Home Technical Talk

ZBrush - Seeking workflow advice / suggestions

polygon
Offline / Send Message
Krato polygon
I've been going through tutorials for Zbrush and Marvelous designer and I want to break into new ground and experiment with new aspects of Zbrush (as well as Marv Designer importing). I have a few concerns that have crossed my mind as I watched through various subject related tutorials that hopefully will be nothing at all. Below I have written out the steps I will follow, if there are any unforeseen issues I could encounter with this workflow please let me know!

1. Make Human Base model (Dynamesh)
2. Create mouth-bag and teeth/tongue (Dynamesh)
3. Zremesh
4. Export model to Marvelous > Make clothing 
5. Import to Zbrush
6. Use Transpose master to pose model.

I think one main issue is how transpose master / previous layers will interact with one another. If I create a mouth-bag and then set up the layers so the model's face has two layers --mouth closed, mouth open -- will it be reset/deleted when I later use the transpose master? If so, how should I augment my workflow?

Anyway, I haven't got around to actually testing this out (obviously) but I like to plan ahead. As such, I think the workflow is pretty straightforward and I'll post here if any future difficulties arise. 

Replies

  • musashidan
    Offline / Send Message
    musashidan high dynamic range

    Workflow for what? Production asset? Illustration?


    you can bring a posed base mesh into MD and use it as a morph target for simulation.

  • Krato
    Offline / Send Message
    Krato polygon

    Workflow for what? Production asset? Illustration?


    you can bring a posed base mesh into MD and use it as a morph target for simulation.

    @musashidan Sorry, I should have clarified. The main goal is simply to be familiar with all of these various steps, however all my efforts are to eventually make a game. So in that case, a production asset with a high poly and low poly. If I were take it out of Zbrush and add a step 7+ it would probably include painting it in substance painter and then eventually animate it in Blender. I have 0 animating experience however. 

    I imagine if I was just aiming to animate it, posing it in Zbrush would be redundant. But I would like to be familiar with that. Perhaps at that point in the pipeline I could branch it off and keep a T-pose/A-pose with clothing then export that to blender and then additionally pose another copy (or the same one I guess) in Zbrush for in-program posing.

    I've set my targets on creating a WW2 German Whermacht uniform. I think it has a good variety of solid objects as well as leather straps and other articles.  I would like to learn to be more familiar with Zmodeler, I know you have tutorials on it. To be honest though, I find it kind of intimidating. 

    Oh well, first thing is first I'll make a base body and head and go from there. I'll D-topo a mouth bag like I saw in a tutorial and then Zremesh the whole thing and post it here once that part is done. 
  • Krato
    Offline / Send Message
    Krato polygon
    Wish I had more time to plug away at this process today, got off done with gym / work by 8:30pm, it's 12am now. I made this base mesh, I've been staring at it for the past few hours. If anything jumps out let me know. I will start actually doing the aforementioned steps tomorrow. 

    https://gyazo.com/06ed3f0bd86215761cd0170ec3b1637f
  • kanga
    Offline / Send Message
    kanga quad damage
    You can stick to your your own workflow if you want, but that is not what you are asking. If you are making a game model you dont want to be messing with layers in ZB. or posing the mesh. Keeping the process simple is really enough work as it is. Just follow a simple process and you should be fine. Moving the character around once it is low poly, rigged, uved, textured is really enough, unless there are special considerations. Your mouth sack wont be a problem if it is present in the final hipoly. You dont need to remesh the final hipoly in ZB, you only do that if you need more rez while sculpting. or you need sub levels for other reasons. Just export the hirez in as many pieces as you need, decimate your model/model parts, send them to the 3d app of your choice, and use that decimated version as a template for retopo. You could also import that desimated mesh into MD as a template for the costume.

    Im not sure if I understand exactly what you want, but the simpler you hold the process the more time you have to spend on your actual product. However not everyone works the same way .

    Cheers
  • musashidan
    Offline / Send Message
    musashidan high dynamic range
    As @kanga says, if your aim is to create a real-time character then forget about messing around with posing and layers and over complicating the mouthbag. If you want to experiment with posing and layers in ZB then do it in a more focused project. As it is you may end up frustrating yourself by taking on too much at once.

    Good start on the basemesh. Once you get the proportions correct then don't spend time on doing a full sculpt, as only the head/hands will show on the clothed character. Unless, of course, you want to practice anatomy.

