I'd like to make a figure I've been working on Talk Designer compatible. Designing the phonemes morphs is easy enough but I'm curious about required naming conventions for the various morph targets. For instance looking through some visememap xmls it appears that a blink morph can be arbitrarily named and the named reference inserted during visememap creation. It appears that other emotion morphs must follow a strict naming convention.

Finally, any suggestions on the best phonemes morphs to look at as inspirational reference when creating my own for my figure. I'm assuming that Poser included content official figures of the legacy variety would be the best points of reference.

Any tips or observations would be appreciated.

The visemes can be mapped to any morph in the "head" of the figure. If you're doing a cartoon for example you might only have 5 or 5 visemes morphs.

The emotion poses can be mapped to any face (FC2/FCZ) in any linked runtime. For sanity's sake those are usually in ...Runtime\LipSync folder but that's not a requirement. The FC2 files can be mapped to any face pose.

Probably the strangest thing about the viseme files is that the use of internal or display names of the morphs is ... kinda random?

Like the blink morph uses the internal name:

<Blink morph="eyelidBlink" openValue="0.0" closeValue="100.0" closingTime="0.167" closedTime="0.034" openingTime="0.134" />

But the eye side to side uses the display name of the dial ...

<EyeSideParm  parmName="Side-Side" maxValue="5.0" minValue="-5.0" forerun="0.1" minimumInterval="0.8" toSideTime="0.1" staySideTime="0.7" backToNeutralTime="0.1"/>

You will be doing a bit of trial and error figuring that out. If some parameter doesn't seem to respond to talk designer try using the other naming for the dial.

To answer one of you questions. You only need 7 different phoneme's to do Lip sync, but you should probably do more than that.

This is from here, https://www.dsource.in/course/lip-sync-animation/process and as you can see many others are basically variations of strength of the morph. The key is to mix them convincingly. As Charles mentioned you can use emotion poses which is key for making things like inflections look believable. The key to making it look convincing from scene to scene is having more than one set to do this based on differences in emotion that is unique to the character.