What I did:
Facial rigging is done! Well, for the 30s John model anyway. I’ve finished the Gustavo Rosa course, Blender 3D – From Zero to Your First Character! – all 83 lessons completed – and I now have a character who can actually express emotion.
Bones:
Controllers:
The technical bit:
Using shape keys and drivers, John can now do a range of expressions: smiling, delight, blinking, etc. – the emotions I’ll need for that section of the story. I’ve deliberately kept the rig relatively simple for this younger version since he doesn’t need heavy wrinkling yet.
Thinking ahead:
The animatic I roughed out has been invaluable; acting out the scenes (pretty badly) on video helped me work out exactly what facial performances John needs at each age. I don’t need loads of expressions, just the specific ones that tell this particular story. This forward planning has been crucial – I know which shape keys are essential across all age variants and which ones would add unnecessary complexity in a limited time.
For the older models (50s, 70s, 90s), I’ll need to expand the rigs to handle sagging skin, deepening wrinkles, and the specific facial changes that come with ageing and dementia. But having the base rig sorted means I can build on this foundation, rather than starting from scratch each time.
I’ve also generated reference images using Microsoft Copilot to guide how John will change for each decade, which should help keep the ageing consistent and believable.
What I learned:
Facial rigging is less terrifying than I thought it’d be! The Gustavo Rosa course gave me a solid foundation, and whilst I had to figure out some bits myself (like closing eyelids and rotating eyeballs), Rosa’s basic principles transferred across well. The key is thinking ahead – understanding what your character needs to do at each life stage before you start building the rig.
Challenges:
The main challenge was resisting the urge to over-complicate things. It’s tempting to add every possible control and shape key “just in case,” but that just makes animation harder later on. I think it’s better to build what I know I need based on my animatic, then add more if required. Knowing I need to create three more age variants also helped me stay disciplined.
Next:
Now I need to actually create those older age variants. The reference images from Copilot will guide the modelling changes, but the real challenge will be ageing John convincingly, whilst keeping his identity recognisable across five decades. The facial rig needs to show both physical ageing and the specific changes dementia brings, without losing the person underneath.
