I’ve been up all night watching sculpting videos…oops. Anyway, I came across several ZBrush instructional videos on Sculpt Cookie. I don’t usually watch specifically ZBrush, but I wanted to this time to take notes.
I don’t know if this is a hot topic or what, but I was wondering what sculpting development looks like in Blender. I will be purchasing ZBrush sometime in the future, but until then, I wanted to explore all of what Blender has to offer. Reason I ask is because it seems there are several big features that Blender is lacking. Again, I know sculpting is not necessarily a focus for Blender users, and yes, I realize there is a reason ZBrush costs $700+, but I wanted to find out nonetheless.
Granted, there are dozens and dozens of features that ZBrush offers that Blender can’t even touch right now, but it seems there are a few that should be at least given some attention to. For instance, remesh modifier. I haven’t really experimented with it, nor do I know exactly what it can do, but after watching a few videos, it doesn’t seem to do very much. I could get more specific if anyone wants me to.
There are a few others that I could point out, but like I said, I’ve been up all night and I need to go to bed. For the same reason, I apologize for possibly coming across as spouting incoherent babbling.
Automatic remeshing is a very very VERY complicated process. Everyone would love a nice remesher like zbrush, but it took their team years and probably millions of dollars of R&D. Not really an option for blender development.
The remesh modifier is the equivalent of creating a unified skin or ReMesh in ZBrush (not the same as ZRemesher which handles edge flow)
Blender has dynamic topology, something Zbrush does not have
The big drag i found in Blender sculpt is the 3d viewport slowing down when mesh gets dense, multires modifier not always be able to sculpt with different level and the way brushes are handled. Apart from that you almost have many zbrush feature in Blender:
Zphere = skin modifier
Zremesher = remesh modifier as people mentioned above doesn’t follow edgeloop flow
Dynamesh = dyntopo; note that dyntopo is better
Blender doesn’t have any autoretopo feature but have many good add’ons make your retopo workflow fast.
In the dyntopo thread you can find a lot tips and tricks. Feel free to ask if you have any specific Zbrush feature that you want ala Blender
There is a way to combat some of the laggyness. I’m not sure what the name is but you can tell blender to reduce the multires modifier to 1 or 2 or something when rotating the sculpt. You can also alt+b around the portion of your sculpt you are focusing on which will hide everything else from the viewport at the time.
Don’t get me wrong, I am impressed what Blender does offer compared to ZBrush being open source and all. It doesn’t hurt to give the developers a little nudge, eh? :eyebrowlift2:
Aside from all of the features that ZBrush offers compared to Blender, the one thing I really like is that you are able to make a smooth and clean final sculpt. Simiple, I know, but I just love the professional look of it. With Blender, either you use dynotop (which has its advantages) but you end up with a relatively messy product. Or you use the multires modifier where the big roadblock I see is the inability to remesh along the way, causing large topology issues which translate into an inconsistent mesh as you subdivide.
Bottom line, the workflow you are forced to deal with in Blender can be very unforgiving.
I’m not sure how much Pixologic spent on getting their autoretopo up to speed, but it helps to point out that the tiny team behind 3d Coat managed to pump out such features along with both voxel and dynamic topology (live clay). So I wouldnt say its a requirement of having millions of dollars worth of funding, but rather just having smart and talented developers going after their goal.
On the Zbrush front, they do have a form of dynamic (HD) sculpting but its not nearly as straight forward as it is in 3d coat, sculptris and blender. The reason for this might be due to the fact that in zbrush you are not technically sculpting in a 3d environment, but rather a 2.5d one. The pros to this approach are huge, mainly with performance. You wont need a powerful GPU with zbrush, just the CPU and some RAM.
I wonder if blender sculpting needs to be built from the ground up once more or if its too complicated for its own good internally. On the other hand, perhaps whats really needed is some other development to pave the way first, mostly for performance purposes. Blender’s sculpting has quite a bit of potential, in part due to its ability to work alongside a dedicated modeling, rigging, animating, texturing and rendering environment.
