Hey Polycount,
Longtime stalker/admirer/hobby modeler here. I'm currently working on my honours thesis in Computer Science and while I'm well versed on the professional pipelines and how long its takes me to go from high poly sculpt to game ready model, for purpose of my thesis, I was wondering what the average time an amateur, or time expected time for any professional out there, spends on the re-topology stage (excluding UVing and baking)?
I'm hoping to keep it pretty loose so re-topology for current/next-gen, hero-characters/mobile/environmental. I'm just looking for any kind of ball park figure (minutes, hours, days or even weeks) to compare with for justification of research.
I've looked around at GDC and SIGGRAPH presentations and papers but other than outlining the pipelines themselves not had any luck with finding any recorded numbers. If anyone knows of some guidelines that are publicly available that would certainly help as well.
Anyhow thanks in advance.
Replies
It could be as little as an hour or so or as much as a few days, even a week depending on all sorts of factors.
What purpose does knowing this number serve you?
I guess for the sake of simplicity the hero character(round 20,000 tris optimized) or something of equivalent level of complexity.
(about maybe a day to a week depending on complexity)
Doing something like a shirt for a modular character system would generally only take around half an hour to an hour, since I'd most likely be using my lowest subdivision level from Zbrush as a starting point. When you've got a load of quads in rows and columns it's easy to roughly draw in quads and use brush-based tools to relax the mesh into clean, even topology.
Will your system take into account animation requirements? Simply lowering the polycount of the mesh is not enough. It needs to have artist controlled parameters for making animation-friendly topology.
Finally, just a personal peeve of mine. "Decimation" means to reduce by 1/10th. Somehow it's come to mean something else. /rant
Monsters and stuff without a base I usually get pressured to hurry up if it goes past 3 days.
Sucks.
Hypothetically works like mayas quad draw using surface complexity (using gaussian curvature) to populate particles that are used as verticies. In the theisis its just proof of concept but if I get to take it further id like it to take a rig as input aswell
Yeah not sure why its often referred to as decimation; in my paper I refer to it more as geometric simplification or optimization.