some Zbrush Advice

I probably should ask this on some other forum but I figured there was enough talented people that would know. so I have been playing around with Zbrush for a day, it is a very nice program and I will probably buy the Core version for $200AUD, it’s pretty good value

I am trying to reduced the face amount, the problems I have found and need some advice on the picture below is the high poly mesh that I am trying to reduce, (not the greatest rock around but not bad for a fist time). The dynaMesh is good because it makes all the faces nice but it destroys the vertices that makes up the sharp edges so when I try to crate a low poly version the sharp edges are gone and will not work once I bake the normal from High to low, I was using the Zremesher to reduce the face count, there’s a Decimate but it doesn’t seem to do anything once lower the faces with Zremsher. the Next option is ReTopology and crate each the individual vertices where I want so when I send the model into blender I can sharpen the particular edge but this is very very slow, and took me a day.

basically the end product that I am looking for is a rock model with not much texture on it and sharp edges, in the picture below I am particularly interested in keeping the lines and the shape of the lines that are pinched and the angle of the large faces of the rock, the rest I will bake out

I was wondering other ways I can uses to keep the sharp edges when I am reducing the model practically in Zbrush but another way I will also look at as well. it’s not this about this model but the method of doing it, so once I crates the cliffs and rocks I want I can iron out any problems fist

also anther question is with Zbrush core itself, I am using the full version al the moment but that will run out in 30 day or so, are there any other features that left out that I would need, I looked on there web site and tried to do some research on it and found the features and so on but I really have I idea what it means it and what they do it could be be written in French for all I know lol


Well in 4.14 Auto LOD was added so the modeled source could be what would usually be considered unreasonable and in UE4 use auto lod to decrease the count based on visual distance to the camera. In terms of numbers the objective is to not lower counts as to a set budget per object but maintaining a well balanced performance envelop as to face counts rendered to the camera in real time. So lets say your render budget is one million faces so up close your rock could be one million and as you zoom out use auto lod to decrease the count of the surrounding objects to maintain the target.

For the source best practice these days is to maintain the count necessary to maintain the fidelity of the object and then make use of feature sets in UE4 to set the output to the required resolution an/or fidelity.

In zbrush there is the decimation master.Use that on your final high poly model.First you have to hit the preprocess buton>than decimate.(its set to 20%, and you can even use 10 for lower).Using decimation master will KEEP your detail and sharp edges.(ive been decimating wood planks and bricks all night xd).
In the end you do need to make a lower poly model(retopo).For that i recommend Topogun.Awesome software.Learned it in 3 days.

FrankieV @ so let see if I can understand this and put it into a practical term in my head, I have my Rock currently it is about 300,000 tris where it still look good, unpack it, paint and so on, load it in to UE4 and let Ue4 do the rest dynamically, what are the disadvantages of this?, wouldn’t I get a extremely large file by then end of the game, and what about graphical errors? being so large and reducing down so much or have I got it all wrong with the way it reduces down?, I really have to nail the topology of the high poly mesh because once it is in UE4 I really have no control over the actual model in a physical sense because I am governed by they way it reduces down, Yes? . is there any ensamples of finish products of this system this around that I could check out?

this didn’t cross my mind at all, I had a quick look at this feature and I was going to do some research on it, and uses it once I understood it but I did think for some reason that the base models had to resemble something useable, but that’s so much I research to do list may have change somewhat.

O_and_N @ thanks I changed the value to 0.2% and I am getting much better edges, I was only pumping it down to 5% or so I really don’t know why I didn’t go lower, thanks for the Tip for retopo, looks good but I might be out of my budget, once I pay the $200 for Zbrush I will still need anther $150 for that, I already Have substance painter, from when I was using blender I did think it was going to do the Job but is was just a big frustration, so I am going to utilizes substance printer still witch was about is about $150 as well, at the moment it is still much of a hobby so time costs me nothing. Zbushs looks like it gives more for my dollar

For a simple rock you can trust your normal map to show the sharp edges, so no need to worry too much about low poly’s topology, just make something which has the similar silhouette. A subdivided cube shrinkwrapped in Blender could do it (not that I’m suggesting to use it actually). Zremesher’s result is good for this kind of assets, but I think core doesn’t have it. Decimating is also good, but usually needs some manual fixing (remove mesh errors, extra long tirangles etc), but again, normal map can fix most of the shading errors.