I was wondering what the maximum possible resolution for dynamesh is . The polycount slider in Dynamesh Master suggests it’s 10 million polys, but no matter the scale of my subtool it won’t exceed about 5 million points , and assuming that each poly has 4 points that would give me a polycount of about 1.25 million. Using the autoscale option in dynamesh master actually lowers the polycount slightly.
I was hoping that some settings in the performance sub menu within the preferences menu could help but I didn’t find good documentation about how it interacts with the hardware I am running Z Brush on. Would it help to install more ram on my machine ?
I can use 40 million point models without problem, it’s just Dynamesh that hits that point count ceiling.
Hi @zzeebee
There is a limit for how much detail Dynamesh can capture, although you may hit a performance wall where it simply takes too long to be practical before that happens, particularly if Projection is enabled. This limit will vary with the mesh and the system. If increasing the resolution slider is no longer getting the job done, a change in strategy would serve you better than a change in hardware here.
You seem to already be aware that the mesh’s size in the Zbrush worldspace can affect Dynamesh performance. Objects that are abnormally small (actual size, not simply zoomed out) may have trouble capturing enough detail even with a maxed resolution slider. Likewise objects that are too large may begin losing detail (read the CTrl-Mouseover help tip for the resolution slider). If you are free to adjust the scale of your mesh, you may try using Tool > Deformation> Unify on it to restore it to an average working size to see if there are scale related issues.
Beyond that, though, if you’re having trouble capturing fine detail on a mesh in the Dynamesh process, you’re probably working outside of its intended range of usefulness. You should transition from working on a high poly mesh at a single level of subdivision, to working on a mesh with a lower poly base and multiple levels of subdivision to sculpt fine detail. You would likely have to do this anyway for best results if planning on posing a mesh or exporting textures for it. Additional levels of subdivision, up to the practical limits of the system, can be added for increasingly fine detail.
Dynamesh works best at medium to medium-high polycount to establish mid range form and detail. By the time it starts to become impractical to use, the need for re-meshing due to drastic changes in form should be minimal, and its use becomes more situational. Fine surface detail doesn’t tend to distort the mesh that drastically. If it does prove necessary, you can simply repeat the above process.
Hi Spyndel,
Thanks for the fast reply. You are right, I am pushing dynamesh to the limit and was trying to forgo the typical workflow of remeshing and projecting because I like the complete artistic freedom and spontaneity of dynamesh and the potential of making major changes even on a highly detailed model late in the process. And dynamesh would be also a great tool for fusing multiple subtools into a watertight mesh for 3d printing if it only would hold those fine details. What’s great about Z Brush is that there always seems to be some other way to get to your goal , in my case a boolean fusion also does the job.
As of hardware, do you think more ram would help ?
Dynamesh speed is a function of CPU. I don’t expect additional Ram (assuming you’re already at the recommended) would significantly improve the tool for purposes outside what it was intended for. It was never designed to replace a multiple subdivision level process for sculpting fine surface detail, but rather to give users flexibility over when and how to enter that process.
You will not reach ZBrush’s upper polygon and sculpting potential relying solely on Dynamesh. This may be fine if you’re working for 3d print and have no topological or texture concerns. But in this case you wouldn’t need especially fine detail either because a 3d printer is unlikely to capture it.
If you understand and are comfortable with Detail Projection in Zbrush, you can switch between the two modes of working at will, and don’t give up any real freedom. After a certain point, Dynamesh is no longer required for frequent surface remeshing, but additional subdivision levels will be required for the finest detail. You may very well, however, work yourself into a trap if you continue to work on an increasingly high res mesh at a single level of subdivision.
Click to read about the trap--condensed for thread readability
Consider:
- Important functions like Transpose Master, and texture creation and export are designed around the idea of a mesh with a lower poly base and multiple levels of subdivision. If you intend to do these things, you will eventually need what I describe for the best results.
- High poly meshes will perform poorly when being posed, or unwrapped for UV Layout.
- Not every tool plays by the same rules. You will be able to sculpt easily on a much higher resolution mesh than CPU-intensive tasks like Decimation or ZRemesher can comfortably handle. If you have multiple subdivision levels, you can throw those tasks a lower res version of the mesh easily.
- If, however, you push too far and end up with a super dense mesh at a single level of subdivision, you’re going to be relying on many of those tools to get you back down to a manageable level, and they may now struggle with your mesh.
If that doesn’t convince you, I should also mention you gain finer sculpting control on a mesh with multiple levels of subdivision.
- Zbrush is designed around the idea of sculpting at multiple levels of subdivision, and has performance improvements to support it, like the way it displays lower levels of subdivision if available while rotating to keep performance fast.
- It is easier to make substantial changes to a mesh at lower levels of subdivision without surface distortion, than it is at higher levels of mesh resolution.
- Multiple levels of subdivision provide finer control over smoothing. High density meshes resist smoothing. Having lower density versions available allow you to smooth at lower levels to reflect the changes at the highest levels.
I hope you can see I’m not critiquing your workflow, but rather trying to explain something fundamental about the program that may save you many trips back here. A lot of questions from new users come down to issues I’ve touched on here. 
