Friday, October 03, 2003

The idea for my SmokingMoose Film came to me in the early 1990s. It wasn't until 1996 that I drew the storyboard.


smoking moose storyboard


To help me "realize" the characters I made plasticine models of the 4 characters.



smoking moose plasticine model



tiger plasticine model



canary plasticine model



raccoon plasticine model





Since then, I've started to make the film in many different versions of different 3D software. Somehow working for a living gets in the way of making short animated films ;o). However throughout the past 15 years I've made a number of traditional animated films and one digital paint animated film (using Eddie), this is my first attempt at a 3D computer animated film. 3D computer animation isn't foreign to me though, I have created a number of interactive 3d pieces for the web.

In 2000, I scanned the storyboard drawings and created a "leica reel" or animatic (a timed video of the storyboard). I recorded and gathered sound effects. BTW, the best place to buy sound effects online is at SoundDogs.

Obviously, many people have heard about "my smokingmoose film" as I've been talking about it for so many years. It seemed appropriate to make it the name of my portfolio website. It is such an interesting and unique name.

In July of this year (2003) my good friend Pierre Tousignant convinced me to restart the film using XSI this time. Pierre also suggested that I "blog" my experience in producing a short animated film using XSI. In this first post I'll describe what I've done so far and what my current approach is. In the upcoming posts I'll describe what I'm working on, problems I'm facing and how I'm solving them. By the way, this is an independent film, and I am not independently wealthy, so if you'd like to support it you can buy smokingmoose paraphenalia at the smokingmoose online store!!!

I spent most of August re-aquainting myself with XSI. I hadn't used it since December 1999 when it was still in beta (before the release of version 1.0). To get back into it I read the fundamentals chapter of the user guide and this helped a lot.

My approach to making the film is to work through the film on a character by character basis. I don't want to get tied down in making the background look pretty now .. I can do that later. For this film, what is most important are the characters, including their personalities. I think it is important for the animation process to follow the modeling process quickly and to stay focussed on each character individually rather than trying to complete each scene in its entirety before moving on to the next scene. This is pure intuition on my part, we'll see in a few months if this method works.

For modeling, I looked at the tutorials provided with XSI but realized they weren't going to help me model the moose, which I wanted to model first. I found a modeling tutorial on modeling a dog using subdivision surfaces on XSI base that really helped me. Once I had done this tutorial, I was able to model the moose and the first results weren't too bad.

For the character setup, I followed Tutorials 6 and 7 in the Softimage documentation called "Beck's Skeleton" and "Beck's Envelope". These taught me quite quickly how to create the IK, apply the envelope and edit the weighting of the points on the Moose model.

September arrived and I decided to put the Moose aside and try to create particles, as the film has many particle effects in it... the first of which is of course the moose smoking a cigarette. The amount of information in the particle property pages was overwhelming at first and although the tutorial for particles is good, it too was just a bit daunting with all the properties and parameters. However, XSI base again provided a tutorial on creating smoke, that was more on-target for me ;o)

When I went back to the moose model to put the cigarette in his mouth, I discovered that the model was not behaving correctly at the chin (underneath the jaw of the moose). There was an indentation between the jaw and the head. A few tests revealed it was somehow related to the number of points on the jaw versus the number of points on the head on the polygon model used to create the subdivision surface. A few more tests revealed I had to remodel from scratch. This time I started with a polygonal sphere rather than a cube. I increased the subdivisions on the polygonal sphere so that I would not have to split polygons and remodelled the moose.

Rather than redraw the IK, I created a Model from the IK Tree of the first moose scene and merged it into the new moose scene. I've applied the envelope and am currently re-weighting the points.

No comments: