In the workflow I do, I tend to leave the really small details to textures and Cycles microdisplacement, it’s non-destructive that way and gives you a lot of room for revision.
3 Likes
I fund both Pablo and Blender by Blender Cloud and Blender Developer fund, Pablo has remarkable improvements to Blender and the sculpt mode, He is being paid by Blender and he alone shouldn’t decide what’s better for Blender or the future of a feature that is not a minor one. Mudbox and Z-brush added dyntopo recently to improve their tools and features, so Blender removing it doesn’t seem like a step forward.
12 Likes
Having the community more involved in sculpt mode’s development wouldn’t hurt, so I would agree with that to an extent.
However, the devs. hate the idea of knowingly leaving large hacks in the code when potential alternatives either exist already or are about to be worked on (because it leads to bloat and unnecessary pressure on hardware). Until now, I had no idea Dyntopo was editmode hacked into sculpt mode with the brush system glued on, I knew there was a sort of conversion that takes place, but I didn’t know that this was why.
3 Likes
Agree, if the code is too messy it has to go. But the problem is, so far we haven’t seen a real alternative to dyntopo, just the mention of possible solutions that aren’t very clear, and on top of that, Pablo already believes his voxel remesher implementation is a replacement for dyntopo. And I think part of that is due to him not really having used dyntopo before, or maybe used it just with constant detail…
1 Like
so many people are scared and screaming. 3 options have been suggested. And no one plans to remove until there is a full replacement. I’m happy with how the dyntopo works now. I have never used it always turned on. I turned on dinotopo only when I needed to add a topology in the right places and turned it off right away. Thats why I never feel any speed issues. Pablo proved his hard work and desire to make the sculpting in blender better. Without him, the sculpture would have remained the same and perhaps would not have changed in the future. I trust his opinion and advise everyone to wait first when he makes a prototype that will show whether it is possible to replace all cases of using a dyntopo or leave it as it is. Take it easy. And don’t pounce on the developer. Nothing has been decided yet. I’m sure Pablo will prove his point.
3 Likes
I usually render with eevee so no microdisplacements.
1 Like
all we needed was to add some autoretopo/zremesher and fix multiress. We did not ask for a different workflow and now removal of dyntopo. Who asked for 20mil dynamic topology? Dyntopo is a concepting tool, same as dynamesh. If you want hyperrealistic details you need multiress not this funny remesher stuff.
1 Like
That’s some bad news.
1 Like
Maybe the removal is temporary (if it’s actually going to be removed). If the code for Dyntopo is that bad, it only makes sense to remove it and iron everything out before putting it back in, rather than trying to work around it.
People are comparing the Dyntopo in Blender to Sculptris Pro in ZBrush, but I don’t think it’s a fair comparison. Pixologic was able to add dyntopo into ZBrush because it was clean code onto an already stable application. Blender’s still undergoing improvements, and Dyntopo (from what I’ve read on this thread) has messy code.
4 Likes
FFS, can people just restrain themselves a bit? AFAIK, no one has proposed to remove dyntopo yet, only to stop actively developing it. It looks like the worst case scenario is it doesn´t get any new brushes ported to it.
5 Likes
i have been a silent reader here for long time. Just wanted to shout my opinion: actually i dont mind if dyntopo gets removed i wanna my mesh converted to quad not tris. Go for it Pablo !! 
2 Likes
Same here. But if even Mudbox offers dynamic topology these days, then it’s present in 3D-Coat, Mudbox, Sculptris and Zbrush, the four major competitors to Sculpt Mode.
4 Likes
The issue is Deprecation or Refactoring of Dyntopo.
Should Pablo let it die or should Pablo change how it works internally ?
It is completely logical that dyntopo is using Bmesh. It subdivides existing mesh.
There was an attempt to preserve as much as possible vertex colors, eddge tags & UVs.
Although Subdivision is always messing up those things by creating and removing some data.
Dyntopo is trying to preserve untouched or masked areas.
So, the whole point about performance due to use of Bmesh (hat support quads and Ngons) could be solved by an initial triangulation step before using a new Dyntopo (only supporting tris).
From artist’s point of view, that implies to renounce to any data preservation when sculpting. Maybe that also implies less flexibility when switching to another mode.
Personaly, I have no problem with renouncing to vertex color painting while sculpting with dyntopo.
(I can do vertex painting when sculpting is finished).
I have no problem with no ability to use dyntopo to make quick adjustments on an object that has UVs.
(I can keep object, destroy data of a copy using dyntopo, then use transfer data tools on final result.)
I am ready to use a restricted workflow for increased performances of Dyntopo.
I think that volumetric sculpting success is not guaranteed to work for everybody.
If it is possible to have as an alternative, a better Dyntopo, I think it is a lot more preferable to a deprecation that will end-up inevitably to a broken Dyntopo.
Strategically, I think it worth it to take an insurance that in Blender’s future ; there is no period where simultaneously one way is not ready, yet and the old way is broken.
We have a past where Dyntopo has really helped a lot of people to pass over multires problems.
I think it will be a wise decision to continue that way and anticipate that volumetric sculpting could know problems in future.
4 Likes
I think best way is to keep Dynotopo as it is (until somebody will fix it;s problems) but add warning message on enable that will tell that this feature has limitations. + short list of that limitations.
For me dynotopo will be still useful even it will only work with most simple brushes (clay, shakehook, pinch)
3 Likes
I really hope they will not ditch Dyntopo, if it came to ZB that’s not for nothing.
It’s better than the voxel remesh when you want to make your object.
I’ll say starting with voxel for the basics forms, continue to dyntopo as much as you can and retopo.
It’s an essential tool, I love it.
12 Likes
11 Likes
And most people don’t know these things. Thus comparing it to a tool that looks the same but functions differently. Functions differently, as in how it runs and works.
A little bit of Dyntopo’s history:
Back before bmesh, there was Unlimited clay instead of Dynamic Topology/Dyntopo. Which subdivided the model similarly, but Unlimited clay could only handle several thousand triangles… Bmesh popped in around 2011 and it helped function the newly made Dyntopo at the time. Nicholas bishop then did loads of work in 2012.
It’s been around 10yrs since it’s debut, if including the unlimited clay time frame. Hasn’t been worked on other than small tweaks, fixes and/or hacks since at least 2014.
5 Likes
Even more info by xrg
4 Likes
People still don’t understand what dyntopo code brings headaches. It’s better to make an alternative than trying to fix an unfixed code. That’s what hurts.
3 Likes