recursive fractal macro I wrote originally for the Gardens round (but never used!). Each of these trees contains 32,418 objects. To save render time in the scene, I rendered each of 5 different trees individually and used them several times as image-maps in the final scene. It's no Giles Tran macro, but I'm proud that I wrote it myself, and I hope to improve it over time.

The lighting comes from a series of area lights on the ceiling. That, combined with all the glassware, the caustics, the anti-aliasing, pushed the render time way up. This is nowhere near as complete as I'd like it, but I really ran out of time.

Development History

No idea yet, beyond general "labby" stuff. But I'll need props. Following Peter Murray's example, here are some test tubes with nice coloured chemicals in them.

More labby stuff. This time a conical flask. This is CSG using cylinders, cones and toruses, and took some heavy geometry to figure out. The glass has a refractive index of 1.4. An early idea was a view from inside the flask, but I rendered that and it didn't look much different to the test tubes above! You couldn't even tell the camera was inside a glass container - but it took 6 minutes to render at this size! Because the glass is a uniform thickness, there is no distortion looking through the single layer, like a curved car windscreen.

I had an idea to put the lab inside the conical flask. And what is in a lab? A table, with a conical flask on it... The big flask is of course on another table... I still need to wrap a bigger flask around that, and also put a tiny table inside the small flask. I don't think any more beyond that will be visible, so that's about as far as I need to go. Then I just need to add other objects to the scene - in three different scales...

Created a beaker and added it to the scene. It's not positioned quite right here, because it's intersecting with the flask, just to the left of the bottom corner of the tabletop.

Haven't done any work on this for a while - but the deadline is fast approaching! The beaker has been moved so it doesn't intersect the flask any more. More importantly, the table is now inside a room, with windows and all. Well, holes where the windows should be. There are also clouds in the sky. And there's another table in the background, with similar equipment on it. Oh, and the wood grain of the table is new and improved. Most importantly, the big table-top has been moved a tiny fraction downwards, so it is no longer coincident with the big flask's bottom surface. This fixes up the anomalous non-transparency of the flask bottom, as seen in the images so far.

This is a zoom in on the windows, which have been fitted with actual bits to hold the glass in. The glass isn't there in this render, nor is the big flask, and all the other glass is white. But you can see the lovely trees outside the windows! I finally get to use the tree macro I wrote for the Gardens round and didn't use. They look better at higher resolution.


Home | DM's Ray Tracing Page | Internet Ray Tracing Competition Entries
Last updated: Saturday, 11 February, 2006; 23:02:21 PST.
Copyright © 1990-2022, David Morgan-Mar. dmm@dangermouse.net
Hosted by: DreamHost