Choosing a 3D Package
So you’ve decided you want to create digital 3D models.
You might be an aspiring movie special effects guru, industrial designer, architect, or mechanical engineer. Maybe you’re wanting to create photo-realistic images of product design ideas, or maybe you want to use 3D Computer Aided Design (CAD) to control a Computer Numeric Controlled (CNC) milling machine to bring your ideas into physical reality. You might be hoping to play with rapid prototyping, mechanical simulation, or 3D animation for art, illustration, or demonstration.
Knowing the right 3D software package(s) will not make you a good designer, a competent engineer, or a film effects creative genius. It will simply be a tool in your belt, there when you need it, making it possible for you to communicate three-dimensional ideas quickly and accurately for virtually any purpose you can imagine.
But when you look for a specific program into which you’ll invest lots of time and money, the number of options can be daunting. How do you decide which 3D program is right for you?
There is no magic bullet in the 3D software world; every piece of software that exists on the market has a niche for a specific type of user with a specific combination of needs. In this set of posts, I’ll try to make the pros and cons of all sorts of common 3D applications easily understandable for the newbie, and hopefully even clear up a few things for the seasoned professional.
In this, the first post in our series, we’ll be looking at the broader landscape of 3D applications, and discussing the basic categories of 3D applications.
Decision factor #1: Why 3D?
The way I see it, there are four basic types of 3D tools:
- Modeling (defining geometry)
- Analysis & Simulation (processing geometry)
- Animation & Visualization (viewing geometry on screen)
- Machining & Rapid Prototyping (physical implementation of geometry)
Engineers and research scientists find themselves needing analysis & simulation, where as designers, artists, and marketing professionals benefit more from animation & visualization. Anyone involved in the creation of real-world deliverables–such as sculpture, product designs, furniture, machinery, or architecture–will need machining and/or rapid prototyping technology in order to bring the digital 3D geometry into physical 3D space.
The one thing all of these disciplines have in common is that they need a way of defining 3D geometry in space.
A computer program that allows a user to define 3D geometry is known as a “modeling package”.
Since everyone who uses 3D needs a modeling package, this series of articles will focus first and foremost on modeling. Once you’ve chosen the right modeling package, everything else will fall into place.
Note: The 3D representations of geometry created in these packages are known as “models”, or “databases”. The two terms are interchangeable. In industry slang, you will often hear people say things like “send over your database, and I’ll take a look.” This may seem cryptic, but it makes perfect sense: the files generated by 3D applications are really just databases full of numbers and definitions. Different programs define 3D objects in different ways, but ultimately every 3D file is just a big list, or “database”.
In industry parlance, there are three basic types of modeling packages (*in reality there are only two, but for now we’ll stick with the conventional paradigms). Within each type of modeler, there are sub-categories as well.
- Polygonal Surface Modeling (“poly modeler”)
- dense poly mesh
- subdivision surface modeling (aka “Sub-D”)
- NURBS Surface Modeling (“NURBS”)
- direct
- historical
- parametric
- Solid Modeling
- direct
- historical
- parametric
Each of these three types of modeling packages does essentially the same thing–defining surfaces in 3D space–but each goes about it in a completely different way. The package that’s right for you depends entirely on what you plan to do with the resulting data. Here’s a basic overview:
2D Deliverables: Design for Screen or Print
Common Applications (alphabetical): 3D Studio Max, Blender, Cinema 4D, Lightwave, Maya, Modo, XSI/SoftImage
Polygonal modeling is ideal for any application in which 3D geometry is a means to get to a 2D result such as: 3D animated film special effects; 3D animated feature films; on-screen graphics for video games or interactive virtual applications; or photo-realistic still images or illustrations to be viewed on the web or in print.
Every surface generated in a polygonal model is comprised of hundreds of individual single-sided “polygons” floating in space, glued together at the seems. Building polygonal surface models is like stretching an imaginary rubber sheet over the desired form. It is extremely well-adapted to defining free-flowing organic objects, but difficult to control with numerical accuracy.
Note: It is theoretically possible to do CNC machining and/or rapid prototyping from polygonal model data, but it is exceptionally rare in actual practice. The reason for this is less because polygonal modeling itself is in any way badly suited to it, but more because the tools and techniques employed by most poly modeling packages are not intended for the level of accuracy typically desired when creating prototypes.
