I just watched the new
ZClassroom tutorial demonstrating more or less the Boolean feature using Dynamesh. He was able to get this shape extremely shortly using this operation where as modeling it in Max would've taken way too long for him let alone me.
I've been told to avoid using Booleans as it simply destroys geo, and that I should be able to model the shape without relying on subtracting primitives. For me, this tutorial threw all that in a blender because modeling this prop in Max seems like a bloody nightmare where as ZBrush made it a 20 min operation.
Take for instance this gun prop I'm making
I can model this in Max easily, but why not just DynaBoolean the thing in ZBrush, decimate the high poly or throw it into Topogun and get the low poly that way?
Replies
Each one has a different workflow, the point is you need to learn each workflow, and make do with it at the end of the day.
You're told to avoid bool operations in traditional apps because you want controllable topology right off the bat. It would be very silly to model something in Max, bool and retopo it when from the start you have the ability of creating an organized mesh.
ZB on the other hand doesn't have this option, just like it doesn't have the option of selecting vertices and extruding manually or creating holes manually, so organized mesh isn't possible right off the bat, hence why you can model stuff with bools, because you have to.
Look at it this way for an idea, the time it takes you model something cleanly in Max = time for ZB user to Model and retopo said model.
Ace has a point that you will have to retopo the ZBrush high poly model but if you build the high poly with an eye towards that process, it shouldn't be too bad.
I'm finding the trade off worth it so far because I can get so much more detail out of the ZBrush booleans than I can out of traditional modeling techniques. That's a personal thing, obviously, but since I'm not that great at sub-d modeling to begin with it works for me!
The native 3dsmax Booleans you are referencing are called "ProBooleans" not Power. I think that is a 3rd Part plugin. Sorry to jump in here and correct you but I just didn't want to confuse the OP.
But regarless, ProBooleans are awesome. People laugh at me when I say that I use Booleans but they are the dumb dumbs that dont know ProBooleans exist.
I think it's look good so far. I just don't know what to do with the low poly on the right. This is the most complicated thing I've done up to this point. Any advice would be great :poly122:
Also, do you often use the boolean method you posted above to start those kinds of models?
Do you do a lot of manual stitching and reflowing afterward and/or is that that just the raw result from proboolean?
Perna, for the parts I've circled what did you do to get the smooth merge?
Wouldn't that extra geo make the boolean and the resulting clean-up a lot harder/more-work?
You also couldn't do this nondestructively or non-linearly, like if you realize later that you need to add that geo in to make the extrusion, so you would have to redo a lot of the work in that case, no?
edit: to answer computron on the destructivity question, proboolean is only partially destructive, you can still modify the operands and their modifier stacks.
Groboto+Modo does something like that, but it maintains great topology and SDS endgeflow. The only problem is the topo is uniformly dense and that can make somethings a little difficult to hand model afterward.
How do you modify the parameters of the boolean members later?
Like how do select and increase only the cylinders edge segements?
I can't find the option in my proboolean:
EDIT: I have to extract the sub-objects, don't I?
Facepalm! :poly127:
I agree about that, it seemed to me like you where discounting zbrush wholesale for hard surface like a lot of people who tried using it before all the new tools in the later releases of version 4. I am curious if you have used it lately as a lot of websites have been posting tutorials and models made mostly with Zbrush in recent memory.
[ame="http://www.youtube.com/watch?v=q4FoEYptQvA"]GroBoto Real-Time Real-Boolean Modeling.mov - YouTube[/ame]
How come I have never heard of this!! :poly121:
Definitely post your tut when you get it up, hopefully with some really complex, somewhat more organic shapes.
As an aside, how would you go about cleaning up the merging on this shape:
My guess is that a more realistic scenario would involve less segments on the big cylinder or more on the smaller one, but any tips in this case?
Here is my attempt, although its got tris and its not really smooth:
I'll be watching for that tut.
Thanks man!
Also, That last thing is pretty cool lookin', can you post the primitives and the wireframe? How much time did it take to clean?
But I understand, I was going for essentially the same thing with my attempt.
It's hard to gauge how much geo is enough, I guess thats the type of thing you hone in on with experience.
And thanks for the Groboto mention. I'll take a look at that today, I've never heard of it before..
