ZBrush Hard-Surface Dynamesh
polycounter lvl 13
Hi peeps,
I want to adapt my workflow between poly modelling and ZBrush so that I may create final renditions of complex shapes quickly from a base mesh.
But here's the thing, after watching a video someone posted on here:
[ame="
http://www.youtube.com/watch?v=MVHPSnBlIrU"]zSummit 2014: The Division | Montr
I want to adapt my workflow between poly modelling and ZBrush so that I may create final renditions of complex shapes quickly from a base mesh.
But here's the thing, after watching a video someone posted on here:
[ame="
http://www.youtube.com/watch?v=MVHPSnBlIrU"]zSummit 2014: The Division | Montr
Replies
In regards to the Dynamesh, I only made the main mesh Dynameshed and kept the subtractive tools as default from importing. Do I need to convert it to PolyMesh3D or do I need to collapse the sub div modifier on the parts I want to merge before appending or importing into the scene?
But if you're in ZBrush, yeah, basically iteration means "do that piece over again".
my counterpoint is that any workflow where throwing away work is part of the process is not efficient, or safe for production.
i find editing precise mechanical models is a lot easier in modeling programs, which are designed for it. most of my highpoly models nowadays are just a primitive and a modifier stack... tough to beat that for ease of iteration
Its also quite common that sometimes the person adjusting the model isn't the person who created it to begin with. This would create a lot of problems down the road if the model ever needed to be updated quickly, by someone else.
my thoughts exactly. everytime i attempt something complex in zbrush, I realize more and more how much I rely and depend on a 3d application. i've found it to be great for experimentation and quick mock ups, but when working in a production setting, the 3d app, for me, always wins out due to flexibility.
In all seriousness, I would put MODOs edge weighting up against ZBrush for fast iteration any time. Iteration time in MODO is going to be many times faster.
I like the idea that they are approaching this as a series of machining tasks. That concept breaks down on any objects that weren't originally deisgned to be fabricated on a mill. Also, watching him do simple tasks with smoothing and polish brushes that rely on zbrush pivots and radial symmetry is a little painful. I think part of why all of this works for redstorm is that they are building real world objects that were mostly drawn in cad programs with similar operations sitting on 90 degree viewing planes. As soon as you have anything based on concept art or your have an AD doing redesigns and paint overs it would become a nightmare.
Right, that doesn't sound like he's doing hard surface in ZBrush.
I wish there was a bit more control in zbrush or I am just missing something in the workflow.
I think the main advantage here as-well, in my opinion at-least, is that it's possibly to create complex high poly models with consistent edge widths for baking. You're also incorporating the detail process into the whole workflow, so you're also essentially eliminating the need for cross-processing the model between applications.
Still, my question that I have yet to hear on is how the guy in the video manages to dynamesh that M1 frame from so many tools when, for me at-least, I cannot dynamesh my piston beyond two times before it becomes un-useable. I want to know more on his workflow or how it's possible to dynamesh the model without it messing up.
Honestly, I find it quite freeing - I can jamb a bunch of shapes together, and not worry too much about edge flow, pinching around holes etc.
I have found that for dynamesh sub to work correctly you want to reset all polygroups (make all objects just one polygroup) - seems to give more predictable results.
And remember, if I exclude my trial and error, then this asset took approximately 20 minutes to make - in my books at-least, that's fast!
Files if you want to go along with the video and see the results yourself...
https://dl.dropboxusercontent.com/u/7741113/Piston_V2.obj
http://www.zbrushcentral.com/showthread.php?177194-DynaMesh-Master!-unofficial-Information-Installation
Also, it would be pretty cool to see some people's own attempts? Maybe we could get a lil Dynamesh thread going for hard surface stuffz.
Post up if you manage to get anything out of this!
This is simply an extension to my workflow, I'm not looking to fully replace my hard surface routine - however, it has to be admitted that it's an incredibly useful solution to adding complex detail to a final model. No, for sure I add edge loops anyway, however what I was trying to explain is that to receive consistent results I'd advise making sure that no matter how simple a shape may be, edge loop it and make sure it can be subdivided while still retaining its original shape.
@Popeye9
Make sure the normals are correctly calculated on all of the tools and then set the layer to subtract. Also, only dynamesh the mesh you're going to subtract from and re-calculate after merging.
Use the file I uploaded to see whether it works for you...