3D Deliverables: Design and Sculpture
Common Applications (alphabetical): Alias Design, Rhinoceros 3D (“Rhino”), ICEM Surf
‘NURBS’ ( and/or ‘Bezier’ ) modeling is a highly accurate means of defining curvilinear surfaces in space, and provides a degree of control and mathematical accuracy that is difficult to achieve with polygonal modeling. As a result, surface modeling is frequently used by industrial designers, automotive and aerospace design and engineering, and artistic sculpture.
When working with surface-modeling, every piece of geometry is a “surface” floating in space. Surface modeling can be compared to building a model out of sheets of paper: every piece of geometry is comprised of a number of ”trimmed” sheets of paper, glued together at the seems. This is a very direct approach to geometry creation which can result in a very high degree of accuracy and precision.
Note: Because of the level of manual control afforded by direct surface-modeling, it is often the tool of choice for full-time professional “class-a” surfacing engineers. In automotive design parlance, a class-a surface is any aesthetically important surface on the exterior of a car (i.e. any surface that the user can see or touch). Because cars are comprised of large, reflective surfaces, it is very important that the CAD geometry be absolutely flawless, and specialized CAD professionals called class-a surfacers know how to define this kind of flawless geometry. It may be important to note that designers do not typically create this kind of geometry themselves.
Polygonal Vs. NURBS, a practical explanation: Polygonal models are similar to pixel-based 2D images in that they are resolution-dependent. With a 2D image, the more pixels you have, the more accurate your image. If you print a low-resolution 2D image at a large scale, the individual pixels become obviously visible. Similarly, with a polygonal model, the more polygons you have, the more accurate your model. If you print a rapid prototype of a low-resolution polygon model, the individual polygons become obviously visible.
NURBS models are similar to vector-based 2D images in that they are resolution-independent. With a 2D vector-based image, no matter how big you print an image, the shapes will always remain perfectly crisp. Similarly, with a NURBS-based surface model, no matter how big you print a rapid prototype, it will always be perfectly smooth.
NURBS has other advantages over polygonal surface modeling that are more difficult to explain here, but NURBS modeling is fundamentally better than polygonal modeling for any type of geometry that requires numerical accuracy. For example, if I want to put a 16mm hole in a surface with a 3mm fillet on its edge, NURBS is the only way to go. Polygonal models can theoretically describe numerically accurate models, but are in practice much more difficult to control with numerical precision.
3D Deliverables: Structural, Mechanical, and Accurate
Common Applications (alphabetical): AutoCAD, CATIA, Inventor, ProEngineer, SolidEdge, SolidWorks, SpaceClaim, UnigraphicsNX
So-called “solid” modeling packages are ideally suited for the creation of geometry intended for manufacture or engineering analysis. In a surface modeling package, a “cube” is thought of as six individual planes intersecting with each other. By contrast, in a solid modeling package a cube can be thought of as a cube, or as a “thick” rectangle. Solid modeling packages work fundamentally the same way that surface modeling packages do, except that they use different types of tools and techniques to define geometry.
Because solid modelers use the same mathematical means of describing 3D geometry used in NURBS surface modeling packages, the two are often used interchangeably, or even in concert with one-another. Often a design team will work with a surface-modeling package, then pass his surface geometry to an engineer who will then use the geometry to create geometry in her solid-modeling package.
Note: Most solid modeling packages include many surface-modeling tools, so increasing numbers of designers and class-a surfacers are switching to solid modeling applications. I am one such designer: I do almost all of my surface-modeling from within so-called solid-modeling applications. Confusing, huh?
Recap & Conclusion
In the end, here’s what you need to know: if you want to make art, movies, video games, or illustration for screen or print, you’ll want to look into polygonal modeling packages. If you’re an aspiring automotive or consumer product designer, consider learning a NURBS-based surface modeling package. If you’re needing to create functioning mechanical prototypes and/or engineer products that really work, you’ll want to look at a solid modeling package.
As an industrial designer, I have found that it is very (very) helpful for any junior-level designer to be familiar with the different types of modeling and their relative advantages and disadvantages, and to be proficient with at least one NURBS surface modeling package, and at least one solid modeling package. As for which of each to choose, we’ll have to save that for a future post.
In the next post I’ll be looking into the various pros and cons of many different 3D software packages, categorized by industry and usage.
24 comments
Trackbacks/Pingbacks
- Computer Related Articles » Blog Archive » Why You Should Consider Buying the GBC CombBind C340 Binding Machine - [...] adam » Blog Archive » Choosing a 3D Package [...]
- 3D Paket Auswahl « - [...] Adam Theoherns gibt potentiellen Anwender in seinem Blog eine kleine Entscheidungshilfe zur Verwendung verschiedener 3D Modeling ...
