Question regarding workflows

Highelf

So, first off, I'm not sure if these kinda threads go here, so forgive me if they don't. (feel free to delete if so)

Hey all, I obviously have a question to ask and this forum seemed to definitely house lots of people that know the answer.
Basically I'm getting into Zbrush (slowly :P) and I was wondering if a sort of "reversed" workflow is used/possible/viable using retopology features and stuff.
The general idea I had was just starting a sculpt - using zsketch/zspheres and stuff. (so a complete lacking of basemesh/lowpoly) and then after the sculpt is done, retopo it to a usable low poly mesh.

The problem - or rather, question - that comes up to me, is since the mesh now still completely lacks UV coordinates, how do you get normals and stuff onto it?

So

>zketch base
>zbrush sculpt
>retopologise to a low poly
>export and UVmap the low

> how do I bake normals onto the lowpoly? Does (for example) the max projection thing not care if the high poly has no coordinates? Or how is it generally done? Is this a possible/workable method? (it seems to save lots of time) and is it used in the industry?


Anyway, please forgive me for these questions, I just couldn't for the life of me think of what the search for in a search engine that would get me answers and I'm still relatively new to all of this :P

PredatorGSR

The high poly does not need uvs, baking programs just bake all the model detail onto the low poly uvs using a ray or cage method that does not require UVs.

If you want to paint color on the high poly in order to display or bake down later, you'll probably need some sort of uvs in order to have something to paint on, but they don't have to match the low poly.

The workflow you described and variations of it are used throughout the industry.

mortalhuman

+1

You would be surprised how much all comes from the same normal map before being put to its own on an asset, let alone how much is really only ever modeled in low poly and high poly is just floating geo etc

You'll find purist who might say that's cheating etc, too. Bless them certainly for their skillz and patience, however for most of us "Means To An End" is the definition of tools, and it's good to use any shortcuts you can - with that, good on ya for thinking outside the box and looking for improvements to your workflow, just makes you better and better the more you find.

You've got the right idea. You can even auto-unwrap your high poly to a bunch of crazy single face islands and paint on it in a 3d painting environment if you want (obviously that wouldn't be recommended - the painting sure, but the auto unwrap no :P)

For UVing your high poly, it might be best to do point-to-point seams then pelt map it or equivalent. You only need UVs on your high poly if you want to bake down diffuse from a bunch of materials on the high poly. Can't really think of other reasons for UV on a HP, but I work this way, and so, I probably just don't know more uses. :P

Roadkill is awesome and what you learn in it will translate to the more complicated interfaces of the mainstay applications nicely for you (eg; you'll see other unwrappers more clearly when you work through some meshes in it), here is a tutorial video I found and a quick run down on what it is:

http://supermortalhuman.com/roadkill-uv-tool-standalone-uvw-unwrapping

Mortague

Currently my workflow is as follows, close to what you outlined.

- Model a base in whatever program is best suited, ie your comfortable with.

- Do a rough uv layout to aid in texturing (not necessary)

- Bring into zbrush and sculpt-add textures

- decimate the mesh (temporarily for this export only to make it more manageable for topogun) export and bring into topogun and build a low poly model. (retopology can also be done in zbrush or other programs)

- Set up UVs on the low in which ever program you like.

- export the low res from topo and the high res models from zbrush, bring into whatever baker, xnormal or max for me, and create maps.

For adding color you can use zbrushes polypainting instead of setting up uvs on the high poly and using projection painting. Once you have your low res set up you can either bake in zbrush or use xnormal to bake the color data. Exporting the high res obj from zbrush does retain the colors.

If you want uvs on the high for whatever reason (they can be handy for texturing or if your baking in a program that does not read the vertex color data) zbrush does have a variety of uv options under tool>uvmap. Recently zbrush released a new plugin for uvs, looks interesting although i have yet to try it. You can get it at http://www.pixologic.com/zbrush/features/UV-Master/

Highelf

Thank you guys for all the good info, really enlightens on the subject. One thing I'm wondering still, I've watched some video tutorials now on Max projection baking and I've also tried xNormal before, it seems in max you need to do a bit of preperation and in xNormal everything is automated. Does that mean xNormal delivers a bit less quality maps (and has you cleaning up more in PS later) or is xNormal just superior in a time/quality ratio?

Mortague

I assume the video you watched on max baking, they were using cages to project and manipulate the rays and in xnormal you set the ray length. Both bakers can do either method. Which is better is probably dependent on the model and your time constraints. Cages allow more control but can be a chore. Setting the ray length is easier but can cause other headaches.

As far as which baker is better I think it is largely dependent on which program the final will end up in. If your doing your final renders in max it may be better to do the baking in max. Some experimentation is necessary to see which gives you the results you want.

To better understand normal mapping I would read through the polycount wiki on the topic http://wiki.polycount.com/NormalMap?highlight=%28\bCategoryBumpMap\b%29 It is in my opinion the best resource available and has helped me immensely.