Advantage of post skinning correctives over pre skinning correctives



What is the real difference between these two, and what is the advantage of them over each other?

Maya help doesn’t help much with the description of the former.



How the data is stored inside the blendshape node changes because of the evaluation stack. Pre-deformation blendshapes store the corrective blends with the joint rotations undone. To view the raw pre-skinning corrective blendshape in isolation sometimes makes for a very strange looking correction. But once the stack is evaluated it looks correct once again.

In general pre-skinning blendshapes are used by games engines as they are calculated on the CPU before sending the data to the GPU where the skinning calculation is performed. You could perform blendshapes on the GPU but it isn’t standard, I’ve seen blendshapes stored as delta data in textures and then driven by shaders, however this is usually for VFX or secondary shader driven animation.



I might be wrong here, but I thought that post skinning correctives were for extremely specific correctives. Because you can’t have a corrective for every possible pose, these are almost like shot specific fixes. I’ve seen this product which is made for surgically fixing bad deformation.


Hi, thanks for responses.

I’m hoping to get more input here to get a more formed idea. Thanks again


Can you tell us specifically what you are trying to do? Then maybe you will get some more in-depth answers.

As the previous answers hint, it depends what you are doing, and what the end result is for.


Typically I use extractDeltas plugin to copy the skinned mesh and correct the shape, then extract the deltas and apply the new mesh as a blendshape target to the skinned mesh. It’s added as pre skinning deformer. These shapes are basic correctives, such as elbow rotation, shoulder rotations and so on.

I wanted to explore Maya’s new workflow for creating correctives and now I notice Maya supports both pre and post.

So in this context, is there a reason to use post? If not, what is “post” used for?



In that specific context, no there is no reason to use post. But in general, are you aware of how the order of deformation affects things?

In audio, if you connect 2 effects to a guitar, ‘distortion’ and ‘delay echo’. Depending on which one you put first, you’ll either here a distorted guitar echoing, and each echo will be identical, or you’ll hear a clean guitar echoing and then being distorted on top. So each echo could be distorted differently. (insert useful illustration here.)

Post-deformation could be useful for doing some global-space deformations. Off the top of my head, maybe you want a shape that stretches your character up in Y to help mix with some kind of teleportation effect. If it was in pose-space, and their head was rotated, the effect would move up in the direction of their head.

Excuse the crappy drawing. I blame Gimp… :slight_smile:

But now keep in mind that you can also animate blendshape sources, AND you can put blendshapes on things like curves and lattices, (not just your final character) and now you can begin to invent useful, practical rigging reasons why you might want world-space deformations on something.

For example, a squid-like character has tentacles that are being driven by spline IK or ribbon IK. You can put global non-linear deformations on those ribbons. And if you want to mix multiple deformations, sometimes it is useful to mix them in using blendshapes. Because then you can paint their influence weights, or control where in the history stack the deformation occurs.


Thank you for an elaborative answer!