Home Technical Talk

Most efficient workflow for creating a high poly character with multiple poses/outfits?

I've spent the last few months teaching myself Zbrush and various other 3D programs. I'm mainly interested in creating miniatures/statues for 3D printing. I've been tinkering around with an idea for a while:

What's the most efficient workflow for creating a high poly character (for 3D printing) with multiple versions of the same character in different poses or outfits?

I believe the workflow looks something like this:
  1. Create or alter existing base mesh for the character's body in Zbrush
  2. Import base mesh into Blender (or Maya) for rigging into various poses, and create animations for each pose that starts in the A-frame pose and ends in the desired pose
  3. Import base mesh into Marvelous Designer and create clothes, use each pose animation to simulate the clothes in that pose
  4. Import each of the posed and clothed models into Zbrush for final sculpting pass
  5. Finally, use Zbrush to prep the models for 3D printing (slicing parts that need to be printed separately etc)
This workflow would be faster than sculpting each version of the model individually in Zbrush alone, but is it the most efficient workflow? I'm still pretty new to 3D, and I don't know what I don't know. I feel like I could easily be missing something. I'm not opposed to learning other tools/software, so if there's another tool I could use to improve my workflow, please let me know.

As a secondary question, is it possible to rig a character and create an animation for each pose using Maya Lite? I've done it in Blender, but I haven't used Maya. Would Maya be more efficient or produced a higher quality result? Is it worth investing Maya Lite for something like that?

Thank you!

EDIT: Just to clarify, I want to go for a more realistic style which is why I am thinking of heavily relying on MD in this workflow.

Replies

  • Alex_J
    Offline / Send Message
    Alex_J godlike master sticky
    that works but exporting with an animation is slightly slower. you can just duplicate the model, delete history, and use that to make the clothes on top of. 

    also you can speed up the marvelous designer part a lot by blockign out the clothes in dcc first. The UV's become the pattern. It works pretty good and saves the tedious setup in MD. 

    Best way to figure it out is do a test of course. Take blockman through the whole process.


    you dont necessarily need to mess with rigging either. Kind of depends how many poses you want to make but if its just a few you can just create selection sets and use soft select to move limbs too. Rigging is a lot faster if you already know how to do it though and have access to some auto-rigs.

    you can also explore the fabric simulation in zbrush. i dont think its as robust as MD but its worth spending a few days messing with.
  • thelittleone
    Alex_J said:
    that works but exporting with an animation is slightly slower. you can just duplicate the model, delete history, and use that to make the clothes on top of. 

    also you can speed up the marvelous designer part a lot by blockign out the clothes in dcc first. The UV's become the pattern. It works pretty good and saves the tedious setup in MD. 

    Best way to figure it out is do a test of course. Take blockman through the whole process.


    you dont necessarily need to mess with rigging either. Kind of depends how many poses you want to make but if its just a few you can just create selection sets and use soft select to move limbs too. Rigging is a lot faster if you already know how to do it though and have access to some auto-rigs.

    you can also explore the fabric simulation in zbrush. i dont think its as robust as MD but its worth spending a few days messing with.

    I suppose animation and rigging is really not necessary if its a simple pose, but for more complex poses I thought it might be necessary since it can be kind of a hassle to try and get clothes onto a pre-posed model in MD (ie arm's crossed).

    I had not thought about blocking out the clothes and using their UV's. I don't have too much experience with unwrapping UV's but it might be worth looking into for this. I can see how that would be helpful for getting the shape/silhouette you want in MD much faster.

    So far I have tried just selecting/moving limbs around in Zbrush, rigging in Zbrush using Zspheres and rigging in Blender using a rigging plugin. I was able to achieve much more realistic poses using a rig, but using a zsphere rig in Zbrush left a lot to be desired. I typically start from the same base mesh for humanoid characters and sculpt it to fit the character's proportions, so I think I could reuse a rig in Blender pretty easily to save some time.

    I had no idea there was fabric simulation in Zbrush lol. I'll have to look into that, as well.

    Thanks for the response! I have a lot to think about and research.

  • pior
    Online / Send Message
    pior grand marshal polycounter
    Learn to skin weight your models to an armature in Blender. That's it - that's literally all you need.
    Dealing with MD cloth sim sounds like a great idea in theory but in practice it will be a tremendous waste of time and will frustrate you to no end.


    Blender allows you to model, skinweight and pose all at the same time. If you ditch the assumption that you need to work "highpoly first and then somehow finding a way to pose that later", and instead embrace the approach of "posable model first and foremost" you'll be able to work 10x faster, and arguably with better results too.
  • thelittleone
    pior said:
    Learn to skin weight your models to an armature in Blender. That's it - that's literally all you need.
    Dealing with MD cloth sim sounds like a great idea in theory but in practice it will be a tremendous waste of time and will frustrate you to no end.


    Blender allows you to model, skinweight and pose all at the same time. If you ditch the assumption that you need to work "highpoly first and then somehow finding a way to pose that later", and instead embrace the approach of "posable model first and foremost" you'll be able to work 10x faster, and arguably with better results too.

    Thanks for the insight!

    I'll definitely learn how to skin weight my models. My only concern is that I want to go for a more realistic style. For stylized clothes/folds, this seems like a fine approach but I don't think skin weighting would work very well for more realistic folds and a complex pose. What do you think?

    While MD does have its frustrating quirks, I have been sewing clothes since I was a teen so it was pretty easy for me to pick up. In my experience, its been great for creating a realistic base that you can clean up or stylize with Zbrush (or your preferred tool).
  • pior
    Online / Send Message
    pior grand marshal polycounter
    Well, you'll see for yourself. There is no concern to have, because cloth sim or not, you need to learn how to do regular skin weighting first anyways. So at that point in time when it comes to acquiring the skills needed for your end goal, the topic of cloth sim is pretty much irrelevant, or secondary at most. I.E. there's no reason to mess around with that if you can't pose the underlying armature and character first. The good news is that the process is very straightforward and logical.

    Knowing MD is good of course, and knowing how to sew is great. Ironically enough I'd go as far as saying that knowing how to sew IRL (and knowing how real cloth behaves) makes one even better at sculpting it, and allows to skip sim altogether :D
  • thelittleone
    I spent a few hours yesterday experimenting with rigging and weight painting in Blender. This is just using Rigify to auto generate a rig, with automatic weight painting on the body. Then, I transferred the weight painting to the clothes. I would need to clean up some of the weight painting, but out of the box, the results are much better than what took me hours of fiddling with Zbrush.

    Zbrush really deformed the joints and the clothes around the joints. And as far as I know, the Zsphere rig is not reusable where as the rig in Blender is reusable.

    Note: I exported the lowest subdivision level to Blender which is why the Blender model is significantly lower resolution than the Zbrush model

    Blender Rigify with automatic weights (out of the box):

    Zbrush Transpose Master and Zsphere Rig:

Sign In or Register to comment.