Wednesday
Jul292015

Noodling in an Ever-Shifting Procedural Landscape

Here's a couple of recent instagrams I posted. These are details of much larger renders which I'm not quite sready to reveal (until my show in October, at the Red Head Gallery in Toronto).

I plan to print some of them roughly 8 feet x 4 feet, at a crisp 300 dots per inch. This requires coaxing out renders 28,800 pixels wide. I'm constantly bumping into hardware and software limitations. Thankfully I can farm out these huge renders, as my machine alone would take well over a week to get through one.

 

The process of making these involves a great variety of stages and components, which range from drawing and virtual sculpture to something more akin to programming. They consist of a set of elements numbering roughly 300 at last count, and which include things like rocks, trees, buildings, abstract forms, etc.. These are then randomly scattered in vast numbers across a given terrain (itself possibly randomly generated).

Each set of randomized elements must follow a set of rules of my choosing. For example, I might say "rocks, you are only allowed to exist at altitudes between 15 and 23.7 meters", or "trees you may only grow on slopes of between 30 and 45 degrees", and so on. The system remains live, and I can change a single parameter and the random arrangement will shuffle itslef into a new configuration based on the tweaked rules. I can see previews of the state of the system at any time. For every proper render I do, I flit past several variations that will never be seen again. After enough noodling I'll get to something I like, I save the settings and do a medium-sized render (as in maybe 9000 pixels wide) which my computer can render overnight. Based on those renders I will opt to refine some of them further into much larger images. Along the way, the scenes acquire ever more computer-straining bulk, which makes the going quite slow.

On its own, any given procedural system based on a simple set of rules will tend to look obviously artificial. When combined with several others you start to get a rough approximation of something natural. If you look closely you will notice that the whole image consists of the same objects repeated in different configurations.

Working with randomness in this way leads to compositions that I would not contrive when drawing. In that way the process is more like photographing a landscape, an ever shifting landscape that exists as a collaboration between myself and randomly emergent patterns. A landscape where I have full control of lighting, atmosphere, and other conditions.

The software I am primarily using to do all this is The Foundry's Modo.

 

 

Friday
May012015

More recent Instagrams

 

 

Monday
Mar232015

Instagram Collection

I posted a collection of my recent instagrams on the Modo forums. These images are actually just cropped down details of larger renderings, which form the basis of some new work I'm doing.

Monday
Oct272014

Future Toronto - Winter Version in the works

I've been picking away at a winter version of last year's Future Toronto illustration. I'm hoping to have prints available by December. The image above is a WIP, and so far I like this version the best. I'm planning to add a lot more life to it, unlike previous versions which didn't have any people visible (except for the tiny guy in the boat fishing). One of the advantages of a flooded Toronto in winter is there's lots of places to skate...

This image and others will be in a calendar called "Fallen Toronto", which will be produced as an award to funders of a successful kickstarter campaign for a sci-fi web series called Haphead.

Large prints of this version will also be available soon at Spacing City Store in Toronto. 

Saturday
Jul122014

Marsh Villages

This is a spin-off elaboration of the top-left corner of my "Future Toronto?" illustration.

A detail from same: