Some of the things I currently dearly miss in Blender sculpting are:
• An upgrade of the Remesh modifier, to make it more like Dynamesh, and not operating as a modifier, but as a function inside Sculpt Mode.
• A solid quad-autoretopology tool, similar to ZRemesher.
• A topological mode for the brushes (like a topological Move brush).
• An alternative smoothing mode that keeps the surface features more intact, like the ZBrush smoothing mode after you release Shift.
• The Carver MT add-on should be usable in Sculpt Mode (not only in Object Mode), for fast cutting.
• An option to create a new layer with adjustable thickness from a masked area.
I have tried that, with other partners and users of this forum and devs put all the problems that they could to that idea and the idea that we proposed was really simple and not really big. So I gave up and accept that Blender important features for artist (UV, modeling, sculpt, painting) are forgotten and we don’t have any way to improve these part in the next 2-3 years (because blender2.8 won’t be completed until near one year at least).
for example, the main problem with multiresolution is so simple like apply to base each time that you go back one level. It automatically solve 90% of problems of the mesh.
Since UDIM is coming, that will change all the painting part, so they will have to redo it tacking cycles nodes into account.
But I really hope they will focus on core features at one point.
Adding features is great but the core features needs more love!
If you were to add features, ok, but see how 2-4 developers are going to redo the interface completely when the community only asked for small changes (and it’s not clear that the main problems, like the left click, are going to be solved).
I don’t see other programs so annoying about making a customizable interface. In Blender we have hundreds of custom themes and it appear not enough.
I do get the impression that a lot of tool development is having to be held back because of 2.8.
It seems unavoidable.
I was looking at those Sculptris ball examples. Quite amazing work and detail. A lot of them seem to be using alphas just like with traditional sub division modeling.
It will be interesting to see how it fits into the more established ZBrush workflows now.
Yeah, I was wondering about that. If it’s a reasonable price, like Photoshop ($10/month) then it might be OK, but I’m not really a fan of subscriptions, especially things like Maya ($190/month!) or Unity ($125/month). I know a few people who bought ZBrush when it was around $300, and have had all the free upgrades, so maybe Pixelogic are wondering how they can get more money out of them. Still, it’s not been announced yet, has it? Perhaps it will continue to be a free upgrade?
Whatever the cost, it does look really amazing though!
Pixologic has announced that the price of ZBrush will go up by $100 at the end of April I believe. But ZBrush is still well worth the increased price. There are no plans to go to a subscription model.
IMHO Blender 2.8 and the new viewport represent an important evolution towards RT rendering, but sometimes I think BF is a little neglecting the most important part, namely the modeling and sculpture tools. I would prefer, for the future, a substantial improvement in the direction of a zbrush (logical that you can not have everything but the road marked by Pixologic is currently the best), and maybe trying to take inspiration from Allegorithmic for the part of texturization (simple and effective). so yes that Blender could be a win-win solution …
Zbrush-like modeling and sculpting tools won’t help you much if the viewport slows to where you can’t work anymore.
The viewport code (the stuff outside of Eevee) is a very important piece that needed attention for over 10 years. Besides, one of the main points of sculpting is being able to push huge numbers of polygons without much slowdown (even with large brushes). It would also be a very nice thing to push large numbers of polygons in editmode.
About sculpting, two very simple things (I don’t know if they are easy) would fix alot.
OpenVDB is already included in the project. Build a remesh modifier using OpenVDB voxelization and rebuild polygons from that voxel base. This in my understanding is very simple and straight forward to do in OpenVDB.
I assume that sculpting in Blender refreshes the entire model. Refresh only those polygons that change. Bam, 10x speed improvement (at least). There are many questions how the Blender internal rendering pipeline works but from a higher level perspective nothing prevents you from just overwriting the Z-buffer with new data without doing the Z-test.
I think that actually it only refresh only affected polygons, because sometimes I had some graphic problem with blender2.8 and I seen how it only paint the polygons that actually are affected by the brush.
But I don’t see that problem with blender, I have sculpt scenes with 180 millions of triangles in different objects and it works well. The only thing important is tell to the objects that in object mode doesn’t have the last subdivision level.
about little things that blender sculpt needs… it true, with a few improvements it will be great. Also in workflow ¿how many times we have needed a good brush selection menu?
The dyntopo version (Sculptris pro) - it seems smart to have a connection between the detail size and the brush size.
As for sculpting and dynamesh in blender - that would be using the flood fill button in dyntopo panel with constant detail.
As for integrating a zremesher in blender - until it happens instant meshes is free. (It does not support symmetry, but you can symmetrize the model in blender afterwards.)
I personally really like zbrush ui though (I know not everyone else does), I find it very fast to work with.
That’s surely a massive step up from the original Sculptris though which struggled to get above 2 millions poly’s.
Wondering how this will go. Guess it’s simply another option. I would have thought if it was being used for super hi res work
then it would be used for blocking and concept before being ZRemeshed and moved into the traditional Sub D modeling and layers
stages ?
I’m sure that some people will disagree, but I’ve found that how much poly detail you use is often tertiary to where you place it and how you use it.
If, for example, you’ve got 10mil polys across an entire mesh, but only the front half has any real detail then you’re wasting polys.
More over, from the viewer’s perspective, once each poly’s size reaches that of a single pixel, even one more added poly is a waste. Depending on where your camera is placed, the difference between 2mil and 400mil polys could be nominal or nonexistent.
This is why, from a production standpoint, it’s still possible to have too much of a good thing. 400mil polys is not animatable. Even static meshes of that size are cumbersome enough to choke the average PC. This is why retopo exists. The average production ready model is nowhere near 400mil polys.
Would it be nice for Blender to have beefier viewport performance? Absolutely. That should certainly be a priority.
I’m just saying that we shouldn’t place too much importance on these obscenely huge poly counts. Even most ZBrush users top out at a certain point. Using ZBrush, I don’t think that I’ve ever needed to exceed 25-30 mil polys. That was across multiple sub-tools and before retopo/optimization. Most of my final models have ended up in the 1-3 mil poly range.
Not every project needs HD geometry. In fact, most don’t. So much of that added detail contributes absolutely nothing to the overall curvature or silhouette.
Again, I’m all for beefier viewport performance. Let’s just realize why we need it. TBH, as some here have said, better remeshing and decimation are much higher priorities to me. The current algorithms are terrible. Too many errors. Too inefficient. Too inexact. Too outdated.
They finally integrated Sculptris’s dynamic mesh into Zbrush, to me it’s the only natural way to quickly block out the sculpture, neither Dynamesh nor Zsphere offers this intuition, Blender’s Dynatopo is good but it’s not completely on the fly because Grab brush simply won’t do with Dynatopo.