I'm working with someone who's been sculpting for a long time and doing 3d printing but hasn't ever needed to worry about making anything to be used in a game before. They're now trying to adapt their business so they can sell game ready heads so they're looking for a way to quickly create game ready heads from all the old heads that were prepped for 3d printing, this is on the order of hundreds of heads. So now they need to deal with uv's/topology/textures and all that. He asked me to UV a head for him and that was fine but now he wants to transfer the good uv's from that head onto all the old head sculpts he's made.
I'm thinking that this isn't possible and we should either retopo each head and UV individually or tweak the shape of the good head slightly to match the old sculpts and then just bake our maps down onto the new head using xnormal or something.
So I'd just like to confirm one way or another, is it possible to get the uv's off the new head onto old heads, or quickly change the shape of the new head to match the old heads? Or is there some other method of getting these old sculpts quickly usable for a game environment?
More info on the quality of the old heads. They are in terrible shape for a game engine. 90,000+ polys, deleted lower subdivs, combined subtools(eyes physically attached to heads,etc), separate subtools for accesories/hair, wildly varying facial expresions, men/women/children, scaling issues, etc. Basically just a ton of stuff that in my mind makes what he wants to do impossible without a ton of individual work on heads to fix them.
Thanks for any help.
Replies
You could also look into shrinkwrap-type deformers in order to have the new head match the shape of the old heads (again it might require some additional tweaking, which may or may not be faster than retopologizing from scratch depending on the complexity).
To get the best quality, you would want to re-topo each asset, by hand, and also uv each by hand. The uving actually should be fairly easy to redo by hand, as uving something like a head isn't particularly time consuming. Its the retopo that is going to take a lot of time.
If these were very similar heads, you could try "shrink-wrapping" type features to morph the shape of a base head asset, but it seems like you're talking a massive amount of variation, so that would still likely require a lot of cleanup, possibly to the point where manual re-topo would be a better solution.
Some programs may have options for transferring an old set of uvs onto a new set, look into "transfer attributes" in maya for example. XSI has some tools for this too I think, i'm not sure about max, but there is probably something.
At the end of the day, you're not going to find anything close to a "one click" type solution that will do what you're after, and if you go too far looking for quick ways to do this, you'll likely just waste a massive amount of time and decide it needs to be reworked manually anyway. My personal advice would be to just start the re-topo process now, and don't waste too much time trying to find a "magic bullet" solution for this.
Edit: Zbrush can "reconstruct sub-ds" from a high res mesh, and there is a script to do it in max too, i forget what its called. This might get you pretty far, depending on what the initial cage looked like. Unless all you have is decimated meshes, then you're SOL.
Sounds like it's pretty much the way I thought though. If anyone else can think of a good method that's quicker than retopo/uv for the task at hand I'd be interested but I think that is the solution I'm going to suggest we go with.
at UVLayout for a 'base' UV-set: http://www.headus.com.au/doc/uvlayout-expert/videos/UVLayout-Reshaping.mov
(prolly take the most extreme 2 of all the heads and set up a 50/50
base UV-set to be used for all the heads...trade-off, of course, is that
stretching/compression will also exist for all heads in the textures, but
if it's baking and you don't need to hand-edit that much for diffuse...)
Dunno if Topogun does it, but 3D-Coat will allow you to bring in a retopo mesh---a mesh that
should be generic to all heads, slightly larger in scale, having the above UVLayout
default UV-set on it upon import---and have it 'snap' to your reference mesh (the higher res sculpted version).
This would allow you to have a very basic default lowpoly template and
spend the minutes adjusting edgeflow according to each unique head.
You can also subdivide and retain the UV's on the above retopo mesh in 3DC, so you could
do minimal (depending on sculpt topology) hands-on adjustment to where you wanted your star-junctions, etc.
before subdividing.
Would be a bit of manual labor to get best fit for each particular unique bust, but
seems like a workable option...
(phew...hope all that made sense)
I watched the video and that was interesting and may work on some of the heads. However having worked with them as a sculptor before consulting I know a lot of the heads don't have the same topology or vert count among a host of other issues so I don't know if that's the ideal solution.
A lot of the work I think would be done by freelance artists so it would also require them to have the software since they don't work in house.
I'll look into 3d coat as well but I think I'm going to recommend they go with buying a couple topogun licenses. After that they can just use zbrush or roadkill for uv's and topogun or xnormal for baking.
@oglu
I'll look into it but I'm not sure how annoying or fast the process would be getting files transferred from zbrush to mudbox.
I know I'm not the fastest with retopo but I could do a head in 1:30-3 hours including new uv's depending on complexity.
My point was to make the retopo head first (the default base head geo you'll use
for all the heads), snap that retopo to 2 of the opposite-but-extreme versions of
heads you need to redo (so you have the extremes of where your retopo geo will need
to be---but those 2 resultant retopo heads have the same vert count) in 3DC, UV
each of those 2 extreme retopo heads, then use UVLayout to generate the go-between
UV-set.
But you know yer project...just hypothesis from this end.
Good luck and share how it goes if you can!
Ok, I get what you're saying now.
It may be one way of doing it and I'll look into it more. I've first got to talk to him and see exactly what he wants the final product to be to see if that has a possibility of working. Depending on how they want to sell/market this will make a lot of difference. They get sculpting but they don't really have anyone who isn't freelance that understands how things need to be for games so it's hard getting the correct language out of them lol.
I appreciate the tip though.