ZBrushCentral

ZRemesher loses details when converting details

ZBrush Usage Questions
#1

Hi all,

See below my mesh:

Originally, the teapot has 1.6 million points (I assume points = vertex points and not polygons), and I previously used Decimation Master to decimate it down to 100,000 points.

Problem is, it’s all triangles, so I use ZRemesher to convert it to quads.
The Target Polygon count setting in ZRemesher is something like, if you set a value of 5, it’s 5000. I set it to 100 for max quality. However, the result ends up looking kind low-res:

How do I convert these triangles to quads using ZRemesher without losing so much quality?

I’m trying to optimize the polygon count so that I can use GoZ to import it to C4D without C4D crashing or lagging.

Oh, and the mesh currently has no subdivision, just Dynamesh 1024.
Document size is at 3014 px by 1807 px, which is rather high-res IMO.

Tks!
Nik

Attachments

1211.jpg

1212.jpg

#2

You’re missing the point of decimation master and zremesher.

With decimation master all triangles is fine (and correct). Why is that bad?

With Zremesher you need to subD your model back up and use project all to get your details back. You then use UVs and maps to transfer detail back and forth to other applications

#3

Tks beta, but I always thought triangles are evils that must be avoided at all costs.
At least that’s what I read from forums and 3D modeling books, which always stress the need for quads (either quads or die). And also, quads look really beautiful when you turn on the wireframe.

“With Zremesher you need to subD your model back up and use project all to get your details back. You then use UVs and maps to transfer detail back and forth to other applications”

> Is there a video tutorial somewhere that illustrates the whole process? I did find a couple of displacement map tutorial on transferring details but it seems kinda confusing LOL

Tks!
Nik

#4

Run ZRemesher on double and look at resultant mesh, you can undo if not happy. Holding alt while clicking the button does it a different way. You can “project the details” from the original high res mesh to the lower res mesh(within reason, you’ll still need enough points) HTH

#5

There are quite a few videos out there that illustrate this process. Are they free? Probably not.

But youtube should turn up a ton of videos for how to use project all and the zbrush docs cover it pretty well too. Basically, you need a video on each of the things I mentioned. Once you learn what it does, then you’ll actually understand why to use it, and when.

Watching a single video of “hit this button, now this button” doesn’t actually teach you anything.

Triangles are bad…for modeling…but you’re not modeling anymore. You’re on to the rendering section. A decimated mesh will work just fine.

If you want a clean wireframe to show off, then you can’t use Zremesher, or any other automatic mesh generation…what would be the point in saying “Everyone, look at the thing I clicked a button for!!”. It would be the same as me making a teapot in max and putting that in my portfolio.

#6

LOL. Scan data or imported mesh is same thing.? Either you’ve actually modeled it or not should be the question? Can’t forget the render cool button either ;).

#7

LOL tks guys! Ok, I’ve imported the thing into Cinema 4D via GoZ, and although the whole mesh looks clean, I must say the details (the embossed flowery designs) looks kinda blurrish and pixellated. I think the alpha image for the flower was 512 by 512. Not sure if using a 2048 px x 2048 px image would give me a super clean and high-res embossed effect.

LOL @ the modeling of teapot via 3DSMax

On a side note, I think the mesh created by ZRemesher looks really clean enough to show off. Maybe not for character meshes, but I think for hard surface non-animated objects, it’s ok. :wink:

ZBrush FTW!

#8

No, the mesh probably looks great, but there isn’t a reason to show it, because you didn’t actually do anything. If nothing else, it should go in the portfolio of the people that wrote the maths behind that, not really as a show piece for an artist.

#9

I’m seriously considering doing a write-up for decimation vs remesher vs retopo etc. Every time I see posts like this I see confused people getting caught up in the hype behind these new 3D tools and techniques, like retopo etc. These new tools and techniques give the artist more freedom and time as 3D software and techniques advance and change. I see people using these new tools and techniques because they feel they have to or think it’s part of the process when they really shouldn’t even be touching them in the first place.

These tools are great and give artists more freedom and save a lot of time and hassle but I also see them hindering users that don’t quite understand what these new tools are really there for. There are old techniques that should still be used today, not ignored and forgotten.

#10

Yup, pls do a write-up if you’ve the time. :slight_smile:
I love how ZBrush has all these legendary buttons like “Fix mesh” and ZRemesher.
It’s like you can close your eyes and sculpt freely and then “Fix mesh” etc and never worry about overlapping faces again etc.

