Perhaps this is sacrilege for some, but I'm coming to the conclusion the once powerful SubD revolution has about ended and SubDs are now only used sparingly for specific types of projects.

Here's my thinking.

SubD's became popular at a time when NURBS and topology tools were either in their infancy or too expensive for casual modelers. They quickly were able to create organic and hard surface model forms with standard poly modelers in a non-destructive way. This was good especially if you were working on dynamic figures for animating in videos where you could rig the base model and then uprez the SubD cage to a higher resolution when rendering.

Topology modelers like ZBrush, MudBox and 3D Coat came along and soon everyone was building their organic models using topology modelers, and their hard surface models using affordable NURBS modelers (like MoI, Rhino, Fusion, OnShape, etc..). It became very clear to many of us it was much easier to create a detailed hard model in a NURBS solids modeler than to try and do it with SubDs, plus there's little reason to ever have to 'rig' a hard model for deformation, and if you do need to do so, there are interesting voxel alternatives as well. For instance take this simple radiused cube with a cylinder subtracted from the corner:



This is a nightmare to create using SubDs, yet is built in only a few mouseclicks in MoI.

Back to the discussion.

As more and more realtime modeling is going on, new workflows and pipelines are created. One thing I'm finding out, even with the 20+ characters we animate for the Alamo Reality project is each of the characters needs different mesh densities at different LODs, and SubD just doesn't work well enough. We're better off using smart decimation + small tweaks from ZBrush and/or 3D Coat to best optimize our models, as literally every vertex counts. Plus, using baking tools like 3D Coat and Substance Painter, you can bake the normals of a super high density mesh onto a much lower one, giving the impression much higher detail resolution for an object.

I see using NURBS + Topology + Poly modelers is a tight pipeline, and with some of the best 'other' tools, including poly filleting, hard ops (Blender), draping fabric and at times SubD, you can get where you need to go. It's just not necessary anymore to start at SubD-- at least not for basic workflows.

If one follows Vitaly Bulgarov, MasterXeon2001, Tor Frick,/a> and other concept and detail modelers, one now finds lots of different tools used, but similar workflows. And for the most part, there's not much SubD stuff there. Sometimes for a quick organic surface while in poly mode, but usually it's FROZEN quickly in order to continue to edit it using standard Boolean tools.

Of course each to his own, but when deadlines, budgets and production realities looming, I'm not seeing SubDs as part of the big future-- at least for AR/VR/MR/XR or any of the recent moves toward realtime effects.

I'm interested in what other think about this?

Hi Chipp, well what's "niche" for one person or industry is pretty different than others. For example people doing waste piping design for nuclear power plants also use 3D graphics yet they have no use for any of the revolutions that are familiar to you... But yes the thing is that different technology bases tend to have a different set of strengths and weaknesses for different kinds of models. Managing a high level of detail like you'd see on realistic character models is where sub-d can get more difficult and where the sculpting/brush-based displacement type approach (like ZBrush, 3D Coat, Mudbox) just manages that type of thing more easily. You can do it in sub-d too but it seems to require a higher level of skill and planning and so a longer investment of time to develop that.

Shapes that are kind of organic in form but don't have small localized bumpy details like car bodies for example seem to be a natural fit for sub-d still though, the super localized detail that brush stroke displacement gives isn't useful for that kind of thing when you don't want any little bumps.

Then if your shape can be defined well by 2D profile curves, that's where NURBS modeling is strong, a lot of man made manufactured parts fit into this category, it's why manufacturing oriented use of 3D graphics has always been based on NURBS and didn't make any use of sub-d during the visual fx revolution era either.

But anyway the best toolset varies an awful lot depending on what specific industry you're in and even what particular kinds of models you happen to be doing as well. As far as I know there isn't any one single tool that spans everything.

- Michael

Hi Chipp,

years ago Professional MoI user Andrei Samardac started a similar thread here and now became
the co-owner of a new subd modeling software and stopped using MoI, so it seems. The Silo guys
just recently concentrated to work more on Silo again. If there would be such a trend that subd
became more outdated i think companies like Autodesk, Maxon or SideFX would focus more on
their included NURBS tool sets. Don't forget Foundry with MODO, where people can see in their
very big Gallery that folks are doing all kind of stuff with MODO.

Best regards
Stefan

Hi Chipp!

This is an intriguing theme.

I think we could spend tons of words and almost all could be right.

