Efficiency of video game art vs. movie art character pipelines
polycounter lvl 3
I hear in the movie industry, the roles of character art is separated by hi-res sculpter, texture painter, and retopologizer. These really are essentially three very different skillsets and it makes perfect sense to seperate them. Why does the video game industry make one person do all three? Is it moving toward the film industry workflow?
Especially retopo-ing, which requires more technical skills rather than artistic (reserved for sculpting and painting).
Especially retopo-ing, which requires more technical skills rather than artistic (reserved for sculpting and painting).
Replies
this is basic eco and ergonomics.
With rising level of details in games you will notice the same move to specialists as in movies (compare for example old games/mobile games to AAA productions like The Order 1886 where you can see far more specialization). But to sum it up as almighty_gir said it is basic economics and getting the most out of your team, being as flexible as possible.
Rigging is also much more complex, there's a special rig for animation, focused around quick responsive performance and intiutive control for complex things like rotational regression. There's a special rig that drives the deformations for the character, and then once that is all said and done there is a 'finaling' process.
I don't know if games will move in the direction of film pipelines, I tend to doubt it will though. In games the art is far less important than it is in film. To top it off profit margins on games have gotten a lot tighter, adding more people to the process so you can have more people in dedicated roles will only increase the number of games that have to sell to break even and make money.
However, I think we'll soon see more specialized roles. When I worked as an artist we had 3 major programs in the pipeline for characters (ZB, PS, Max/Maya) and a few plugins. Now you can find as many as 6 (add MD, Retopo tools, Substance). Software like ZB too became more powerful, but also more complex. In the future there may be guys who just work with Marvelous designer, or Substance artists - similar to shader artists found in film.
anyway i was expecting when clicking on the topic that this is about pipeline, as in tools, asset transfer between apps, preview capabilities, quick and predictable export to engine and so on. i believe film has a leg up - or two - on games in that regard to streamline the process.
far too many hoops for my liking to (manually) jump through in any setup i've come across so far in games and it seems to only get worse with more complex assets these days.
edit: come to think of it, this split of responsibilities for creating character assets was done at a previous place in another team in the early PS3 period, so almost 10 years ago. about half the artists even came from film as far as i remember.
the results didn't quite look as tight as when created by one artist however. there seemed to be a bit of a disconnect between what should be modeled in and what should be in the texture only. also it took longer overall. the team switched back to have one artist handle the process afterwards.
Both of those people can do both of those tasks. So this is likely a "let's get as many high polies done in as quick a time as possible" kinda thing.
Hopefully game and film studios are all fairly cognizant of this stuff at this point in time, and they try to balance things out to keep the quality of the work high enough, the people diversely skilled and happy enough, and the cost effectiveness of their pipelines high enough.
I started in the industry long long ago, when an artist could be expected to do a bit of everything, and I must say, as the roles gradually became more specialized the roles became a bit more boring to me. I have absolutely no interest in just doing retopo work all day long every day, for example. One the reasons I'm currently focused on being a one-person indie 'team' right now, and not frantically trying to line up my next full time gig, is that for me it's pretty fun to bounce around between lots of different tasks. Sometimes effectiveness is not the top priority.
Artist burnout is the last thing you want because it will show in the final work.
The issue with that is that in order to get tight and clean game models under strict budget constraints, there is a need for a close to perfect symbiosis between the highpoly phase and the lowpoly.
What I mean by that is that more often than not, sculpted models (even if impressively detailed) can become a production nightmare if they are done without knowledge or care for the next steps. Surfaces that bend in a barely perceivable manner (because the Zbrush materials hide everything) causing the retopoed low to not have clean, solid forms ; sculpts with seemingly unnoticeable gaps between parts causing *very* noticeable oddities once baked, and so on.
Now of course these two steps can still be done by two different people in some cases, but to get there requires serious training because everything is intertwined. The best gameart sculpts are done with the lowpoly in mind ; good UVs (hence clean bakes) are a direct consequence of a well planned model topology ; and so on.