3Ds Max/ Zbrush: Industry vs personal workflow
polycounter lvl 12
Hi all, After reading, watching and searching on Zbrush workflow I am facing the point of decision making on the subject of workflow when using Zbrush and Max together.
I am going to prefer making low poly with enough details than importing to Zbrush for more stuff to get Normals and other maps. I dont want to end up wasting time using optimisation and simplyfying mesh. Its just me I dont like computer do calculations for me and give me a mess to deal with. I have used optimization thing before and I really hate them.
I am using UDK so I love my grid setup in Max, I am new to Zbrush and havent found precise calculation thing yet.
Compared to what experts in Zbrush are doing.
Making super high stuff, importing in Max using Polycruncher, sorting out messed up vertices, edges, Unwrapping etc.
I know in arts there is no right or wrong way and patience is the key to victory, but i dont have time as a luxury at hand and would like to know what u guys think before I fully commit myself into Painting stuff like a pro
.
I am going to prefer making low poly with enough details than importing to Zbrush for more stuff to get Normals and other maps. I dont want to end up wasting time using optimisation and simplyfying mesh. Its just me I dont like computer do calculations for me and give me a mess to deal with. I have used optimization thing before and I really hate them.
I am using UDK so I love my grid setup in Max, I am new to Zbrush and havent found precise calculation thing yet.
Compared to what experts in Zbrush are doing.
Making super high stuff, importing in Max using Polycruncher, sorting out messed up vertices, edges, Unwrapping etc.
I know in arts there is no right or wrong way and patience is the key to victory, but i dont have time as a luxury at hand and would like to know what u guys think before I fully commit myself into Painting stuff like a pro
.
Replies
There's theory, and then there's practice. The latter should give you a good indication of how the process works, which should then give you a better indication of when you may (or may not) want to use such methods. It's not something that should require a full commitment or a lot of time; small tests and experiments will suffice.
It may be worth noting that the automatic calculations (decimation master, meshlab, polycrunch, etc) are not meant to create the final product (at least for a game asset). Depending on your system and level of detail put into the sculpt, they may not even be needed. When people use them with a sculpt, it's generally just to make the file lighter so that there isn't as much data to load and process while retopologizing and baking maps. The topology is a mess, but it's only the surface that matters for the high-poly. Then manual retopology kicks in, which is actually a pretty straight forward and quick process now with plenty of tools dedicated to speeding things up.
It still may sound like extra work or an extra step you don't want to waste time optimizing (which really isn't a waste of time as it can be an important process with real-time assets), but if you want to sculpt then you may be stuck with having to optimizie things regardless. Your base mesh may need control loops to hold its shape, and will need evenly distributed (square-ish) faces to help maintain consistant behavior in areas where you want to sculpt. As a result, the mesh created for sculpting can often be different than the final mesh used in the game.
What kind of assets were you looking to create? (Characters, environmental props, textures?)
build low res version around it
bake maps
any other way of working just results in you repeating yourself or getting crap results in my experience
This kinda reminds me of good old days when i first started out in 3D, trying to make a perfect car model which I did after 4th attempt
Thanks guys.
1: You have a bad basemesh for sculpting because it's not even enough and you have to retopologise in order to get a decent high res - which then means your low-res mesh is out of synch and you have to build a new one anyway
2: your basemesh doesn't fit the shape of the high res very well and you have to build a new one anyway anyway
basically - you'll end up rebuilding something which is a pain in the arse and wastes time.
There's no reason not to use a basemesh built in max or whatever - just don't spend time making it the right shape when you can chop it about in zbrush much more quickly.
With characters i'll usually build the body out of zspheres (then sculpt it obviously) then accessorise at full res using max/zbrush/whatever .
Once that's done I'll retopologise (i use wrapit as a rule) based on the final shape , unwrap and do my bakes
for environment type stuff then a max basemesh would be the sensible way to start so's your proportions etc are correct - I'd just say don't expect to re-use it for your lowpoly later
For anything hard-surface, modular etc, the grid is very very important and usually you can get most of the form with basic modeled geometry (like walls and basic forms like that). If you can get the form that way (and don't need to bake any form) then you can do a lot with NDO2 and photoshop to get the surface details and pretty much skip high-poly work.
If you are doing a clean hard-surface that needs precision and has a lot of complex form, then model it high-poly/subd in a modeling app. Zbrush can do hard-surface but for the most part if you want any precision you're best off modeling by hand.
If you want to use Zbrush for modular pieces I'd make the low-poly piece first, have it snapped to the grid and ready to go, use Zbrush to add details that don't change the form, and use the original low-poly for the bakes and as the final mesh. There are many ways to go about detailing a low-poly without changing the form of the mesh (NDO2, per-pixel painting, baking from a high-poly, etc).
Yes I loved the method you mention Gestalt , make low poly and add normals like nuts bolts as separate layers in Photoshop, It gives me more control of what i want to add, but that is an easy way. I have done that before. I am working in 3D arts for long time and I wanna go full on with Zbrush and Kevin Johnstone has inspired me alot with his work for Gears.
Also I have just found Grid in Zbrush which is great news lol but have no clue and no forums that mention if this grid spacing is exactly like in Max/ Maya/ UDK. like 1 unit = 1 UDK unit.
Neways a long road ahead of me to learn this Zbrush, but i am really upset not many people are into Hard Surface using Zbrush, hence not much info. But hey I might share my knowledge once i get there :P. lol
Try and fail, dont fail to try
What u mean? If i make High res (decimate it use GoZ to bring in Max) than use it to bake on lowpoly. How it can not be reused again once things are baked? oO
The only hurdle is that you want to make your low poly in even Quad spacing in order to make it work Zbrush correctly. Correct me if im wrong on this one.
It helps. It doesn't have to be perfectly spaced, and the more you subdivide the less it will probably matter. Consistency will probably be the more important thing if you expect to sculpt across a surface (if one area is dense and another has much wider faces, trying to sculpt across it may look like the stroke stepped down a subdivision level. There are times when you may even want to do this on purpose, if one area of a model just isn't going to need the same amount of detail as another).
It really doesn't even have to be all quads truth be told. It's just more of an ideal, but isn't always needed to get the job done.
So today i have finally finished all Kevin Johnstone's posts and came to realise that what i am looking for needs to b modelled super high in Max than Zbrush than Decimate or Polycruncher and Low poly. Sigh I will miss Low poly stage first lol. But on a possitive note I get my pefect units I want (scale, measurements etc).
I am ready to jumpl from a plane with closed eyes (with ofcourse Parachute)
I believe in English you say it as "Leap of Faith" ? XD
HAHAHA, I think that might be where you and me share a problem! Smoking weed, and trying to make game levels and 3D Models while your stoned out of your face! Its a lot easier when your making music stoned, because music allows you to be more spontaneous with the production, and be experimental, get carried away by a vibe, trail off into a new tune entirely. With 3D Modelling though, maybe its cos I'm a beginner 3D modeller, but an experienced "music man", everything becomes so much more confusing, and I forget what I just did, so I do it twice without realising! Yesterday I made another UV map when I didn't even need to... With music though, I've made a lot of tunes with friends, and when your sharing a spliff, the whole room is just ALIVE with ideas. We can make like 6 tunes in a day, locked in the studio for about 10 hours, nothing but ganja and a lot of tea & coffee. I could never do this with 3D modelling though...
Thanks man I'll keep you informed then, cos I'm also using UDK, concentrating mostly on the "ins & outs" of asset creation right now! Used to make maps for the UT games, but always wanted to work on something that could be more than just a map or a mod. As in a game or some kind of interactive movie.