Question about a thought I had. What would be more efficient:
A) A standard ~20k poly mesh with approximately 150 different face and body morphs
B) A low-poly (under 5k) mesh that has no deformations, but instead its material map has all the morph data in the form of heightmaps, which get used through scalar parameters and lerps to represent the different adjustments made to it.
I know it’s difficult to say without actually putting it to the test, but what would everyone’s guess be? For both few characters on screen (5-15) and many (50+).