    Here's the steps I would take personally from where you're currently at:

    1 work until happy with proportions and do a first sculpting pass on head/hands
    2 Zremesh topo as proxy MD avatar(not too high res as it will slow down sim, and not too low res as facets will show up in sim)
    3 create clothing and proxy straps/belts/etc in MD
    4 refine clothing meshes in ZB and block out all hard-surface elements:belts/boots/straps/pouches/helmet/etc
       replace proxy belts/straps/etc from MD with blockouts
    5 re-evaluate design when all blockouts in place. Revise if needed.
    6 create final hard-surface parts when happy, using either sub-D modeling or sculpt and retop.
    7 final passes on head/hands sculpt.
    8 retop clothing and props(if needed) and retop head/hands using manual method. Model the mouthbag at this stage.
    9 unwrap uvs and bake
    10 texturing

    This is ONE workflow you could use. Remember that this is an iterative process, always. Gone are the days of such a linear workflow. This is a rough guide of steps I might take but certainly not set in stone, and not to be followed in a linear fashion. Things always change and you must be prepared/comfortable working flexibly/iteratively.

  • Krato
    Offline / Send Message
    Krato polygon
    @kanga Thanks for the input. You're right, I will stay more focused and on task. I'll mess around with z-layers at some other time perhaps. I'm curious how the whole high-poly / low-poly mouth bag thing will work. I think Pavovlich has something covering that with his "creature sculpting" tutorial chain, I'll see If i can find something on it there. Thanks again!

    @musashidan  I've refined it down to just the steps you provided. They provide a pretty good guideline and I'm not really experienced enough to bounce around in a more non-linear manner, at least not yet. I decided to work through the anatomy, something is still off with it but at this point I need to move on, the objective is after all to learn all these steps.  

    In regards to marvelous, when I make these proxy straps/belts/etc will each belt/strap be a unique subtool in Zbrush or does it all import as one tool? 

    Anyway, here is the current stage. Going forward will be the part where it starts getting cloudy. Hard-surface modeling in Zbrush, making straps, using marvelous, etc. These are all things in the next series of steps, I think I have the tutorials needed to guide me through it and I'm looking forward to the plethora of challenges I'm sure I'll encounter.

    Edit: I've been having on ongoing issue of trying to get these objects to scale correctly when they are imported into Marv. Importing the below mesh in meters makes him massive. Importing in the next step down, inches, he is an ant.  I managed to get him into what seems to be a decent size using feet but in previous test runs meters worked best and in tutorials they've used cm. Is there some consistent approach I should implement before I send him out in Zbrush? I typically unify my objects in the program, but it still seems to scale all over the place in marvy. Strange. 

    Current stage, I'll duplicate it, Zremesh, and export it into Marvelous for clothing. Critique is welcome!
     
  • kanga
    Offline / Send Message
    kanga quad damage
    Yep there is a problem, or strange tick in max, or with ZB.
    Make a box in max at .5x.5 and 2meters tall (make it transparent, its easier). I stay with cm as both viewport and system units in Max. Use GoZ to send your model body to max, its too small. Go to ZB amd at the bottom of the tool menu open the export dialogue, play with scale and the Y setting until your GoZ model fits in the Max box perfectly. Save the Ztool/model, all the settings for export are saved with it. When you work with Max on this Ztool in the future everything should be fine.

    Your model needs its hands to be in a flat pane in the front view. Later when you rig your model the wrists should have no twist in them. This model is bent slightly at the elbows to help establish the bending direction for IK. The wrists are straight though.

    Cheerio
  • Krato
    Offline / Send Message
    Krato polygon
    @kanga thank you for the continued help! I will revise the model as suggested and set up the template for marv. I probably wont post again in a few days since I'll be learning marv and modeling in Zbrush. 

    Thanks again!
  • kanga
    Offline / Send Message
    kanga quad damage
    Well judging by tour model you are off to a great start! I like the torso especially.
    You dont change the model though, just the export settings in ZB for Max or whatever app you will be using. From memory those setting will be good tor MD.

    Cheerio
  • musashidan
    Offline / Send Message
    musashidan high dynamic range
    For the proxy belts/straps it doesn't matter if they're a single import or not. It's easy in ZB to split/merge subtools by elements/polygroups.

    My scale workflow between ZB/Max/MD involves the Styx plugin. Because it gives me a perfect scale factor between Max and ZB I will import my real world scale working basemesh to ZB and do any work on it that I need, export a mid level poly proxy from ZB to Max and then export from there to MD. Import to MD using CM (this is also my scale in Max)
    Do my marvellous work (in quads is much better than tris and gives much better results back in ZB)

    Then I import my MD meshes back into Max (which I'll use as the hub of the project) And finally back to ZB. Because of the Styx plugin I always have a live link that maintains perfect scale and avoids any hacks.

