Archive for October 5th, 2011

Maintaining open horizons: The key to finding answers isn’t to always look in the box marked “answers”.

I’ve been working in CGI rendering methods for approximately two years, creating a handful of characters and each time the approach in modeling has been slightly different, but almost all of them started out similarly.  I’ve come to accept this as part of the territory when learning anything new and potentially complicated.  Textbook examples for modeling something like say, a face or a head will only usually take you far enough just to get your feet wet, unless you only find yourself creating the same kind of character each and every time.

My ‘book of choice’ at the time illustrated a method for creating a face by creating the mouth, nose, ears and eyes as separate entities and connecting them afterwards.  This works fine as long as you want to model a realistic human face.  But what if you wanted to create faces like these?

head comparison

Well, suddenly that method doesn’t seem so sufficient.  How do I attach the mouth structure to the nose structure when they’re integrated by design?   The book illustrated how to go about the business of crafting each one separately, not together.  Having nothing else to go on at the time, I muddled through by pushing and pulling the mesh into the ‘right’ shape and produced faces much like you see in the top image.  And while there’s nothing obviously wrong at a first glance, when it comes time to move the mouth, the deformations become… difficult to manage.  I don’t want to get bogged down in the details here but essentially it has to do with the flow of edges and polygons as they wrap around the model’s surface.  Another problem was that the character I was modeling was based on my comic strip character which in turn had been composed mainly of soft shapes that lacked a certain amount of defined structure in the face and head.
Time travel to this past July when I started thinking about creating a new character (bottom image).  In the period of time between the two, I ended up viewing some particularly helpful tutorials.  Hand-drawn art tutorials.  There was one in particular that addressed ways to stay on model by clearly envisioning the planes of the face and consequently adding the structure that my previous work had so sorely lacked.  It wasn’t until I started applying that knowledge in the hand-drawn design of the character that the thought occurred that perhaps one could model the head and face as a series of very low-res planes (basically one polygon per plane), merging them and then going into the requisite detail.  You can see the effect most clearly in the mesh between the bridge of the nose and the corner of the mouth.  As it turns out, using planes rather than parts provided a level of control over that tricky area of the face that develops in these kinds of muzzled characters, as it ends up dictating how well the face will wrinkle, fold and deform when it comes time to animate.

There are broader implications here which will become a recurring theme on this blog:  Finding solutions is not unlike good design.  The answers do not necessarily come from looking in the obvious places.

Advertisements

, , ,

1 Comment