All in all, is think that this sentence from Michael could recap the whole situation :

"As far as I know there isn't any one single tool that spans everything."

This is the brutal truth :)

Rather, I think that the biggest and, for the moment, UNRESOLVED problem is this :

HOW, WHEN AND AT WHAT COSTS ALL THESE DIFFERENT MODELLING TOOLS/TECHNOLOGIES CAN BE INTEGRATED SEAMLESSLY INTO THE SAME "PIPELINE".

I think that this is the crux of the matter.

For example, let's say you have to model something.
You can do it using any of the current modelling tools (subd, poly, nurbs, voxels, 2.5D) based on the kind of the thing you want to do.
OK. This is not a problem at all.
After doing this, let's say you now need to render the model, for example.
Ok. You will get almost no problem with this task. Simply export the model to a software capable to create and manage UV and 3D painting.
Export to a renderer and the deed is done.

Now let's say that you created your model using a 3D Nurbs tools, like Moi or others, and now you want to simply add a new fillet or change a blended area.
Ok. Simple...right ? ... it's not that simple :(
You have to go back to the 3D Nurbs tools, modify the original model, export back to the UV and 3D painting software and this,
most of the time, implies that you have to redo the UVs and the 3D painting.

But this is not still a real big issue.

Let's say now that finally you have the 3D model painted and ready for rendering, and 5 minutes before doing that,
your customer ask you to add some "organic" details to the model or your have simply to do some other "boolean" operation into the 3D Nurbs side.

What a nightmare! Many times you have to start all over again!
Export to UV, redo UV, redo 3DPainting.

And this is only a very simple use case.

Very often, doing a very little modification to the original model, implies that you have to do a not a so "little" modification in all the rest of your "pipeline".

And if, on the same model, you have to apply different techniques then everything becomes a NIGHTMARE!!

And if you have complex scenes with many models each using a different approach the thing becomes very very complex...

I think that the main problem is still how different tools can be used together into the same "pipeline" without wasting time and money!



What do you think ?

Thanks for sharing.

Ciao!

Hi Marco,

just wondering... have you ever worked with clients?

Clearly communicating with clients and approving the design steps until the product is finished
should be IMHO mandatory if you want to avoid such thing you have mentioned, especially if
you use toolsets not having parametric input options, where you can do changes on the fly, in
every step you do.

I would let your clients approve every design phase (sketching, modeling, texturing and rendering)
to avoid surprises. Otherwise start all over again and make it clear to your clients that they will be
charged accordingly.

Best regards
Stefan

Hi Stefan.

@You : "...just wondering... have you ever worked with clients?..."

Oh yes! It's happened many times.

My main job is not 3D creation but software developing, but during the years many times I happened to do some 3D/2D creation
for some of my customer of desktop/web/Android apps.

You're right. A clear communication with clients before doing the job it's mandatory. Anything to say about that :)

But what I wanted to say is that even if you take every "commercial preventative measure" with the client,
this not erase the "technical" problem. The crux of the matter.

That is, integrating all the different approaches into the same pipeline.

This is what I wanted to point out with my previous post. Not the trade matters with the clients.

Thanks.

Ciao!

Interesting topic and discussion!

When I started to shift from 2D to 3D work around 1998, my tool of choice was the mainly polygon-based 3ds Max (it has some NURBS tools, but they're limited). I clearly remember how I longed for a modeling method that would be able to combine primitive shapes and seamlessly blend them together. The problem with subdivision modeling for me personally is that you're carefully planning and constructing in stead of sculpting. Constructing doesn't allow for spontaneous creation, with "happy accidents" as the late Bob Ross used to call them.

In those days, the late 1990s, there was a cool plug-in for 3ds Max called Clay Studio Pro. It was a metaball system, but you had many different shapes to work with, including 'Clay Splines', which allowed for a lot of modeling freedom, and everything was animatable as well. Clay Studio Pro was ahead of its time, kind of a very basic 3D Coat avant la lettre. But the problem was that PCs didn't have enough processing power yet to turn Clay Studio Pro into an all-round solution. To achieve smooth meta-surfaces you had to increase the polygons higher than my system could handle.

When I discovered MoI many years later, it was the first program I encountered that combined the accuracy of constructing with the spontaneity of fast and easy creation. And for organic modeling, ZBrush and — for me only recently — 3D Coat turned out to be fabulous tools.

