Well said. I agree. Pablo is doing a fabulous job pushing Sculpt Mode to new limits, but up to now it feels like an accumulating collection of miscellaneous features that’s not yet integrated in a very structured way, while something essential like Multires is still broken.
Having said this, Pablo has recently written a comprehensive Sculpt Mode UX / UI essay, so he’s working on that bit as well.
Is Dyntopo at least so unusable that it’s not possible to get results? I just tried a bunch of brushes in Dyntopo mode and they appear to work fine (as long as the subdivision is low enough to subdivide geometry).
Though I do agree that Pablo should probably try to convince the developers to make Multires more of a priority, though with the state of the code I don’t think it will be possible to fix the issues in one release cycle.
Also about the clay brush, the clay brush does have hard stroke edges now, but the regular draw brush seems very smooth now with one of the new curve presets.
And also about Multires, at least the current version is telling you to not use it in the first place (as opposed to you spending hours on a sculpt which suddenly gets destroyed), perhaps it could be considered a mercy. The modifier should probably just be removed from the UI for now.
They can’t remove it! I still need it! Even as broken as it is, I need it.
Like, as a game artists doing organic stuff, I NEED it.
ya’ll don’t get it. It’s cursed
It’s broke to hell, but it’s the backbone of the high-low poly workflow.
I’m in hell. This is hell, guys.
Devs reading this, do you know what you’re doing to me? To my soul? I can’t be alone here. Surely blender’s larger user base aren’t people just sculpting a bunch of topologyless one offs to be 3d printed or something. surely there’s a silent majority of people who are just getting flogged right now, right?
Right now, it’s broken to the point where you can get it to glitch out without even sculpting anything, just play with the buttons.
If you are somehow able to make it work, then you are of a very rare breed.
I’m not. That’s the thing, I’ve got a lot of ongoing projects where I used it and revisit em often. As of 2.81 they don’t explode much but I increment save just to be safe. After these projects are over I’m never gonna touch the modifier till BF does.
the likelihood of which is looking so astoundingly grim. Tbh It’s such a pressing and important issue that I really can’t think of anything other than undo functionality that’s more important and actually worth even thinking about working on.
Personally, I’m hoping that the development fund rocketing at least ten-fold over the past year would allow them to prioritize these things better (as they would now have the resources). The BF also has a new management approach with the aim of making development more efficient.
If not, then it may be time for the community to determine the BF does not need the extra money coming in (ie. everything above the 2.7x level, at least until they start listening to the users on critical issues). If not even that, then we should get a team together and make BforArtists the new big FOSS solution.
Then just stick to 2.79 & 2.80, they still have the old Multires, there is no need to move up in version if you still need it, I myself use 2.80 just for Eevee but I found out that I can get better and faster high quality with marmoset instead.
Eevee/Blender still has a long way to develop before I would call them ready.
Pablo is a mega army but they still need someone with more experience and a vision.
So where can one find an explanation of where we are at in terms of developing these sculpt features?
IE: Is everything basically on pause till multires? If not, what are the priorities on stuff like adding vdb modifier back, continuing to refine and add brushes, etc?
Dyntopo is not the problem, it works just fine. Could and should be improved but at least it hasn’t been destroyed.
But dyntopo is just a part of the workflow. I (and I assume others have similar workflows) start with dyntopo, then do a quick retopo or remesh depending on the needs of the project and after transfering any dyntopo detail into the retopo mesh with a shrinkwrap I usually use multires to add the finer/smaller detail. That last part of the process is completely destroyed right now.
I am not a Blender developer, but I just took a quick look at how I would start a multires sculpting tool and the situation is sightly disappointing.
What do people mean by multiresolution sculpting?
There are two common ways to refine meshes. Subdivision and Tessellation. Adaptive versions that refine the mesh depending on the current view exist for both and are probably desirable for a fast / real-time experience.
Blender’s dyntopo mode uses tessellation. I do not think the mesh is currently down-sampled depending on the view, but the feature is theoretically possible. That might be a solution, but…
ZBrush seems to use subdivision for its multires feature. The “Subdivide” and “Cage” buttons are strong hints for that. So lets implement that, because that is probably what everyone expects.
Luckily there seems to be a great open-source library by Pixar called opensubdiv. Even better, they are already supporting adaptive rendering. And as far as I know, Blender already uses this library. Great!
Hey look, they also have a hierarchical edit (aka multires sculpting) feature… or do they?
Hierarchical Edits have been marked as “extended specification” and support for hierarchical features has been removed from the 3.0 release. This decision allows for great simplifications of many areas of the subdivision algorithms.
So, in order to support hierarchical edits, someone has to either use an old, no longer maintained version of opensubdiv or they need to re-implement the whole library and re-add the feature… Neither solution looks promising in my opinion.
Yeah, what I find myself doing a lot when working on characters is putting together the basic shapes in edit mode and using the voxel remesh to combine them all together. I sculpt on that for detail until I have the look I want increasing the resolution of the remesh as I go.
Then I retopo that by hand (I don’t have the fancy new plugin) and UV map that model. Then, I take that model and lay it over the higher detailed sculpt, add a multires and subdivide as high as it can go without killing my machine and apply a shrinkwrap that’s targeted to the sculpt. This is similar to re-projecting the details in Zbrush. At that point I can bake the displacement to the lower res retopo mesh and drop the multires or just disable it till I’m ready to render.
I know the mere mention of this method send people into a tailspin. It’s considered unprofessional I guess. I work professionally in VFX and it works for me so whatever.
That’s not exactly true. It only causes issues if you go up and down in resolution. yes, I realize that is entirely the workflow that a lot of people want to use. But there are other uses for it still.
Just subdivide it a bunch and don’t touch the settings again. I don’t really use it that much in my workflow until the end though as I said above.
On a couple of my creature models, I just use textures and Cycles microdisplacement. Now with UDIM, the resolution issue is being taken care of while remaining far more memory efficient. All Blender would need to truly take advantage of that approach is realtime tessellation shaders. Sure, Multires is important, but that data eats a ton of RAM unless you add baking steps which need to be redone with every revision.
Then there’s the hypothetical future where we just don’t worry about topology anymore, Dyntopo is your modeler (while actually keeping data) and all of the painting, UV, and animation tools are updated to no longer rely on quads for best results. It’s 2020 and we’re still wrangling topology like in Blender’s early days, imagine a future where there’s no such thing as a step which locks you out of larger changes.
well, I was wrong, kind of, Blender has amazing clay brush, but it’s fill/deepen brush with those settings:
it’s fluid, with proper accumulate effect, and it’s feel amazing to sculpt
but, also, there are couple of cons:
don’t even try it with mouse
there is not inverse option, you can invert it to scrape, what, actually, for the most part I want my clay brush to do, so you will need another brush to digg
Speaking of which: did y’all notice the ‘Smoother’ curve preset that’s been added recently? I frequently select it when I’m brushing.
Yes. I also thought it will make more sense for clay. That is a Pablo’s preference.
I think it would make more sense to let user decide for all brushes which other brush is called by pressing shift or ctrl key.
But currently, you don’t need to create another brush, you can simply uncheck Invert to Scrape option.
Another thing that should be a priority for sculpt/paint modes is the ability to assign a shortcut to a custom brush.
if you uncheck this option invers option would be deepen, which is useless
and you can assign custom shortcut to brush: right click on brush-> assign shortcut
Yee-haaww! 
Man, just reading some of those complaints makes you realise how undercooked Mudbox is. Apparently Mudbox doesn’t even have a way to preview PBR in the viewport. Really goes to show how far Blender sculpting has come after all the content releases in 2019.
