And to point out the jaw line issue:
Consider those planes flat for understanding their shape, and pay attention to the back of the jawbone that is quite pronounced in the front view. A thing to keep in mind is that if you made a head out of a box, the points on the jaw are as close to the side of the box as most of the features on the side of the face.
This is helpful. I guess I need to print loomis facial planes reference and hang it onto my wall⌠or somethign like that :-\
You may want to hang up a skull too. Right now youâre mostly missing what the underlying bone structure is like.
:lol:
Every unity developer must own a skull.
But youâre right. I have not properly considered bone structure when sculpting.
Who doesnât aspire to eventually have a skeleton their a closet?
Looking good neggy. As they say, practice make purr-fect 
When you get good nothing beats admiring your handy work. If maybe a little creepy.
Alright, how about this one?
bf5ea3e83a8a452b8027faff2ea62b86
I think youâre at the point, or at least close to, where Blender is holding you back. As much as I love blender, itâs just not as good for sculpting as zbrush. Theyâve recently launched zbrushcore which seems to be a kind of entry-level zbrush without most of the special features youâre never gonna use anyway. Might be worth a look:
http://zbrushcore.com/features
I canât tell you any details, I âonlyâ have regular zbrush and Iâve not done that much with it.
Their Wacom Intuos bundle is very tempting. Iâve been thinking of picking one up to start learning artwork and that bundle is almost the same cost as a tablet by itself.
Care to elaborate? As far as Iâm aware, zbrush easily handles larger polycounts, but you still have to manually worry about regenerating geometry.
https://www.youtube.com/watch?v=y8hHiV9XWEY
Blender require more tweaking to get proper brushes, though.
Easily handling larger polycounts is part of the reason why you donât need to resort to dyntopo sculpting in zbrush, and dyntopo is imho the reason for much of the âblobbynessâ problems in sculpting. I would elaborate if I could, but comparing different sculpting tools Blender just felt the worst for me. My opinion is somewhat based on subjective experience.
The recommended zbrush workflow for avoiding the blobbyness issues is starting at lowres, sculpting as much detail as resolution permits, subdivide, and repeat the process till you are done. That way you get clean results and meshes with sensible topology. Imho topology does matter for sculpting, not in the sense that it would matter how it is arranged in the end, but in the sense that the topology has influence on how the tools behave when interacting with the mesh. I think dyntopo has its place for concept sculpting where you donât know what it is going to look like in the end. But if you are going for human heads I see no reason not to start with a normal sculpting basemesh and use the standard subdivision workflow (multires modifier in blender).
This mesh here:
https://forum.unity3d.com/attachments/low-poly-male-png.218445/
Actually would be pretty bad as a sculpting basemesh imho. You need roughly even polygon topology with ideally all quads, and the least amount of distortion possible (~2:1 aspect ratio quads are bad for sculpting, those should be cut into two ~1:1 aspect ratio squares).
Edit: taking a quick look at the video you posted, it actually is an excellent example of how every hard edge in dyntopo looks terrible up close. Thatâs why you have to retopologize such a mesh to get a decent result. But with that kind of model imho you should do that anyway. Itâs a lot of work to get a super clean hardsurface sculpt and then you have something thatâs super hard to make any further changes to, compared to regular hardsurface modelling with standard subdivision techniques. For that stuff Blender is fine imho.
Hmm. Interestingly, I have opposite experience. I had a chance to use zbrush and workflow for me did not click.
Dyntopo is something I âgetâ almost on intuitive level, but clearly need more practice with it.
Basically, I have a gut feeling that I need to stick with this approach for some time, but Iâm not sure if the âgut feelingâ is correct. Blender sculptor seem to lack certain zbrush features mostly related to painting, and zbrush models pop up on boards like pinterest more frequently.
And Iâm more interested in concept sculpting at the moment.
For âexplorativeâ concept sculpting where you donât know what youâll end up with, dyntopo certainly takes out some pain. But for a human head I donât know what there is to âgetâ about multires sculpting that dyntopo solves. If you donât like the limitation of having to focus on big shapes first, and gradually increase the resolution and scale of detail youâre working on, then there you go, that might be your biggest problem with sculpting. Having a well structured approach from big to small helps with most art tasks. Workflows that allow you to break out of the well structured path more easily (or accidentally) might be to your detriment.
Knowing when youâre âdoneâ with one sculpting step before you increase resolution further is very important and not easy. It might help to either take a break, continue the next day, or look at your work in another context (different, shader, different lighting, different software, different size, different perspective). Itâs like the âfliptestâ for painting, but with x-axis-symmetry-sculpting youâll need something else to make your brain look at it more critically.
Iâd personally prefer sculptris over blender when it comes to dyntopo sculpting. To me the brushes feel better. But the tool is⌠barebones, unstable, and no longer developed. So your milage ma vary and Iâm sure youâve tried it already.
Zbrush has so many features, I donât know even half of what it can do. It certainly is tons more than âmostly related to paitingâ when compared to Blender. Itâs also pretty much the least intuitive program I know, so I donât blame anyone who doesnât like working with it.
Youâre doing high frequency work without mastering low frequency. Thus you have zero experience in the basic forms first.
Only go up a detail level when forced, clearly youâre going up detail levels when there is more than enough detail previously to express a form.
This is a common feeling. Iâve tried zbrush trial offer no less than 4 different times to try to get accustomed with the workflow. I have strong dislike for the workflow and will never be able to work the way it requires.
I was also going to suggest sculptris, for concept sculpting, though importing base meshes into it is still the suggested workflow so a solid fundamental grasp of poly modeling is still needed - unless you can find a base quad mesh online to begin with. I know polycount has some base quad poly models available that can be used as base meshes for sculpting.
Sculptris has a default sphere which is fun to piddle with, and can be shaped into anything, but a base mesh import is so much easier for any sculpting package - imo.
Iâm not sure why people are saying you canât lower your mesh density in Blender in dynotopology Just use collapse short edges with low brush strength and you can reduce your mesh density.
In my experience dynotopology can get much more detail without performance lagging than multirez. Use multirez if you arenât going to retopo it, or if the density is even throughout the mesh. I prefer dynotopo though myself.
Perhaps the idea is to preserve high frequency detail while working on low frequency portion of the mesh?
Iâm aware of density reducing trick. With relative detail, zoom out, touch any area where you want details to be reduced, and dyntopo will greatly reduce number of triangles there.
Tried multiresolution sculpt.
I gotta say that I definitely strongly dislike this workflow.
It might be great for making normalmaps on existing lowpoly models, though.
I think Iâll stick with dyntopo and destructive detail removal for now.
Why exactly?
The short version is that having to deal with individual points and topology pisses me off.
Basically, starting with simple shape then going through levels doesnât seem to give me any advantage, subdividing it alters whole model (one thing I didnât like about subdivision surfaces - slice a face somewhere, and half of vertices change their position), and after subdividing model quickly arrives at state where face density is higher than what I would have in dyntopo.
I might revisit the approach in the future, but it is not gonna happen any time soon. I think I might need to consider a voxel tool or an actual clay.
