How many times have you seen a hole in one? I mean, can you think of a time when you got it right the first time? Did final work in the first pass? I've never seen it.

Isn't it amazing that we all still approach work this way? Every demo you've ever seen shows it working the first time (unless it crashes) - but work you get paid for isn't a demo.

My way around this is using a 'template' or 'jig' approach to the work I do in 3D. I can hear you saying:

'The shortest distance between two points is a strait line - work from where we are to where we want to be.'

This is not always so.

Tiger Woods is not just a good golfer - he's a great golfer. He doesn't go for a hole in one - ever. His first shot is not strait at the hole - but to a midpoint he knows he can get to. Why?

For the same reason the experienced artist uses 'templates' or 'intermediate steps' - the simple truth that the shortest distance is ALWAYS a curve in 3D space. Rainbows are not strait lines - the pot of gold is never where you thought.

First, size up where you THINK you're going. Next, move strait to a midpoint (that you can clearly see from where you're starting). Now that you're close enough to completion that you can see the end clearly - you move strait from the midpoint to the end. Looking back, you'll see that your midpoint was far off the course you would have taken to go strait to the end - but if you only do two steps EVERY time to get finished results, you'll be ahead of everybody who's struggling to recover from their first attempt. And their second. And third... Tiger Woods would LOVE a par 2 on every hole.

I know this sounds simple - and it is, but you'd be amazed how many people (including me) frequently have gone off into oblivion with the belief that they were on the FAST path.

Okay, enough theory - here's an example:

• You want to create a character geometry quickly
• The model needs to have strong proportions
• There are some fun ideas you have for the model that are NOT in the design you'd like to play with
• You need to have results that you can recreate

Where you're going can only be measured from where you are. The first step is to make sure your scale works with the rest of the environment.

I use the most primitive method - I build a 3D yardstick geometry as a visual reference. This lets me share with different machines, setups and applications as a initial test to see if there are any 3D space differences.

For most characters, I'll scan in a photo, reference, or sketched image as a background - but this isn't required.

Now that I have size and some idea of look - I create a NURB sphere and place it to represent the biggest part of the character (torso). This extends the sketch idea into 3D space. Remember initially drawing circles to get the mass / proportion of your character right - then to create the final drawing on top of that? This is the same idea. Remember to start with the LEAST number of spheres to get the shape or volume you need to 'sell' the character. Continue on until all of the major parts are represented in rough.

Now you have a shape that you can look at in 3D perspective and see what it communicates to you. Making changes to this rough is easy, fast and you're attachment to what it looks like doesn't keep you from working the overall shape until it's solid.

Finally, you can use the NURB geometry to build a Sub-Division surface or poly geometry around it that helps keep you 'in bounds' of the whole shape while working on some specific detail. When you turn off the NURB spheres, bind the skin and animate the character - things continue to work nicely and you have a 'history' of it's shape that you can go back to easily at any time.

Two steps to a finished character geometry that is on par with what you want.