The steampunk chess robot project asked us to design,
build and animate a Chess robot and its environment in the steampunk visual
style, and whilst the brief asked for a limited “set”, I made the decision to
use a slightly more versatile street setting, as I felt that an alleyway could
feel more industrial and contemporary to the brief.
I took in a number of inspirations, and thankfully
steampunk was a theme I was vaguely familiar with already thanks to films such
as “9” and “Howl’s moving castle”. Other inspirations included an unusual piece
“future forms of life”- David Lance, which featured a strange shuffling
biomechanical beast, and somewhat influenced my animation angle regarding my
steampunk spider.
During the project I came across a number of challenges
and issues, each of which was overcome through problem solving and internet
searches. For example, the scale of the set required me to create separate
instances in which I could efficiently render separate scenes without having to
worry about extra lights and artefacts slowing down the rendering process.
Another technique that I used to maximise productivity
was to only render whilst I was asleep, I’d check the first few frames, then
when I woke I’d have the full collection of stills from which to make an
animation.
However this worked against me at one point as I’d accidentally
selected the “perspective” render view, and wasted 4 hours rendering the entire
animation from one angle. Thankfully, I learned from this mistake.
Another approach I took to make the scene more cinematic,
but avoid overcomplicating the scene for myself, was to only use a single
camera, and to use key framing in such a way that to the audience it would seem
that the shot had cut to another, separate one.
Similarly to this, I struggled with a certain linking
constraint- where a chess piece refused to be unlinked from a hand constraint,
so I made the decision to splice it out of the shot at a certain point, and
stitch in a copy in its place. This worked well.
One issue that I would like to overcome is the way that I
textured my objects, I’d like to be able to map textures onto surface in a way
that looks entirely convincing, rather than just adding bump maps and using a
simple UVM map modifier. I hope to achieve this by practising with the Unwrap
UVM modifier.
Secondly, I believe that despite the cinematographic
aspects of my final animation, my use of pacing and timing could use some
refinement, I’ll look to address this through practise and exploring other
examples to see how they tackle this.
One aspect that I would like to improve on is the way
that a character has ‘character’, after seeing other examples I realise that
there are certainly ways of effectively adding depth and strength to a character
profile, in similar ways to Pixar and Dreamworks.
Overall I am pleased with the outcome, though next time I’d
like to aim for completing the project well before deadline in order to add
more post-production time to refine and polish the work into another level.
