View Full Version : phonemes

05 May 2003, 06:09 AM
out of curiosity, what techniques do u guys use to model all the different phonemes and expressions of the face? do u set up a bunch of clusters first, or joints, or just move the vertices around? my character's face has so much detail so im looking for the easiest and quickest way to do this..any suggestions will be great!

05 May 2003, 08:23 AM
The method I used was for low-poly game chars in cut-scenes - but it should be similar for more complex meshes.

First I created groups of verts and saved them in quicksets so I could easily select them with a nice script called Selection Wizard (from highend) or in the Outliner.

These groups were of points in the corners of the mouth, the middle upper lip, top right eyelid etc.

From these I created clusters on my "master" head.

I duplicated this with upstream - so the clusters work on the dupe.

A very annoying thing Maya does when you create a cluster is to create an icon (a little C) that has nothing to do with where the cluster is. You are supposed to select them via the icon - but the icons don't move with the verts they are supposed to represent, so it can get really messy. Also I like to reposition the pivot for the clusters - the little c doesn't update with this. I group the bottom lip clusters under the jaw so that they move with it (no this doesn't contradict what I just said) but you have to have "relative" checked and I think you must create them in parallel (not 100% sure about that)

You name the duped mesh "smile" - which is your blendshape source

Push and pull the clusters around to get your shape right

The advantage to clusters is that you can zero them out and get back to your original without too much trouble.

I usually do some individual vert tweaks as well

Do a straight duplicate (no clusters) of your master head and use it as a target for your blendshapes.

bind the mesh

switch the order of deformations so that the skincluster is before the blends - first blend should be a parallel.

I created blend groups for eyes, mouth, brow and a separate one for phonemes.

I found this process one of the steepest learning curves in Maya - mostly due to deformation order and the whole relative non-moving cluster thing. It's not really that difficult it just seems to be the area where they have the least amount of info on "how to".

i'm planning to use this on my subd models - can anyone shed some light on this process - does it vary much?


05 May 2003, 11:35 PM
hey thanx magilla that helps a lot!:beer:

CGTalk Moderation
01 January 2006, 05:03 AM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.