visemes help morph targets

[needse]

hi guys

im setting up some morph targets but im not sure excaly what targets i need to be able to make my character talk in English if anyone knows or has any info on this it would be very handy also any info on lip sync would be a plus

heres a few ive set up they are just made up ones

[poopipe]

I settled on this set after trying countless different setups over the years.

At 7 targets it's about the minimum number you can use and still do english properly. Used in conjunction with various mouth targets I've not had any complaints with regard to not being able to make a word with the required emotion (it's always a good idea to overdo the targets a bit - give the animator more to play with than they need)

You can see all the names in the layer palette. I tend to use the generic sounding ones on most face rigs and add character specific ones depending on the situation - I think I said before that it usually comes out at around 36 targets (for realtime stuff) - For more detailed setups I've gone well past the 100 mark (which raises it's own problems as max tops out at 100 targets)

In this particular case I had a jaw bone to play with as well as the morphs

I tend to rig them using spline joysticks and wire parameters - plenty of tutorials floating around the web on doing that so I won't explain it here.


with regard to your setup.

I'd advise thinking of the face as split up into areas and making targets that only affect that area

eg. left brow, right brow, left lids, right lids, left mouth, right mouth and so on.

your winks are nice but will be inflexible (eg. cannot be used together as a blink) meaning you have to make the blink which is no good if you want one eye half closed and the other fully closed etc. .
If you build them in a modular fashion, you don't have to make as many targets and you get more flexibility.


(pro tip - to get modular targets from your existing ones use skinwrap with a vertex selection )

hope that helps

[needse]

kool thats what i needed thanks

yeah im having trouble blending some of them together plus ive been poses the faces using the bone rig then cloning and collapsing and when i say close his eye it effect''s alot of muscle all over the face and im getting some undesired results so i think i will redo them all again using your list (thanks) and pose them manually

just loading a vid on utube

[poopipe]

I use zbrush cos the masking really helps avoid all that stuff.

you can be creative with soft selections, morphers and skinwrap to knock the unwanted vertex movement out though so you don't need to start again

[needse]

http://youtu.be/fa2zqcsIC_c

just a grab of the morphs no blending as you can see his head moves all over the place but ive just took all weights off of the back of the head and antennas so i might be alright to pose using the bones again
if not i try the Mudbox approach

[needse]

have you checked out voice-o-matic anygood ? looks cheap

anything to speed things up coz im doing all this on me own its gonna take me ages

[poopipe]

I've tried a few automatic ones, good for speed but that's about it.

to my mind, the lipsynch itself is the easy bit. Just do it before you do any other animation. you canbput expressions on afterward

[needse]

thanks for ya help AGAIN

[poopipe]

no probs dude - i put 6 or 7 months hard research into facial setups for work a few years back, so it's a bit of a specialty.

It was all towards creating this stuff - everything that moves is procedural and driven by emotional simulation and events from the world , the head shape is hooked into facegen so you can do photo generation of models or tweak away manually and create your own. there's something like 120 morph targets in that version.

What's really nice about the system is the tiny subtleties like eyelid flickering and nostril twitching etc that you just wouldn't think of doing if you were hand animating.



http://www.youtube.com/watch?v=Tx8qo...feature=relmfu