Modeling for smoothing vs modeling for clean UVs/Normal maps

Mambo4

kinda a noob question:

I have noticed that certain details on a smoothed mesh can require extreme distortions at the low poly level, for things like creases and ridges, sharp folds along eyelids etc. The smoothing algorhythm sometimes requires severe overlapping and angles on the low-poly cage. I've been getting the hang of exaggerating my low poly cage so that the smoothed version looks as I wish, but now I notice this tends to lead to UV overlaps, and artifacts in the normal map (even where the UVs are not overlapping, but the low-poly edges do overlap).

When you guys model for creating normal maps, do you avoid these low-poly cage exagerations?
If so, how do you typically achieve such details as eyelid creases and ridges/folds -just leave it to the high poly version?

Mambo4

thanks!

Mambo4

Its finally sinkining that there is a HUGE difference in approach between
1.)lo poly modeling as a base for hi poly sculpted models
vs
2.)lo poly modeling as base for a normal mapped ingame object or character.

Mambo4

this leads me to a question: with so much detial being baked into the textures, in what situations does edge flow really matter for game models?

Dylan Brady

for every part of the model that will be animated. edge flow is extreemly important.

Valandar

Like Bonebrew said. Proper edgeflow allows the polygons to bend in ways that effectively mimic how real-world objects tend to fold and bend.