    And if I need to iterate or resim in MD this workflow is really efficient as I can jump between the 3 apps very quickly, without having to worry at all about the scale.
  • pior
    Offline / Send Message
    pior grand marshal polycounter
    "The main goal is simply to be familiar with all of these various steps, however all my efforts are to eventually make a game. So in that case, a production asset with a high poly and low poly. "

    In my opinion you are tackling the process the wrong way around. If your goal is to make a game, then your focus should not be on the Zbrush side of the pipeline at all. Instead the focus should be on the necessary and sufficient things you need to do to get a character working inside your game. This means building a light, rough, lowpoly model loosely representative of your intended game character, picking a game engine, and getting started right now on the topic of animation, player controller implementation (first person ? Third person ?), and all the million other things you will eventually run into.

    Besides that, judging from this thread I would say that your other issue is that you seem to be using Zbrush as your main "hub" app. This is misguided and I think you are starting to see the issues that this can cause : counter-intuitive workflows for scene management, various scaling issues between apps, and so on. You need to pick an old-fashioned 3d app (Max, Maya, Modo, Blender, XSI) and establish it as your central hub which will be used to export assets to your game. This is how 100% of games (indies and AAA alike) are made ! Once (and only once) you have this clearly established it will be time for you to play around with Zbrush, MD, and other modern asset creation tools.

    Now of course I am taking your statement about making a game at face value - if you were only interested in creating cool standalone character assets my advice would be very different. Although a case could be made about the value of having a solid grasp on the whole game pipeline even for people who are merely creating assets.

    I hope this makes sense. Aim for your actual goal and don't waste your time. 
  • Krato
    Offline / Send Message
    Krato polygon
    @kanga Thank you!  

    @musashidan Thank you again! I'm out of town visiting family for a few days, but once I am back I will give those processes a shot, it sounds like a much more efficient approach than what I am doing currently. I did google the Styx program, maybe I can get it to work in regards to blender as well. Perhaps not. It is free, so perhaps I can troubleshoot and maybe at a minimal I can still use it between Zbrush and marv. 


    @pior Thank you again for your advice! I am certainly taking it to heart and everything I have been doing has been aimed towards the game. I would consider Blender to be more of my 'main' program. I have seen artists make beautiful characters --low poly and high, in Zbrush and I figured If I could complete that whole process in Zbrush and then only re-import into blender for rigging and animation, then I would attempt that. I may have been misguided. Perhaps I can explain my situation. My friend, who is a programmer, had the idea of making a game with me. I started up in and April I've been plugging away at it since. I have been learning Blender, Unity, Zbrush, Substance Painter/Designer, and now trying out marvelous. in April I practiced using Blender, I spent most of May  building on that foundation and becoming familiar with creating building structures in Blender, importing them into painter for texturing, and then sending them into Unity and testing them out in engine. While I am no expert by any means in those processes, I've been trying to get an introductory exposure to the respective programs and becoming familiar with both the program itself and how it fits in into the greater picture. Now, starting the 3rd month into this foray, my thought process is since I need to make pretty much everything, I can continue my practice with modeling and texturing, but instead of applying those techniques to buildings and objects, I can now further practice those skills on characters. That way I can introduce myself to rigging and animation while improving my texturing techniques. I fear I may have given the impression that I want to jump straight into making a game; my programming friend has certainly been pushing for that. However, I am very resistant to that idea. Instead, I have been trying to utilize this time to learn, experiment, fail, and try again. When we start making the game, I want to have enough of an understanding of this process, that I can give my friend a reasonable deadline and also minimize the amount of unforeseen difficulties and time spent re-doing assets.  Hopefully, that shed a bit more light onto my situation. 

    As always, love your work and thanks again for the suggestions! 
  • pior
    Offline / Send Message
    pior grand marshal polycounter
    Heya - I totally see where you are coming from ... and my suggestion remains unchanged :)

    It's of course up to you to spend your time as you so desire. The one thing you'll have to keep in mind though is that the things you are leaving for later (that is to say, everything related to actually building your game with your programmer friend : providing him with working skin weighted characters and properly looping animations, figuring out all the elements required for him to build a character controller, and so on) are actually the most complex and time consuming parts of the process ... piling up on top of a full month required to build a typical sculpted + MDed + retopoed game character.

    Courses and tutorials will of course cover one thing at a time in a linear order, because the teacher is always here to steer things in the right direction. But if you are figuring things out on your own, working that way is not efficient. If you don't really want to make a game right away you are better off telling exactly that to your programmer friend, otherwise he will likely get frustrated real fast since he will not be able to get started on implementing your character until 2 months in the future at best.
  • musashidan
    Offline / Send Message
    musashidan high dynamic range
    I highly commend your enthusiasm and passion but I fear that you're taking on way too much for someone who is only just beginning to learn the entire pipeline, including the software imvolved, and all at the same time.

