I wanted to concept and model a character from scratch, ending up with both a high resolution model and a LoPoly engine-friendly mesh. I learned more working on this piece than with any other (more on that below). There are clearly some Lucasfilm shenanigans afoot, and on the design side, its becoming obvious that I get little pleasure from the over designed, hyper-fractalised approach.

I started off with the basic shape of the head gear, and it went through some iterations as more organic and cicada-like before I reminded myself to stick to a design which best for an overall concept rather than pursuing an infinity of designs for each asset. This is a very difficult balancing act – to explore options and variants while remaining true to some vague spirit of an initial concept. Of course this is becomes redundant when working to someone else’s design brief, but the line is always there at smaller scales.

So this is subdivisional modelling. This is the technique I’ve been using up to now. Never having (digitally) modelled traditionally, (either polymodelling or box modelling), there are certain factors I’ve never had to consider before. I’ve always worked from a LoPoly mesh upwards, always with the intention that the lowest mesh is somehow untouchable, that it is definitely bound the high resolution work. As a result of this, I’ve been considering the  geometry needs of a perceived “game engine asset” as a sort of baggage as I worked, failing to grasp that , although the base mesh which I start off with can often be used as the geometry of a game engine asset, this is never the case with more complex meshes (such as this one). As usual, this is going to sound incredibly obvious to some, but I have to say that it’s massively liberating for me. I needn’t keep my base meshes half as simple as I have been doing, for the simple reason that the process of sculpting is and should be a completely separate pursuit to the process of making a sleek, memory-efficient asset. Having said that, working within confines is always rewarding by its nature, and I wouldn’t have learned half as much if I hadn’t found out the difficult way. I think.

Finally, these are the six layouts I used for the finished mesh, and some of the maps employed. I had to make two variants for each tile, as I was working on a LoPoly and a High resolution mesh at the same time. Let’s just say that its an interesting experience learning about object space, tangent space and smoothing groups while juggling 6 mesh layouts (with the software constantly flipping them for its own undisclosed reasons) with no grasp on when a normal map or geometry is best employed and, if changes are required, “how far back along the pipeline can I afford to make those tweaks while destroying as small  an amount of work as possible?” – all while using new softwarez on a rig with Less-Than-Sufficient steam power which sometimes decides that its had enough and packs it in just as you’re stupidly passing the two-hour  mark since your last save.

“I’m Learnding”

-R. Wiggum