Hi Spyndel,
This particular project I am working on is for 3D print only, but nonetheless you gave me great in depth information and I really appreciate you taking the time !
You’re welcome! Sorry for all the text that doesn’t actually answer your original question. In my experience, though, trying to squeeze increasing amounts of detail out of a single resolution mesh is often a sign of a workflow that may prove problematic at some point. Just trying to arm you against that.
One final reminder though. It’s impossible to say without seeing your mesh. There may yet be something atypical about it that is reducing Dynamesh performance, and it will vary with the size and complexity of your subject. But generally, detail that is too fine for Dynamesh is often going to be out of range for many 3D print processes as well. With this is mind, a purely dynamesh workflow can be useful for keeping yourself from over-detailing for that output.
Also, while 3d print output doesn’t have the same topological or texture concerns as a project for digital render, a multiple subD process can still sometimes be useful. Particularly if it’s a character mesh that you’ll be posing at some point.
Good luck!
I find the information beyond my original question very helpful. I am fairly new to ZBrush and still have to figure out the workflow that works best for me. I think what threw me off the multiple sub d approach is the fact that some functions within Z Brush don’t work on it and the user is asked to delete lower levels.
I am doing environment work, a house, with a lot of individual subtools that are supposed to be merged into one watertight mesh at the end, but when I made a test to see if ZBrush still can handle it I found that working on one big merged dynamesh object lets me make changes more spontaneously . The detail is not overly fine, but the edges of the dynameshed house are slightly jagged due to lack of resolution. I tried clay polish but did not find a setting yet that really works.
I noticed that Z Brush overall is still better with organic shapes than with hard surface modeling even though great improvements have been made there.
Btw, I found a glitch with zremesher and polygroups , is that already known and if not is this the right spot for bug reports ?
It makes more sense for longer term users who have seen the changes layered in. If you’ll forgive the brief history lesson…
Click for brief history lesson if interested.
…a multiple subdivision level workflow Is the “classic” way of working, and still the way to sculpt fine detail on a mesh, but Zbrush used to have fairly limited tools for actually creating a mesh from scratch. Most often what would happen is you would import a low poly base mesh created in another tool, subdivide it up, and sculpt your detail.
Now, though, Zbrush has robust tools in its own right for actually creating a mesh from scratch. Dynamesh is one of those tools. Many of the newer tools have a slightly different way of working, and focus on working on a mesh at a single level of subdivision, because multiple subdivision levels don’t allow you to make rapid changes to the underlying topology. Mesh creation in Zbrush requires frequently changing topology, but the need for this diminishes over time. High density mesh sculpting is a separate process.
The line gets blurrier with each new release, but those newer tools are still there to support that original way of working. The idea is still that you would use these tools to create your base mesh, rather than importing it, and then transition at some point to a multiple subd process for the sculpting of high end detail, posing, and the creation of textures.
3d print output can be free of these concerns, because it typically isn’t concerned about those aspects of 3d work. It trades them for a bunch of different concerns though.
Detail projection is the key. Get very comfortable with it! It’s what gives you the freedom to work back and forth between different processes and different meshes.
I actually have a list of top 4 poorly misunderstood concepts in Zbrush that keep people from moving to the next level. Numbers 2 and 3 are:
Click to read about 2 and 3 if interested.
-
The helpful balance between low poly and high poly work, why you benefit from having both available, and the ongoing need for multiple subdivision levels. It’s not all about extreme polycounts.
-
Projecting detail from one mesh to another, and why you need to be comfortable with it to get things done between different versions of your mesh. If someone is reluctant or unskilled with this process, it can cripple them. If you’re confident with it, it gives you so much freedom to work fearlessly with many different approaches.
So now you understand those two things, and have that going for you.
This is true of any program to the extent that hard surface stuff is just more complicated, requiring a much more deliberate way of working, and often more deliberate topology. Organic stuff is more forgiving. I personally work exclusively in Zbrush for HS, and ZB is quite good at it. As with many things in Zbrush though, sometimes it requires a slightly different philosophy. This is what I’m working on in another window:
Go for it. It might be known, or it may not be a glitch. Be sure to lead with screenshots illustrating your issue, and your ZRemesher settings. Make sure you are using the most up to date version of Zbrush (currently zb2019.1.2) before reporting issues. Obsolete versions are not officially supported.
This is all great information, thank you so much ! Even though I worked through a lot of the tutorials on the Pixologic website ( which are, by the way, excellent ) I did not get the importance of reprojecting detail.
I am starting to understand that the benefits far outweigh the very long wait time for projecting dense meshes. Hopefully the developers can further optimize the process.
Thanks for the rendering, that is a beautiful piece of SciFi tech ! I agree, ZB is very good for HS modeling. I am using it more than Maya now even for Low Poly stuff.
I should have been more specific about ZB’s HS weakness - I mostly had Z Remesher in mind, one of my favorite tools , which seems to have problems especially with rectangular shapes, and the fact that Z Modeler lets you build meshes that are “unclean” but pass the check mesh integrity test without being really optimal. I will open a new topic with an example.