It makes amazing job to make a real human face from base mesh , with 4th version even animation in possible. Congratulations !
To illustrate whole point , let’s make 2 real human characters, Andrew and Mike.
For example I have already done Andrews face animation, also done Mikes face static
now I need Mikes face to perform same as Andrews, but Mike should look like Mike, not Andrew.
I mean retargeting from one face to another , Is that possible with 4th versions ?
This would be a very inefficient workflow. What you are describing is wrapping every frame of a vertex animation. You don’t need to do that. (although this is a function that is in the new versions, according to this comparison page: https://www.russian3dscanner.com/wrap4d/)
Instead you’d give Mike the same blend shapes as Andrew ? maybe 50 or so shapes that cover all possible FACS-based facial movements.
However you created the animation for Andrew, keyframes or motion capture, it’s documented as values with these 50+ blend shapes, not as Alembic 3D vertices or animated obj-sequence ? that kind of data would be huge (hundreds-of-thousands of vertices each frame), if it can even be captured in realtime in any consistent way. (Alembic node is coming in Wrap3.5) ?
So the part where Wrap3/4D fits in the workflow is giving the 2 models similar blend shapes. the blend shapes, like bones, data can easily be retargeted in realtime, so yes you can do this (soon), but I don’t think this example is a good use of that workflow.
I will try to rephrase , that is important for me to understand what that software capable of.
see that video , you will understand what I’m talking about.
face on the left is animated , so I would like to transfer that animation to face at the right (gray)
but without changing face , in other words instead of conforming faces what R3DS does, I need to retarget animation from one face to another. I want that Right (gray) guy to talk like that left guy talks.
Is that possible with R3DS?
PS
Sometimes content are blocked and video is blank.
Either needs to be unblocked from browser or use that direct Vimeo link https://vimeo.com/443320986
You still have the wrong idea. Explaining it to me again (as if I didn’t understand) isn’t going to fix your bad workflow.
AS SHOWN IN VIDEO: Andrew base mesh has animations. Mike is the desired head shape.
wrap Andrew (basemesh) to Mike (3D scan)
route the resulting wrap as a blend shape on Andrew base mesh (Andrew’s topology, with 1 Mike blend shape)
RESULT: Andrew’s base mesh is animating, but it uses 1 blend shape to look like Mike.
There is no ‘retargeting’. I am telling you the best/fastest way to do this specific example with Wrap.
If you insist on doing it wrongin a very labor intensive and elaborate way no one is going to stop you.
pf… Hello!
“Mike is the desired head shape.” NO it’s not
Tell me/show me where have I said that one should look like another.. which your “desired head shape” means.
They should look as they look, ok stop calling them mike or whatever , lets keep it like left one(animated) and right one(gray). they should stay as they are, remain same personality (no morphing needed) they are two different persons.
I don’t want to change their faces and make one look like another.
“wrap Andrew (basemesh) to Mike (3D scan)” If I will do that Andrew will look like Mike.
Instead, I just want Gray character to say same speech, like animated one does. Hope now it’s more clear.
I want just make right(gray) character to speak and perform same.
Is that possible? retarget his animation to right characters face…
I’m looking for that simple answer for weeks, will somebody give a clear answer ?
The dumbed-down answer is yes – but by the time you figure out how, you would have already done what I said.
The problem is you don’t understand the workflow. The way you ask the question (and reject the answer) tells me you have misperceptions how it would be done.
Good luck getting anyone to help you when you are too belligerent to hear the answer.