So right now I am a student studying 3D art. Rigid body and visualization are my areas of expertise. I mainly use 3ds Max and I feel like I have a good grasp on 3D modeling itself, just not a solid vision of the overall workflow. I can do insanely sharp high poly models, because I have the mindset of "if I am just rendering a beauty shot or a short animation, nobody is going to see if I go nuts with detail." Well currently for my internship, I have my eyes on a local company that specializes in 3D visualization and prop renders and I want at least one piece that demonstrates my understanding of the workflow. Keep in mind that I have not baked a model yet.
From what I have heard, this company requires a low poly, unwrap, high poly, and then baking for each model they do. From those who are fluent with this, does that sound standard for a low to high workflow? With characters for example, I believe the artist needs to retopo to achieve a low poly version second, followed by a bake. Is retopo ever used or even needed for a simple rigid body prop pipeline?
I am choosing a simple chair scene using Max and Vray for texturing and rendering. I would greatly appreciate any advice before I begin. I understand the pieces, but I haven't demonstrated the whole.
Replies
Let's say we make a Gun.
I first do a lowpoly-blockout model. This is used to get the basic shapes and some detail (but not too much). After the blockout-phase, it's time for the high-poly model. This will be done with booleans and detailed modeling (i.e. to get shapes of the screws, fine lines and small shapes that can be baked onto the lowpoly model and of course smooth edges). For the highpoly I work with Maya and ZBrush.
After the highpoly is complete, I take the blockout-model and reduce the geometry as much as possible (that means deleting edges, combining lots of vertexes etc.). This will give me my low-poly version of the gun.
After that I unwrap the low-poly. Then I bring both the low and the highpoly into Marmoset Toolbag 3 and bake it.
Maybe it sounds a little confusing, but I hope that gives you some idea of a baking-workflow.
Cheers
You mentioned booleans and that is concerning with topology. Is it common for the high poly to be riddled with a bit of messy topology as long as it produces a clean bake?
As far as producing the low poly, I like working off an earlier version of the high poly model, or just tweak the current high poly if that's faster. Scrolling through some AAA dev's portfolios will give you an idea of the polycount you need to be hitting and what details to keep on the model or place in the texture.
So when I modeled my chair, I used an edit poly and blocked out the cushions and armrests the best I could, paying close attention to the turbosmooth so I can work in sort of a hybrid workflow for when I need my high poly later. I am confused though why people do not use turbosmooth with 1 iteration and turn that into an editable poly and call that the low poly. Wouldn't that be a naturally better low poly mesh with a cleaner more lined up bake? If it is a matter of high poly count, I can't see that being any different from a blocked out edit poly on it's own.
This workflow has thrown me off from what I'm used to. Like I said I have always shot for the high poly first, placing as many edges as needed for the smoothest subdivided model possible for a render only. That was when I thought low poly was just turning off the turbosmooth lol. There is so much to learn.
http://wiki.polycount.com/wiki/Texture_Baking