Excellently put. Great overview.
Though I do wish to add that solid modeling, under the hood, is a little different from surface modelers because of the added boundary representation. But it’s still very similar for the most part.
Adam found your website via 3d World. More than anything else I would like to say that you come across as a genuinely nice guy. Love your blog/website. Very informative so keep up the good work!
I have a question for you if you could take the time to answer please. I am an ID guy who graduated when 3d software were just coming in. So just by a year I missed the boat with regards to Alias training at university.
Anyway I worked as a concept artist at ID studios and later on started doing more CG/Entertainment industry kind of stuff and picked up Maya.
Fast forward to present time and I am kind of not here nor there. I am not a full fledged CG/Entertainment industry guy and neither a full fledged ID guy. Sometimes I do ID concept and visualization projects other times CG/Entertainment related projects. I basically like doing concept/CG/Entertainment work more that pure hardcore ID work.
To cut a long story short I am thinking of adding another software to my Maya seat to widen my horiuons. Was looking at Rhino as Alias seems too expensive for freelancer. Your post that people are moving from Rhino to Solidworks type of programs due to surfacing capabilities getting added got me thinking and slightly confused. So I ask….
- If visualizing creative concepts with “maybe” some possibility of passing on the data for manufacturing was on my mind then what would you suggest? Rhino, Alias or Solidworks? I feel Solidworks might be too out there for my needs. What do you think?
- Also any opinion on Maya Vs Modo as far as modeling and visualization side of hard surface concepts is concerned? Is it really worth adding Modo in the mix or am I basically well covered with Maya?
Thanks for your time and would really appreciate your advice.
Hi ID_CG_guy:
Thanks for the kind words. I enjoy what I do :)
All of your questions are valid, and I’ll do my best to give straightforward advice based on my own experience. You’d probably get 10 different answers from 10 different people, but here’s my $.02:
- Learning another package is ALWAYS a good thing. No matter what it is, learning new software gives you new points of view, and new ways of thinking about old problems. So bravo!
- If you know Maya pretty well, there’s no point in learning another sub-d package. There are tons of them out there, each with its own niche, but Maya is up there with the best of them as far as I know. I’m not a full-time sub-d guy, but I’d say if you already know Maya, stick with it.
- If you want to learn NURBS modeling, download the demo version of Rhino. It’s free, and it does everything the full version does except save. So you can learn everything there is to know about Rhino without purchasing anything at all. Once you’ve gotten the hang of it, you can decide if it’s worth it to you to buy a seat. Also note that as of today, the commercial Rhino license is $900, but the student/faculty version is $200, and there are *no licensing restrictions* on the student version, meaning it is perfectly legal to use a student license for pro work. So if you’re teaching anywhere, you could pick up a license that way.
SolidWorks (or ProE, or any of the others) is an amazing tool, and depending on the type of work you are doing, it can save you a ton of time and headache. But based on your particular situation, I might recommend giving Rhino a go first. You can’t beat the price, and if nothing else it’s a great learning tool.
Best of luck to you! Keep in touch.
Adam
Thanks for the post Adam :-) Will continue to follow your site/blog.
Hi Adam,
Just read through this post twice and watch the video. My current 3D “needs” definitely place me in the Polygonal category.
Concept art/design of consumer products, and some dabbling with the human form at comic book type character level just for fun.
One downside is that this category seems to have the most contenders. But it also has some very affordable contenders.
So in the short term I have no need for mathematical accuracy, nor (mass) production. Maybe in the future, because I have plans for computer cases. But that is long term. Although I also said that less than a year ago about concept art/design.
I wish there was an all-round 3D modeler which offered all 3 categories of modeling.
BTW, whatever happened to Shade outside of Japan ?
I can only find the Japanese homepage : http://shade.e-frontier.co.jp/ and thew Wikipedia Shade page in Japanese.
Strata and Shade were the first programs which make me even think of getting into 3D.
My only advice would be to avoid Blender. It’s a perfectly capable package, but the interface is so unorthodox that it’s extremely difficult to learn, and will probably just frustrate you. Go for one of the major packages: Modo, C4D, 3DS Max, or Maya. Any of these will do the trick, and all are very common, and very well documented.
Thanks for the heads up on Blender. That definitely doesn’t fit in with the smooth ride into 3D I was looking for.
Alas 3DS Max and Maya are out of my budget.
So it’s going to be a toss-up between Modo and C4D then.
I might add Rhino later on when I decide to seriously start with the computer cases.