#11

My thoughts are the same as everyone else. I think you’re a little confused with what the tools are for.

My tuppence worth is…

Decimation Master - for taking a very high poly model and cutting the polys down dramatically (and I do mean dramatically), without losing any of the detail.

Uses could include taking a model and sending it to a renderer for a quick render (without UV textures), or cutting polys down to export the tool for 3D printing. It also is useful to do it to a tool before using Zremesher sometimes too, as it reduces the time that tool takes.

ZRemesher - is for taking a high poly model, either dynamesh or lots of division levels, and reducing it to a much much lower level of detail. It can sometimes give you very good results in terms of poly flow, but is rarely going to give you something good enough for animating organic shapes (in my opinion). It is more useful for taking a model you are not going to animate, that either has bad topology (uneven) or is a dynamesh shape, duplicating it, running Zremesher to get a good low poly version (all detail gone). Then to divide it as many times you need, while projecting the detail from the original back on. This can low poly can then be used to create displacement maps, or normal maps.

Uses are going to be more about having a lower poly version of your model, which has UVs and displacement, normal or texture maps that you apply to it.

Retopology - either the old way, or with the retopology brush is for the same thing as Zremesher, but were you have control of the new topology. Much better for situations where you need specific poly flow, like for animation etc.

This is how I use them all any way :slight_smile:

#12

~BetaChannel
“No, the mesh probably looks great, but there isn’t a reason to show it, because you didn’t actually do anything. If nothing else, it should go in the portfolio of the people that wrote the maths behind that, not really as a show piece for an artist.”

I don’t fully agree with you on this statement Beta… Zremesher dose not design, sculpt and model for you. It certainly is not like you are borrowing a pre-made mesh and calling it your own or Rendering a bought figure in Poser/Daz3d and saying you truly made that.

ZRemesher is a Topo tool. A good one at that.

It’s much like saying that in using Clip art as an alpha completely invalidates a character sculpted from a sphere. :rolleyes:

Sorry if I misunderstand your statement. Perhaps I have your context wrong.

#13

There’s basically two main types of modelling, Box modelling and Sculpting.

(I know there are others, nurbs/spline, procedural, etc. but I’m just using these two for the purpose of my post.)

Before sculpting became more mainstream and viable for computer hardware then modellers used basic box modelling and gained more detail as things went along. Current sculpting software and tools are just starting to bypass the older box modelling techniques. This is great for artists as they can model ‘on-the-fly’ and make changes as they go along, changes that in the past would have been time consuming. This gives the artist more freedom and clients more variations of prototypes and idea changes within their time and budget constraints.

However one thing people tend to forget is that the older techniques are old but not worthless. Box modelling is still very useful today. Proof of that is in Pixologic incorporating it into Zbrush 4R7. They wouldn’t be putting old techniques into their next release if they weren’t useful.

I’m starting to ramble so I’ll get to my point. As people see things in 3D, or CGI, they tend to imitate what they see or hear about. Posting images of wireframes is one of those things that people post without asking themselves why. In the past models were more restrictive on polygon counts as the output rendering is limited by current hardware, this is most noticeable in older video game systems. Topology (wireframe) is very important for rigging and animation. This is still very true today as it is in the past. I won’t get into the reasons why. So with very restrictive polygon budgets and rigging and animation limitations the topology of models was extremely important. This is why wireframes are shown. I could have two different versions of the same model, one with good topology and one processed through a decimation sequence. Rendered out these models would look the same and have the same polygon count. Animated the decimated one would be horrible. This is the main reason wireframes are shown along with polygon counts. A low polygon count doesn’t mean everything if the topology flow doesn’t work for the intended animation.

So ask yourself why you’re showing the wireframe.

  1. Showing the proper flow of topology for a model that is going to be rigged/animated.
  2. Showing off your modelling skills, showing that you know how to properly model, especially highly detailed low poly assets
  3. Showing comparison of high poly and low poly assets

I think what BetaChannel is saying is that Nikar shouldn’t show a wireframe just to show a wireframe. The model isn’t complex but the sculpt is. Focus on the beautifully sculpted model instead of its wireframe. That’s where the talent of this particular piece is. Now if he was planning on creating a low poly asset for a game or something I think the wireframe would be worth showing.

#14

It basically goes back to my earlier post. Some people are using things; posting wireframes, dynamesh, zremesher, decimation, etc., just because they are there.

It reminds of when my old professor talked about the same thing with PowerPoint slide transitions or Photoshop filters, “Don’t use them just because they are there.”