    Trying to learn all that software, the pipeline, and the mechanics of getting a fully animated AAA character running and jumping around using a custom character controller is quite a monumental goal for a 2 person team with very limited experience.

    You're basically covering the entire gamut of knowledge required which is quite a lot. :)

    I'm with @pior on this one. Walk before run. Smaller, more focused projects will set you up for much greater success in the long run. Even if it takes you 6 months to get up to speed, and your friend the same familiarising himself with the mechanics, it will be well worth it.
  • Kevin Albers
    Offline / Send Message
    Kevin Albers polycounter lvl 18
    As others have said, it seems like you are setting yourself up for tremendous frustration by trying to learn a huge amount of stuff in a short period of time, in what amounts to a single iteration of work.

    Breaking off a much smaller piece of that puzzle, and then going on from there, seems much more realistic. For example, putting together a very small environment in Unity, with no characters.  Or making a small game with simple stylized assets.  Or learning how to make AAA characters, but without learning how to get them in-engine, animated etc quite yet. 

    Absolutely nothing will turn you into a combination AAA character artist/animator/tech-artist over a period of a few months. Like most skill sets, working on multiple projects over time is how you gradually pick up skills. 
  • Krato
    Offline / Send Message
    Krato polygon
    @pior @musashidan
    I'm back from visiting family out of town. so...

    Thanks again! You guys are right. I'll follow prior's suggestion and make a basic model, perhaps stylized, and work towards animating it and setting it up for unity. I would like for it to be third person, I'll leave the mouth-bag part out and just keep it simple. This will be my primary aim. During my free time / off time, I'll tinker around with a few other programs just to continue developing my understanding and keep it 'fresh', but they wont have my primary focus. 

    edit: To clarify, I will keep my "free time / off time" studies focused on one or two things for a period of time. For instance, I want to get better at sculpting, so I will attempt to do a head bust a day from reference and then perhaps I will take it into substance painter for painting If time allows. These are things I'll do if I feel like I need a break from the main flow for a few hours. I won't let myself get distracted from the main goal. 

    I'll post a new, simple character soon, skip marv, and set him up for rigging / animation in blender. If anyone has any good tutorials regarding that I'd love to see them. I understand Blender is the red-headed step child to Maya, hopefully I can find some quality stuff on it regarding animation. I have a few ideas. 

    Anyway, thanks again! I'll post soon!

    As per kevin's response, perhaps following this animation cycle I will plop the character down in a test environment. 
  • Krato
    Offline / Send Message
    Krato polygon
    Threw something together.

    Decimated version.

    Retopo. I dont think I did it that well. I hand-topo'd the head/face and combined it with a Zremesh for the body. 

    GIF of face. 
    https://gyazo.com/842604fd9ae45b0a2a09145ab4ea602a

    I'm not entirely sure how eyes are supposed to be combined together with the body for the end product in animation. 
  • Krato
    Offline / Send Message
    Krato polygon
    Made my first rig for the guy, it has a good naming convention and soon I'll be weight painting him. I've been giving this process more thought

    Make a new character that will be intended for the game. Practice rigging on him (just to get it down) and then once I start animation I wont be wasting time on a mesh that I wont be using in the game. At the same time, or perhaps before I make the intended character, make a few untextured houses that will provide a base environment for the character. It will provide scale and allow me to better visualize the project as I work through it. 

    Now, a good question for the community...

    I don't intend for the people of the game to have expressions or to talk, it's aimed to be an isometric game. Additionally, the town is intended to have a population of 50+ people as well as houses, shops, and items that can be placed inside the buildings and otherwise. As such, It's important that I keep the polycount low, but how low? If I want the camera to potentially zoom out and look over the town, while still having the fidelity to zoom into individual houses, what parameters should I set? What would be a good target range for the characters, and lastly, what kind of animation-friendly retopo should I use for the people? 

    The studies  and people I have been working on weren't intended for game use, as such I didnt really cut down on the polycount. It was more of practicing the concepts. 

    Anyway, thanks for the help!


  • kanga
  • Krato
    Offline / Send Message
    Krato polygon
    kanga said:
    @kanga That was actually super helpful! I wonder if in that case, 5k polygons per person is suitable. If all 50 people are on the screen for whatever reason, that would be 250,000 polygons just for the people. Then there is occlusion culling, which should bring that number down or perhaps Zooming so far out might not be necessary at all and at most I could see 25% of the map at a time. 250,000k still sounds steep though for just the people, some of those games had massive amounts of polygons dedicated to just one character while others seem to have less. Would having 1 million polygons on screen at any time be considered demanding for a mid range pc?

    To add to that, I feel like 5k polygons would still be generous for what I am thinking about making. Perhaps Sims level characters would be the higher end, I'll see if I can find a polycount for them shortly. 
Sign In or Register to comment.