A little confusion with BlendShapes/Morph Targets workflow for Unreal

DustyShinigami

Hi

I have an animtion planned for a project I'm working on, and I aim to have different facial expressions that I'll be making in ZBrush with Layers that I'll either export into 3ds Max as Morph Targets or Maya as BlendShapes. However, I'm a little confused with some of the workflow.

In order to get these animations to play with the rest of my rig/animation, do I just export the rig straight into Unreal, or do I need to bake out additional Normal maps? Are Wrinkle Maps the same as using Morph Targets/BlendShapes or are they different? Are the Wrinkle Maps the only thing that need additional Normal maps that you blend between inside of Unreal?

Once the rig is done in my chosen 3D package, I just aim on getting it play/loop in Unreal using the Sequencer as opposed to any fancy Blueprint setup.

It'd be much appreciated if someone could clear up my confusion and steer me in the right direction.

pior

Take any skeletal mesh (even just a sphere linked to just a single bone), create a handful or blendshapes for it by moving some verts around, and export it to UE as FBX. The morphs will appear as channels in the mesh properties, which can be driven by curves (imported from the FBX, or manually created in UE).

From there you'll be able to figure out everything : import settings, your use of the sequencer, and so on. 

DustyShinigami

Awesome. Thank you. It sounds incredibly simple. I just hope I don't run into any issues when I get to that stage. I'll report back here if I do, or have any other questions.

pior

Well that's the thing - you don't want to "get to that stage" as that is the recipe for disaster and frustration for everyone involved. It is not incredibly simple *at all*, as a matter it is really quite complex and the only way to understand how it all works is to try it now on a test model, rather than waiting for the final asset to be done.

In other words : figure your tech stack first, as opposed to waiting for the art to be done to figure it out.

DustyShinigami

Okay, in the end, the Morph Targets plan for that project didn't happen. However, I've been working on them and facial expressions separately. So far, I've managed to figure out how to import/export the Blend Shapes into UE5. I also keyframed an animation that cycles through them all in Maya. That animation works no problem with UE5's Sequencer. However... the problem I'm struggling with at the moment, and I just can't process or get my head around it (I caught an incredibly mild cold/sore throat, but today I just feel incredibly weak and sore all over, which isn't helping) is blending between the different Normal maps for each expression. I've checked various sites/images, but I'm struggling to 'get it'. Or it's just not working for me. :-\

Here are each of the Normal maps:

I've looked through each of these, but - 'does not compute'. 😅

https://www.youtube.com/watch?v=5wWGaJL_Q_8&ab_channel=UnrealEngine

https://www.artstation.com/blogs/leahmcewen/0Yl7/dynamic-wrinkle-maps-in-ue4

https://2.bp.blogspot.com/-GoymYzUbfKY/V50Dal4uWZI/AAAAAAAAEvg/5QKRBWDdFU8sxfCO5pRb4vxcd9XtNJfjwCLcB/s1600/11.png

https://cdnb.artstation.com/p/media_assets/images/images/000/521/043/large/MatBreakdown.jpg?1575801136

I think it's the masking part that's throwing me off. I'm not quite sure how to approach it. Do you copy the Normal map and paint directly on it? Or do you do it on a separate layer for each Normal map? Do/can you just put the painted masked areas in one of the channels (R,G,B, A), replacing what's already there? Do you do it in Greyscale? Or RGB? Stacking all of my Normal maps in a Photoshop document, and then toggling each one on/off, there's differences all over, not just specifically around the eyebrows or mouth etc. Each one is different. Try copying/pasting the images above in Photoshop to see. Therefore - is a mask even worth it seeing as the next Normal to be blended to is going to be completely different?

My head hurts... 😖