Okay guys

heres an overview of the modeling process from the beginning. Now those of you who are familiar with my work will already know some of it, but theres a couple of differences with these models being for aniamtion production.

Okay well the first thing is that you get the original artwork either from the concept designer or the modelling production supervisor, either will decide on who models what depending on the complexity of it, me having experiene of a few years got the hard ones, but needless to say I was still doing them in 3days, 4 maximum.

As ever my stuff always starts from a cube then its cut n shaped as you would. Also I'd like to point out that the concept artist has made front and profile images for you to load into the maya viewport, I was also provided with folders of photographic references of each person/character to be modelled. These were vital to capture the persons character or areas of exageration and other details that may not be clear on the original artwork

Clothing was done again either from a polycube or plane, just depends on how I felt doing specific things.

Areas of deformation: this always depends on the character. Most of the time shoulders, elbows and knees had between 3-5 loops, this is adequate enough especially when some poor bugger is rigging all day! Its got to be clean in those areas allso the stomach areas. As with everything try to keep in mind about the processes after modelling. Basically best results come from good edge loops

Textures: All pretty hi res, between 2048 and 4096, all file types were either targa or tif, mostly targas as it was for Renderman. JPGs just don cut it at the end of the day. Rumour has it that everytime a jpg is opened it loses quality because of compression. Ive still to get this theory sussed.

Because of the style for the show most of the texturing was pretty simple.


Texture Tip: Dont use the pelt style of mapping because it doesnt give you an exact/true replication of your model. Just because its fast doesnt make it correct. If you are going to use high res pics for a facial map you want the eyes lips etc to be in the right place, no lips in the area of the forehead if you know what i mean. Alot of the time I found models that I corrected by others either had badly layed out uvs (pelted or auto mapped!) or none at all, if the model was in need of huge tweaks, ie: re defining edge flow, removal or additon of geo then it was usually a case of redoing uvs from scratch as it was actually quicker than arsing with old ones. Use the tools available by all means but to get great results you have to do it the more traditonal way, of pulling uvs about. And use the whole uv area, thats what its there for!

Approval:Before sending the model off to rigging the model has to have all the holes plugged up, this means noses throats, sleeves, trousers or neck areas usually need some polys to fill the holes. This is basically for the camera, (most models are single sided once theyre rigged, so you dont want to be looking down a mouth cavity and see the background or props etc) once finished, its usually checked by the character concept artist who in turn will then either say yes or no to whether it matches the design or not. If yes then you do a turntable render or front, profile and perspective renders that will go to the Director and Producers. these will be approved accordingly then they are off to rigging.

The rig was a generic one to fit all, you can probably get the Digital tutors dvds to attain something similar if not the same. A floating gui was created for the facial setup along with some blendshapes, again all available from DT it wasnt rocket science or Muscle plugins etc, these are caricatures so they can behave slightly different.

Once rigged, its off for animation. This comes at a variety of stages. Blocking, then animated with the base model or mocap was added to specific characters, lots of playblast too! then its final animation with Lip sync, then rendering.

Rendering was just awsome to watch happen. You see your character go from open GL then come to life. I cant really elaborate on it too much but all I will say is renderman is magic, the Tds had written specific stuff for it to work for the show, its basically like Maya in those respects. Also alot of custom shaders came into play. No Standard shaders were used as such so...

Well there you go, hope it helps...any other questions I'll gladly answer if I can

cheers
Jay