I'm trying to figure out the most efficient way to model game assets as an indie developer. From what I can tell, sculpting is much quicker than manual polygonal modelling for things like characters. However, it requires a retopology step to get the polycount acceptable.

I've seen the following methods of retopology:

  • Shrinkwrap edge modelling - can take an hour
  • Tool-assisted shrinkwrap edge modelling - speeds it up a lot, but still a lot of work
  • Decimate (automatically remove vertices) - instant, but ugly and inefficient end-result
  • ZRemesher - automatically creates a near-perfect low-poly model

But the part I'm confused about is that ZRemesher came out 4 years ago and other programs are still lacking any similar built-in features. I did find an open-source alternative called "Instant Meshes" which seems to work okay, but not as good as ZRemesher.

Why is this feature still unique to ZBrush after 4 years? Is it a company secret nobody has been able to replicate, or is it not as important of a feature as I think it is? It seems like it can save 10-60 minutes of tedious work for every model. I'd like to use Blender because it's free and I've already gotten familiar with it, but this makes it seem very inferior to ZBrush.

  • \$\begingroup\$ Part of the issue is that ZRemesher doesn't produce near perfect low-poly models for animation. As models they're great, and potentially with no additional work, you may also make ones great for animation some or even most of the time. But outside of ZBrush which is relatively expensive software, this kind of thing is generally done by hand. My method is to use Blender and do it by hand. Which does offer some shrinkwrapping tools. TopoGun is another great one for remeshing. In your question about whether it's worth it though, just realize that worth is dependent on your project needs. \$\endgroup\$ Jun 6 '17 at 19:24
  • \$\begingroup\$ @FullyLucid Can you elaborate on the animation problems? I've heard that before, but in my quick tests I couldn't figure out where the problems would appear. It seemed like any stretching issues were fixed with weight painting. Even shapekey animation to make the eyes widen seemed to work fine without edge loops (Zremesher lets you add edge loop areas though). The only problems I can think of would be if you were going true low-poly (e.g. Playstation 1) where the top of the shoulder ended up as one face, but with characters that are at least a few thousand polys that shouldn't happen. \$\endgroup\$
    – Tenagra
    Jun 6 '17 at 19:44
  • \$\begingroup\$ IN the cases you provide I don't think you'd have problems and I recommend you use it, it's an awesome feature. I think @Chris's answer is way too open ended to be complete as it's primarily hypothetical whereas you were asking a very specific question about ZBrush, but on the other hand I think they're probably right. It's obviously not an worthless thing to solve or ZBrush wouldn't have spent the time to do it anyway, let alone anyone else catching up. \$\endgroup\$ Jun 7 '17 at 14:23
  • \$\begingroup\$ To elaborate on the problem though, I believe you're correct that at least with a sufficiently complex model, at least several thousand polys you really should have no issues with some work put into the weight painting. When it comes down to it though there's a ton of hand placed edge loop tricks that make a huge difference come animation time, I can't provide any of them by name as I have no clue, but it essentially comes down to manipulating edgeflow by hand, and this is difficult with thousands of polygons. I'm talking about really expressive areas like chest, hand, and face creases. \$\endgroup\$ Jun 7 '17 at 14:26
  • \$\begingroup\$ Sometimes, time will keep passing and passing, and the competitors seem to never implement the features. That can happen because of a patent that the end users never knew about. I don't know if that is the case with the software at hand. \$\endgroup\$
    – Theraot
    Jan 15 '21 at 5:36

when it comes to retopo there are a few questions that need be answered first is it hard-surface or organic? second is it gonna deform or not? from what distance are you going to be looking at it ? so based on these questions your method should change. yes Zremesher is a very powerful tool BUT it sometimes makes issues that fixing them would take u longer than doing retopo by hand like creating spirals enter image description here(your model should not contain spirals if you want to animate it)

sometimes after using zremesher u see the topology and think it's good enough but when you test it u notice some issues like the image above.

lets go over some scenarios : 1- organic + deformation : if it's something like a face that require the edge loops to be well defined and the animation is really important it's better to use manual retopo as it will give you cleaner animation and it can speed up the unwrapping and weight painting as well enter image description here

the right one is done with zremesher and as you can see the loops are not quiet at the place they should be you can still use it if it's going to be in the background and not really focused on but the middle one is the best. the left one is decimated and works if you want to use the face as an statue .

2- organic without deformation : you can use decimation BUT its can sometimes make issues with normals and texturing for example an statue

3- hard-surface + deformation : you might think that hard-surface models do not deform but as we have seen before in games they do because it's really not an efficient way to rig them realistically like in the movies , for this type of model you should definitely do retopo manually because the edge placement is very important and the zremesher is not really good at placing edges where u want them , there are ways to control the edge flow BUT it sometimes take more time than doing retopo manually .

4- hard-surface without deformation : the example would be a gun or a rock , you can easily decimate it and not worry about anything. in this case decimating is even better than other methods because with same amount of polys you can get more detail. enter image description here


A general answer from the development and business side of things for why Feature X isn't available in every software program.

Hypothetically, if the feature was released 4 years ago, it may have been 2 to 4 years before that when it was started.

So, if you consider that it's been in development for 6 to 8 years, for any other company or project just to catch up to the same quality would take years and in the same time it would be improved even more. Playing catch up on unique features like this can often not provide enough return on the the cost of development.

  • \$\begingroup\$ Any reason on the downvote? \$\endgroup\$
    – CLo
    Jun 13 '17 at 3:16

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .