Facial rigging/Morph Targets advice needed
Hi all,
Just about to rig my character. Firstly, I'm not sure whether to do a fcial rig, or just rig the character and then use the morph modifier after for facial expressions?
I've found some good facial rigging tutorials but they all just use a head...I assume I would just connect the bones up through the body as I'm going to do and then continue them up into the head for the jaw bone and head bone?
With morph targets, every video I've watched just uses a head. My character is all one mesh, so how do I go about creating the morph targets for just the head? Do I just select all the head polygons, shift-drag and do it like that?
Thanks
Just about to rig my character. Firstly, I'm not sure whether to do a fcial rig, or just rig the character and then use the morph modifier after for facial expressions?
I've found some good facial rigging tutorials but they all just use a head...I assume I would just connect the bones up through the body as I'm going to do and then continue them up into the head for the jaw bone and head bone?
With morph targets, every video I've watched just uses a head. My character is all one mesh, so how do I go about creating the morph targets for just the head? Do I just select all the head polygons, shift-drag and do it like that?
Thanks
Replies
Alternatively, if you havent created the rig/skinning/morphtargets yet, you can separate your head from the body if you have a convenient area to do that in to hide the seam, just to make it more managable.
As for morphtargets or a bone driven face setup, they both have their pros and cons.
with morphtargets, you are limited to linear motion, which leads to you having to add inbetween shapes to try and break up the linear path of the corner of the mouth from, say, a neutral expression to a smile.
You are also locked down with what you can do based on what morph targets you have created. you may discover mid animation/posing that you havent created a certain shape you need for the mouth to hit a certain expression/look, and you will have to go back and create these before continuing.
On the other hand, with morphtargets, you are bascially modelling each expression, which gives the freedom to create the exact look you need. You can can achieve a lot of skin sliding and folding a lot easier when modelling compared to moving a joint and relying on your skinning job. Also, it can be relatively easier to change one slider vs x amount of controllers to achieve the same effect.
With a bone based face setup what you will lack in well sculpted/modelled final poses you get back in freedom to create any expression you want (given you have enough bones etc to start with). But it all relies on your skinning job. And the more bones you have, the trickier the skinning will get since all your bones are fighting to influence the few vertices available.
And then, you have another option. Combining bones and morphtargets!
I was actually thinking about bones and morph targets together. My main concern was how to do the morph targets in the first place...not knowing whether to take off the head or not. I got confused and thought that if I copied the whole mesh then it would morph the positioning of the hands etc... Whcih I obviously didn't want. Am I right in saying that the morph modifier only morphs around the surface of the mesh and not the positioning?
An example...My copied mesh morph targets are just at a t-pose with different faces...I'm mid animation in a run cycle and want to change the face. When I try morphing this to the t-pose target, will my mesh morph back to the t-pose, or will just the face change?
So, I can rig the whole body with bones , apply skinning, then copy the whole mesh and create morph targets?
Thanks again!
That depends on when you apply the morph node/modifier
You should make sure it is below your skinning in the stack.
Also, make sure you dont freeze the transforms on the morph meshes! if you do this, your base mesh will slide to the position of the morphtarget!
I've just found this video:
http://m.youtube.com/watch?v=fgrqJCMDHHg&desktop_uri=%2Fwatch%3Fv%3DfgrqJCMDHHg
He detaches the head, creates morph targets and applies the morph modifier. He then re-attaches the head to the body and proceeds to rigging and skinning from there. This seems to be the best way I've found that I undestand. Would you say this is a good method?
Even inside of max it's dicey if you have to go back and and make some changes to either the morph targets or the body. Incorporating model updates is just a part of life for an animator and it's easy to screw yourself over using the reattach method. It's best to keep them separate and cover the seam or make the normals match.
What ways could I go about to cover the seam?
Thanks for your input.
So you dont need to have separate skeletons for the head, body etc.
As for how to cover up the seam, you should ideally find a natural cutting point, like the collar if your character is wearing clothes, or a scarf etc.
Alternatively, as mark mentioned, you can match the border normals so it appears as if there is no seam.
Make sure the verts are snapped to each other, match the boarder normals and match the skin weights.
Applying the skin modifier to a selection of meshes lets you treat them all as one mesh but they are separate objects. This is less likely to cause a problem with other applications or engines and is widely used.
Using an instance of skin, you can copy the weights from one vert and paste them to another, using the "weight tool". This makes sure they match and won't tear apart. You click on vert, click copy, then area select both verts and hit paste. But you can only copy/paste within the instance of skin with both meshes selected so it's a bit of a pain, but it works.
You can do it the long way of writing the vert weights down and manually making sure they match, but that takes a lot of time. But if whatever you're exporting to doesn't support instances of skin it might be your only option.
To be a bit faster, you can also create a copy of your mesh, weld everything together and skin weight it so it's prefect. Then skin wrap your separate meshes to the joined mesh and convert to skin. Skin wrap will copy the weights of nearby verts on the target mesh and those verts that are doubled up will only have a single vert to copy from, resulting in the same weight being given to separate verts.
So when you go to rigging the neck area that meets the body, you will know that the vertex skin data will be consistent with each other, avoiding tears during animation.
That way. You can edit the head separately and do what you want with it. Sometimes it may go all wrong and you might have to delete the head mesh and bring in another one. Usually you can just copy and paste the skin modifier from the body mesh and paste it onto the head mesh without any problems.
I'm going to use the method as shown in this series:
http://www.youtube.com/watch?list=PL3C6BDFAA684F5EEC&v=nobV52tIF5U
This video, at 3:07 in. He mentions, "it's best to bake all the envelopes down to the vertex level." Is this done by selecting each bone and clicking "Bake selected verts" within the skin modifier?
Thanks again
If you feel you should keep the head and body as one however I don't think it would cause any major problems. Just hide the polys you don't need when creating your morphs.