[ame="http://www.youtube.com/watch?v=XtCD9-OsogA&feature=player_embedded"]fRay Boolean Master 1.0 - MEL Script Demo - YouTube[/ame]
Real-time booleans in Maya
Under "advanced options" you can tick on "make quatliterals" which makes it possible to add some support edges easily.
I think for personal art this is a good policy, but for studio production it falls apart.
Really it should be "that any artist can make any changes requested in a reasonable amount of time" because quite often someone else will work on an asset you made, either because you're working on something else, you've left the studio, someone else has left the studio and you're tweaking their art, etc etc etc.
So if you're the only guy in the studio that is comfortable doing hard surface stuff in zbrush it really wouldn't fly. Unless you literally never have any revisions or never have more than one artist work on the same asset.
Traditional sub-d methods can be viewed/edited in virtually every standard 3d app, and should be easy for any artist to pick up on even if you just send them an obj of your cage. Zbrush content is an entirely different story, and doing certain things like a quick tweak of some edges will be time consuming and difficult, whereas with traditional sub-d its only a matter of sliding some edge loops around.
Another common situation:
Lets say you have an asset, and that asset has 20 identical screws. Lets say you go to do your bake and find out the the edges are too hard on every one of those screws and it doesn't read in the bake.
In your 3d app, you tweak the one source floater you have instanced around your mesh and rebake, takes only a couple of seconds.
In zbrush you.....
This is true but mainly because most guys who do that are married to doing everything in ZBrush. To be fair, building shapes in a modeling app and sending them over to ZBrush makes that process a LOT faster. Trying to shadowbox or clip the correct boolean sources is a waste of time, IMO.
On another topic, I have a question about your process...
You advocated booleans earlier in the thread, which I agree with. There's really no reason not to use them. You also said that by generating a high poly mesh you essentially get the low poly for free by removing edges, which I also agree with.
However, do you find that still holds true if you combine both of those thoughts? The boolean'd meshes look nice but they don't seem to have any kind of nice polygon layout that would allow for quick creation of the low poly mesh. Or do they? Granted, I'm not all that experienced so that's why I'm curious...
I think people are just impressed that you can make stuff like that in Zbrush because Zbrush wasn't made for such things (which is really funny logic. Impresses because it does things it doesn't need to do). Sculpting app that can do hardsurface seems impressive but only as a concept. Also people think they can just make hard surface just like organic stuff. "Topology sucks ? Just divide it and it will look ok". When I started learning 3d I created heads in Zbrush. They needed 2 milion of polies to look more or less ok. Now by using topology right and knowing how to use geometry I can do the same thing with lots less polies.
I really like technical aspect of subd modelling. I have still much to learn but it's a great subject and I really don't understand people who try to avoid it.
Booleans are indeed awesome
The damn cleanup still takes forever, I couldn't figure out how you boolean'ed in support loops as you mentioned, Perna, so definitely elaborate on that in your guide.
Did you try using Intersection with Imprint checked? That just cuts edges into the operand above it.
Sound advice.
Also awesome info in here perna, looking forward to that tut!
I gotta give this stuff more love, I need to find a girl to cover in hard surface bits that might motivate me...
Yes please.
As a result, the radio in the first post is pretty easy to do in a few minutes, just model and boolean one corner of the radio shape and then mirror the rest once you are done cleaning the instanced corner.
And yeah, people who say avoid boolean are just plain ignorant. I admit, use it at your precaution (keep edges/vertex cleaner), so theres a cleanup process, but the benefits awarded with using proboolean are huge.
I had to make 40 highres weapons for a AAA game, and boolean was and still is my most used tool.
This is the exact reason I steered clear of Zbrush until I had a firm grasp of topology/subD modeling. A good modeler can use Zbrush and take their creations to their modeling package of choice and use them. It doesn't work the other way around.
Although this is also partly the reason I avoid Zbrush a lot in favor of just subD Modelling things that would be more sensible and faster in a sculpting package. Swings and round abouts, I do love subD modelling though. It really is incredible fun
I have Polycount set up on my RSS reader, so I get every new thread loaded in quickly glanceable chain and if I see a thread I like but I don't have something to post I set it up in my subs and check on those threads later with the sub box. super effecient and I never miss any cool art.
Now if only I could get the WAYWO and Lowpoly threads posts in my RSS reader the same way...
I am learning to use modo and I am really missing the modifier stack and the 'edit poly, show-results, turbosmooth' features and haveing proboolean at the base of the stack as well.