The bottom line is that I really love not having to worry about polygon structure planning, keeping everything quadrangular, and having to keep an eye on edge flow everywhere. So my best friends are MoI, 3D Coat and ZBrush.

Having said this, I do love the polygon-based Blender, but mainly because it fills every gap that's left by MoI, 3D Coat and ZBrush, such as animation and high-level rendering (although I also use Keyshot for rendering).

In the field of sculpting freedom, the only area that's still a bit on the weak side currently is automatic retopology. I dislike processes that require tedious manual work, such as manual retopology, manual UV mapping and manual rigging and weighting. I hope AI / machine learning will improve those areas soon.

So, the bottom line is that I think subdivision is not yet outdated, but it is becoming slightly dated as powerful alternative modeling methods are on the rise.

All the best,

Hello,

In my opinion the whole SubD thing was an ultimately poor solution to an existing problem. Tools to perform easy control point modeling and continuity operations are lacking so this was an attempt to address that issue. I do think there are some interesting developments on the horizon but the first toolset to make those operations highly usable will win.

I’ve been using Nurbs for illustrations maquettes for 10 years now.
It’s ridiculously faster than subdivisions, but of course polygonal software is useful when needed.

Being focused on user interaction and simplicity, Moi makes a striking difference in the landscape.

I guess it also comes down to preferences, my brain must be curve based.

Marc

https://community.foundry.com/discuss/topic/136219/art-with-ice-cream

Sorry couldn't resist. :-)

Regards
Stefan

I think we all know this is not only SubDs... And could easily be done in 3D Coat or ZBrush-- perhaps even more easily.

I see a big philosophic difference :)

When you use Nurbs you know exactly the form that you want!
When you use SubD you don't know exactly the form that you will obtain! :)
it's always a surprise (good or bad :)

There is an interesting thread about that on SketchUcation! (34 pages growing of examples of what I say! :)
https://sketchucation.com/forums/viewtopic.php?f=397&t=63826

"You need to be a member to view images." :( I can't see any of the posted screenshots / renderings in the subd thread.

Ahhhh yes!
Sorry you must free register for see images who are directly published on their forum ! (not external linked)!

:)

Here's a 154-page alternative forum thread about subdivision modeling:

http://polycount.com/discussion/56014/how-the-f-do-i-model-this-reply-for-help-with-specific-shapes-post-attempt-before-asking

Hi Pilou,

//
When you use Nurbs you know exactly the form that you want!
When you use SubD you don't know exactly the form that you will obtain! :)
it's always a surprise (good or bad :)
//

The problem is that most of us need to use SubD in a predictable way, where there is no room for surprise.
And the only way to archive this control is with a lot of practice (and planing) ... creating edge loop in every direction, fighting with surface distortion etc.
It's not a very funny process, at least not any more after a couple of years.

In the other hand SubDs is the only modeling technique that can tackle any subject, from organic to mechanic, and that's probably the reason its far to be 'dead' yet.

Fortunately we can now convert SudDs cage model into a Nurbs surface, and somehow take advantage of the best of those 2 modeling method.

Still much easier to UV map a sub'd mesh than a somewhat messy mesh generated from Nurbs Surfaces.
Also you have to be really careful to keep your normals data intact before rendering a nurbs generated mesh.

With Sub'ds you can model anything you think of. Modern retopology tools are overcoming many of sub'ds shortfalls. Its still time consuming though.

Nurbs are really at their best for hard surface objects.

I don't know if it will ever happen but a program that has the speed & power of nurbs and the tweakability of sub'ds would be really cool. Modo's meshfusion is an attempt at this but still generates a messy mesh.

Yes, you are correct about UVs. Still, have you tried out 3DC's UV tools for NURBS polys? The auto mapper is pretty darn good and can often suffice with a single button click. I use it all the time. Here's a quick video of some of the easy to use features.

Ah, some kind of flashback :)
Couple of years ago I started topic like this and there was huge battle:)
I just want to say what I learned from that time, guys do not waste your time posting on forums to debate with each other, just do your job, it will be much more effective to learn things:)
There are thousands of designers that use NURBS, there are thousands of designers that use Polygons, there are thousands of designers that use Sculpting, everybody uses the instruments they need, because there is no any universal tool!!!

Just to be clear, I'm not advocating one over the other for everyone. It's just my observation that there is a shift in focus among many of the top concept design modelers that I know and follow.

Of course everyone should use whatever tool they like.

That said, paying attention to shifts in trends makes sense as the technology in the space moves so quickly.