FWIW, Pixologic have patented the key feature that makes their 2.5D approach work as well as it does. I don’t agree with the concept of software patents, but they have such a patent on 3D ‘pixols’ and they will shut down anyone looking to implement an alternative that breaches it. So any alternatives, including Blender’s, will need to stay away from what gives zBrush their edge.
I honestly think that Blender’s sculpting is a pretty good foundation, at least compared to non zBrush alternatives, so the idea of scrapping it and rewriting it from scratch isn’t terribly appealing.
I hear ya on that one, to be honest I just do not really know where it stands or what kind of technical mess they would have to wade through to improve upon it. One does get the impression that they (BF) and its developers are avoiding it or at the very least improvements to the multirez modifier. At some point though I think all the sculpting apps run into performance issues outside of Zbrush. I would love to imagine some kind of alternative to this. I know the Foundry finally got their hands on the Adaptive Field Sampling patent/rights to use it, and have some consultants (author behind sculptris) working on how to best use it with sculpting workflows, this could offer some unique competition to Zbrush performance. (http://www.fxguide.com/featured/whats-the-foundry-buying-the-tech-of-adf/)
Would be nice to see Blender devs focus on the auto-remeshing feature. Let’s face it. Retopoing a sculpted object will always become part of the workflow. Otherwise, the texturing options are limited.
When I say auto-remeshing, auto-retopo, whatever you call it, I don’t mean the one-click button. Would be nice to see them enhance what Ecclectiel started, which was the Bsurface addon. Grease Pencil is unique in Blender. Wouldn’t it be nice to use strokes in guiding the topology flow before clicking the auto-retopo button? It seems not practical to expect a one-click retopo feature. It needs at least a few inputs before clicking the button like using pencil strokes to guide the program.
Something to note, the two main things that Blender sculpting needs right now (a largely rewritten multires sculpting system and automatic retopology tools), is not something that you can expect someone to do in their spare time. The complexity of those tasks is such that it will require an experienced developer (Nicholas Bishop?) with a paid contract.
It’s much like how there wasn’t much prospect of Blender ever getting Ptex until Ton laid out a paid development plan for it, it will be the same for things like retopology as an automatic system (that works well) is not even close to being trivial.
Blender is only where it is today because the foundation pays out tens of thousands of Euros annually to make paid development possible, that’s not going to change (or not anytime soon).
If you didn’t know Pixologic didn’t actually develop Zremesher. It was called QRemesher in the beta versions. It was developed by a software company called " Verold " . Pixologic either bought it as is and implemented it or they asked Verold to do it; either case it was couple of releases I think before they renamed it Zremesher and finalized it.
my advice; buy ZBrush and be happy, incorporate it into your workflow
it won´t hurt you + it offers a lot of stuff which Blender could so far simply not deliver.
Blenders Sculpting mode offers much which would make it ace especially compared to tools like Mudbox and Sculptris but there are other areas which are immensely crippling the Sculpting area and imo that is the slow IO - or do you guys really want to always wait a couple minutes for importing and exporting files?
some people are also expressing a lot of interest for a ZBridge to cycles (a great idea)
to some extend it is already possible - if you have a lot of time on your hands…
but without tackling the big issues first which would benefit all or a real big chunk of Blender users,just to mention it again; quicker IO, Blender won´t be that much useful for Sculpting.
I would happily again pay a developer if a quick IO gets added to Blender since it benefits a lot of blender users and also the sculpting users.
so yeah … we should try to look for the bigger issues/points for sculpting and tackle these first.
Maybe ? Dr Petter was a very good move made from Pixologic. The problem here is it took some time and resources to get to this point. The failed attempts or implementations we don’t know about might have opened the way for the correct system. Maybe verold failed and Dr Petter understood the problematic path and found an alternative algorithm?