Rhino has the glorious advantage of being cheap! Keep in mind, however, that Rhino does NOT have a built-in rendering engine (well… it does, but it sucks). So if you want to render things from Rhino, you’ll need to either buy a plugin like Flamingo or V-Ray or export your Rhino geometry to another application like C4D or Modo.
Re: “I wish there was an all-round 3D modeler which offered all three categories of modeling.”
There kind of is… sort of. Catia V6 has all three types of modeling available within a single application, but that comes at the cost of $30-100 *grand*, a really extreme learning curve, and even then has really limited animation and texture-mapping capabilities.
Your statement is kind of like saying “I really wish there were an all-round vehicle that was great for hauling freight like a tractor-trailer, but also a fantastic race car for street racing, and got really excellent gas mileage as well.” Even if a company wanted to make a single app that did everything–and there’s really no reason to believe that they would–it would be so riddled with contradictory goals and features that it would probably be completely useless.
Hi Adam,
It’s me again. I went through your Rhino tuts and they were very helpful. I am close to diving fully into Rhino. Are you beta testing Rhino5? Any views on it?
Also do you have any personal views on Rhino Vs Alias Design? How is the wind blowing these days with regards to Alias 2011?
Thanks
I haven’t used Rhino 5 yet, nor do I know much about Alias 2011. I don’t really use either program in my own studio, so I don’t keep up with them as closely as I might.
Rhino has the advantage and disadvantage of being totally dumb. Everything you do in Rhino is completely manual, and any changes you want to make to your model will basically involve a re-build in one way or another. The upside to this is that it’s a very “direct” approach, and everything works (or doesn’t) very consistently, and in a very easy-to-understand way. The learning curve is very shallow, and forces the user to learn the ugly under-belly of down-and-dirty surfacing technique. Rhino is a brutish hack-saw of a tool that will let you do anything you want to, but will allow you to make a total mess of your model if you’re not careful.
Alias is a tremendously powerful tool in certain ways (e.g. great surfacing control and historical associativity), and a shockingly limited tool in others (i.e. total lack of dimensional parameters and solid operations). Its learning curve is much steeper than Rhino, but it makes up for it with power and agility. It is not as useful of a “swiss army knife” as Rhino, but many studios will have both packages around for this reason.
In my experience, Alias Design has fallen by the way-side in the wake of the rapidly-improving surfacing capabilities of engineering packages. Do a survey of Alias users, and you’ll find that the demographic is heavily weighted to designers between the ages of 35-40, who learned CAD back when Alias was the only surfacing player in town. Younger designers are tending toward combinations of tools like Rhino and SolidWorks, and unless something changes drastically, I expect the trend to continue.
As of today, in my studio I use a combination of SolidWorks, Rhino, and Modo. I do the vast majority of my modeling in SolidWorks, some minor hacking in Rhino, and all rendering and animation takes place in Modo.
I was going to add Rhino later, and getting Modo or C4D first. So the rendering would be covered.
Ehrm, Catia will be able to sell their software to others. :)
I get the analogy your making, but it was just a wish.
And yes it would quickly become a messy app. Didn’t think of that.
I should clarify that Catia doesn’t have true Sub-D modeling, but a (better) technology very similar to “T-Splines” for Rhino. You can get a somewhat disjointed approximation of the Catia workflow by using T-Splines, Rhino, and SolidWorks together as though they were a single app. For improved “direct modeling”, throw a seat of SpaceClaim into the mix. Even with all four of those tools, you’re still only looking at a total of under $8k in software, where a seat of Catia will easily run you five to ten times as much.
Wasn’t aware of T-Splines. Thanks for that one.
The selection process is getting complicated again.
Rhino + T-Splines plug-in + V-Ray looks pretty tempting.
Rhino + T-Splines plug-in + C4D looks pretty tempting.
Rhino + T-Splines plug-in + Modo looks pretty tempting.
Both offer NURBS modeling + Subdivision modeling + HQ-rendering. The last two offer subdivision modeling twice. :)
For now I’ll get evaluations of Modo and C4D as per your first recommendation. And adding Rhino + T-Splines plug-in later on. It will give me both types of modelers.
T-Splines and true Sub-D are very different. It would be really cool to have both. If your primary goal is product design, go with Rhino/T-Splines first. Then later buy a sub-d modeler for your human projects, since those are “just for fun”.
Concept Art/Concept Design is what I want for now. And that is mostly not tangible. So I thought no accuracy was required.
But you’ve got a good point there. Maybe it is better to have the Concept Art/Concept Design accurate from the start, to skip having to completely remodel in NURBS.
