Dynamesh -> Zremesher workflow?

Greg DAlessandro

I want to try this workflow, but I have a few questions:
From my understanding you start by sculpting the base form with dynamesh, then use zremesher, and then divide a few times to sculpt details onto the surface. But my question is: where in this process do you start to go asymmetrical? (while you are still working in dynamesh, or after you use zremesher?)  Thank you.

I am currently sculpting the C1 Atlas and C2 Axis bones (working from a medical skeleton model)
I have been sculpting details while using dynamesh

My goal is to have a good workflow, to get better at sculpting, and to not have a model that looks like a piece of gum ( ie. blobby) 

Alex_J

Depends how big of a change the asymmetry will make. Will it require different topology? You'll have to test it out to see. I think that's the crux of the decision. If not, the longer you can keep symmetry, the better IMO. Saves time.

Looks like with these reference bones the asymmetry is on the surface level, so maybe that can be added at the end in a "detail pass."

Greg DAlessandro

Would you only go asymmetrical after you use zremesher?

Alex_J

Probably.  But no reason to make a rule from it. Usually exiting from the dynamesh part of the workflow means I'm certain about proportions/forms. And by certain, I mean I'll change it seven more times .

Like I said, I think the crux of the decision is solely, "will I need to make changes that will require new topology?" if the answer is yes, might as well stay in dynamesh, regardless of symmetry. But I've only done this workflow a few times, so somebody with more experience might be able to explain some different situations better than I can.

Greg DAlessandro

I'm not really sure about when to make the transition, (I find myself getting caught up in 'surface level details/asymmetrical differences')

Alex_J

You've just got to take time to test it out on a case by case basis, and pretty quick you'll develop a feel for it. 

So always save out your scene or ztool first, make a duplicate, and make a remeshed copy and see if you can sculpt the detials you need without topology limiting you. If you are fighting with the topo, then you know you just need to build that part and dynamesh it before remeshing.

Looking at those vertebrae, I'd add asymmetry in the final detail pass. Didn't look super close at them, but first glance suggest they just have differences in little bumps on the surface. A normal map is perfect for that. If one vertebrae had a lump only on one side and it protruded out enough to show on the silhouette, then I'd add that in before remeshing. Otherwise you'll need to subdivide too much, and matching your lo and hi for baking will be troublesome. 

In short, there is no rule for this. It's just judgement call that comes from experience. So invest the time to experiment and don't worry about breaking anything.

Greg DAlessandro

Does anyone else have any suggestions/insight? Thank you.

But what if the asymmetry is pretty significant? The C1 Atlas's transverse processes are twisted differently, the C2 Axis's Inferior Articular Process are shaped very diferently, and the  C2 Atlas's Posterior Arch and Spinous Process are very asymmetrical

I updated/organized the pictures, although there is a lot of distortion and field of view differences between the reference/my model. The model pictures were taken with the perspective setting turned off.

carvuliero

Ask yourself first whats your end goal => just trying new workflow? Would you use those models for anything else but studies ? Are you satisfy with Zremesher topology ? and will you need to retopo again ? Why would you prefer model with subd levels ?

As far as asymmetry goes its up to you both a valid :blocking symmetrical mesh and then add asymmetry or blocking asymmetrical right away

Dynamesh to Zremesher : I tend to stick with dynamesh for as long as I can if you need more poly just up the resolution x2 x4 [trimdynamic is a good way to keep your models from looking blobby or the other polish brushes ] I will only transition to zremeshes topology if I will go all the way to fine details with the model or model have parts that are interpenetrating and I want one model not pieces

Greg DAlessandro

End goal: to use the skeleton to add muscles for anatomy study (almost like an echorche sculpt). I want to try this workflow to see if it would speed up my process of sculpting. I feel like it would force me to only work on the major forms before I am able to work on surface level/micro details. I find myself sculpting in dynamesh and not making any progress: I'm spending a lot of time by constantly making adjustments, but not getting it right. I feel that maybe I am working at too high of a resolution and by lowering it - it will cause me to get the forms right faster then work on how the surface looks. 

sprunghunt

Greg DAlessandro said:

End goal: to use the skeleton to add muscles for anatomy study (almost like an echorche sculpt). I want to try this workflow to see if it would speed up my process of sculpting. I feel like it would force me to only work on the major forms before I am able to work on surface level/micro details. I find myself sculpting in dynamesh and not making any progress: I'm spending a lot of time by constantly making adjustments, but not getting it right. I feel that maybe I am working at too high of a resolution and by lowering it - it will cause me to get the forms right faster then work on how the surface looks. 

You can always work at a low resolution in dynamesh and then increase the dynamesh resolution to work on finer details. 

Greg DAlessandro

True. I am currently working at 320 resolution in dynamesh. What resolution would you work in?

carvuliero

Dynamesh is size dependent so no magic number here .rule of thumb keep it as low as it give you barely enough poly to model what you need

I will work for the most part on something low like middle one and when I am happy with overall shape then crank one up for smaller details like the 3rd one

If second is 128 , 3rd is 256 .., i work in power of 2s

If you give this workflow a try I suggest not to use smooth but trimdynamic instead

Greg DAlessandro

Thank you. I would increment by 64, and found myself zooming in to work on surface level details before nailing down the larger shapes.

Also, is it bad that I would use an eye-patch to view the physical model in a monocular way? I would place the model directly in front of my sigh-line to view the model from an orthographic perspective and then sculpt in an orthographic perspective (to model what I see) But then I would move the model/my sculpt into another orthographic perspective and make adjustments. I find myself changing views a lot, but never sculpting in all different views.

carvuliero

I have seen painters use monocle so if it works for you use it

Greg DAlessandro

I was told that it is better to frequently rotate your model around while sculpting (rather than select an orthographic view, sculpt a bit, then select another orthographic view, and repeat

Alex_J

You are not a soldier following orders. After you collect advice, try it out, then it's up to you to decide if it's something that works or not. You've got to think for yourself. Getting feedback on your process from people with more experience is a smart thing to do, but I think you are taking it too far. Overall, it looks like you are doing just fine and I encourage you to keep on carefully considering your process, but not relying on outside confirmation for every small detail. You've got this.