Good luck ever using DynTopo with those numbers. I can barely push half a million vertices with Dyntopo before the lag becomes too much. Do you mind sharing your magical version of Blender, because I would love to have that kind performance.
On my hardware Voxel Remesher is capable of reaching around 8-9 million vertices on a human sized sculpt with somewhat reasonable amounts of lag to be able to sculpt. Only with a MultiRes workflow have I been able to push the 20-30 million vertices numbers. I don’t have top of the line hardware, so people who have that will have to weigh in on that discussion.
I don’t recall anyone who has said that 20 million vertices with Voxel Remesher are possible. Only way that I think would make that possible would be if you cut up your sculpt into pieces and worked on individual parts with high density remeshing, but I guess that’s not what you meant by “20 million polygon sculpt bullshit”.
I agree we need better performance. I don’t think there is anyone here who would disagree with that. Still, there are more features that are needed besides performance. Features like booleans, vertex painting, and a destructive version of the Decimate modifier for Sculpt Mode just to name a few.
One of the reasons people have mentioned why they want to sculpt in Blender over ZBrush is that the development for it is super fast and is starting to have some pretty great features that are better than in ZBrush. I.e. Pose, Elastic Deform, Cloth, Slide Relax, etc. Also, Blender is a lot easier to quickly jump into and sculpt with than ZBrush, so there’s that.
Besides, it’s not like people have stopped using ZBrush if they need it. There is no stopping anyone from using both Blender and ZBrush. Based on your “trolling” comment, it is pretty clear to me that you have only skimmed through this entire thread and completely ignored all the times ZBrush users have discussed legitimate grievances with using Blender and how to improve it. You are just poisoning the well by suggesting that people have tried to be disingenuous about what they have said and that they are “trolling” when talking about their own observations from using the software.
No, I’m serious, but I’m a stylized sculptor, so I can live without the tertiary detail level that often still needs ZBrush.
I do still love Dyntopo, and still protest against its development being frozen, but I expect Pablo to come up with a solid dynamic topology sculpting alternative. It will probably be voxel-based and use the upcoming new nodal brush system.
I am pretty sure someone said it in this thread when it looked like dynamic topology was getting the axe and am pretty sure Pablo has said something to that effect that the performance of remesh was significantly better than dynamic topology but to me at the moment it’s only marginally better, I would say 20% to 30% better.
FYI I’m going from dynamic topology to remesh, I could sculpt very well on dynamic topology at 3 million polygons.
That’s complete bull, where is it said what remesh was meant for or not meant for. Dude am an artist ( at least I fancy myself as one) You give me a tool and I will use it for what ever I please and push it to get what I want.
In general most Blender sculptors are not sculpting high frequency detail with any of the sculpting modes, Dynamic Topology, Remesh or Multires because performance is piss poor and has been for a long time.
The standard Blender work flow has always been to paint high frequency detail in as a displacement map because you can’t really hit high poly counts and it shouldn’t really be this way.
Remesh in my eyes produces a nice clean surface that should be very amenable to adding high frequency details if performance is brought way up.
You can use remesh whatever you want, but there are phisical limitations for amount of geometry that can be processed in cpu/gpu. The only way to get ridiculous amount of geo is via multiress (there was even a build of blender that could handle 200mil poly meshes and still be kinda sculpt usable) or use zbrush 2,5d hacks.
Im not sure if you are familiar with sculpting workflows for production but general idea is to block sculpt with remesh/dynamic topology then when all shapes are there make retopology or use autoretopology then go to multiresolution mode, bake all details from blockout phase and then add details. (or just use basemesh and go straight to multiresolution)
In blender autoretopo and multiress are missing. Thats why only stylized sculpt are doable. But they were doable before Pablo work.
hey i have really bad pc (pentium 4400, 8 gb, 1050ti) and i get like 5 mil per object. i just tried after your post it takes time to calc the remesh but it works it won’t crush blender but it takes time to calc i get around 17.9 mill poly. but its not possible to sculpt in my pc because of 4.5 gb of blender file, no ram left to calc the brush strokes. if you havent tried new builds just give a try. if i use dyntopo my pc lagges for 500k polys.
You only seem to use the old Sculpt Mode brushes, is there a reason for that? I don’t see you using newer brushes like Draw Sharp or Multi-plane Scrape, while those brushes are very suitable for hard-surface sculpting.
I guess it’s quite hard to see what features I’m using from this timelapse:) Yes, fortunately I didn’t found good reason to use DrawSharp and Multi-plane Scrape. I tried few times, but no.
But I’m using a lot NormalRadius, MeshFilter, FaceSets, Transform tool, Masking, and Remesh of course:)
Would be nice to see how to use DrawSharp and Multi-plane Scrape for hardsuface workflow(no idea at the moment)
It is planned for 2.90. However, commits have been pretty slow for every BCon1 release since 2.81 because they keep developing a Beta and Alpha simultaneously. Once the 2.83 Beta is over things will move along faster.
Personally I wish they were a bit more aggressive with pushing changes to early Alpha and worried about fixing stuff later after the Beta is done so the Alpha testers can find bugs and issues as early as possible. If something is breaking too much and can’t be fixed in time the devs should be able to pull stuff out of Master by the time the Beta period rolls out. At least that’s how I see it.