While I was trying to find out a bit more about compatibility of T-Splines with Parallels Desktop, which I could not find at their site and forum, I read that the T-Splines modeler makes sure that the models are “water tight”, and also exactly converts T-Splines to NURBS. But, to me, that would suggest that during modeling with T-Splines one has to already put in that accuracy.
Which normally isn’t needed for a concept design.
I hope that T-Splines works without issues on Parallels on a MacBook Pro. Otherwise I’ll have to hangon to this WinBox for a while.
I’m shooting tsplines an email later on. (I’ll probably forfeit all support.)
Rhino mentions BootCamp and Parallels specifically in the system requirements and that “sometime after the release of version 5 for Windows” they will add support for OS X (http://www.irhino3d.com/).
This newsgroup thread seems to confirm Rhino’s compatibility on Parallels Desktop : http://news2.mcneel.com/scripts/dnewsweb.exe?utag=&group=rhino&xrelated=337845&cmd_related=View+thread
I told you that I found this 3D stuff to be complicated.
Maybe I’ll better dive in the 3D World tutorial you did, and get things going. Even if the modeling goes well, I’ll probably mess up with the lighting, or rendering. Good old Illustrator.
Haha, you’re really over-thinking this, Robert. If you’re making super-organic stuff like humans or clothing, you need Sub-D. If you’re making hard-body stuff like cars or products, NURBS is better. T-Splines is a happy middle-ground that kind of straddles the two. Just try out all three and see what you think suites your workflow best.
Rhino and T-Splines both work fine on Parallels, and I do run the Mac version of Rhino in beta (and it’s fine). Modo is cross-platform, and if you buy a license, it works on both Mac and Windows.
Look at what you’re saying “If you’re making hard-body stuff like cars or products, NURBS is better.”
But you’ve setup the tutorial around Modo, a Sub-D modeler.
Or maybe it was a requirement by 3D World Mag?
3D is complicated. :)
BTW, I couldn’t get issue 127 of 3D World anymore yesterday. It was still in the shops last week though. (The Netherlands is getting the mag almost a month later.) Held it in my hands. Should have bought it then.
I’m trying to get a new pen for my Wacom. Because the warranty exchange is going to take 2-3 weeks.
I’m starting your 3D tutorial on Saturday. And post a link to my result.
I watch the first couple of videos with the rims, and I almost chickened out again. :)
I’ll stop thinking now.
Like you say “Enough jibber-jabbin’. Let’s make stuff.”
Jibber-jabbin’ from my side of course.
Hi Adam,
Just letting you know that I’m postponing my start of your 3D World tutorial. I need to learn the Modo interface first.
Just watching the videos is one thing, but actually doing what you are (saying) that you are doing is “different”. I need to know where “everything” is first, and also some of the shortcuts.
I’m making the same mistake now, that I made when I started with Adobe apps. Then I had to go back, and first learn the toolbox of each App I wanted to use.
Rotating the backdrop items like you do, doesn’t work for me for instance. And I can only see one backdrop at a time, not the cross section view.
Also the viewport and menus are quite dark in reality. (I know, whinge, whinge). I cannot find where to change the colors of the Modo menus, only managed to lighten up the colors of the viewport to an easier on the eye grey. But then the contrast with the darker menus turned out to be even worse.
OK, I’ll stop now before I monopolise this post again with off-topic stuff. I have a lot to learn.
Hello Adam,
I have one question about Rhino/Modo. Say you model a detailed phone or any product in Rhino then you want to render it in Modo or say Maya. Maya can read Nurbs so not that big of an issue but how does the transfer work for Modo?
Do you convert it into poly or something inside Rhino first? Or does Modo read it directly and then automatically convert it?
Does the process mess up the original design intent? I mean does the phone still visually look 100% same inside Modo as the Nurbs model was inside Rhino? Do the blends, fillet, chamfer etc still look visually same/accurate?
Hi ID_CG_guy:
All renderers use a polygon mesh at render time, it’s just a question of when the meshing takes place. Maya can read in NURBS, but ultimately will be “meshing” the NURBS surfs for render. Modo needs basically the same thing, but it can’t do the meshing itself.
Think about this: in order to display anything 3D on your screen, your CAD package is using OpenGL or DirectX, both of which use exclusively triangular meshes. OGL knows nothing of NURBS; all it knows is triangles. So if you’re looking at something 3D on your computer, you’re looking at a triangular mesh, whether you’re aware of it or not.
So you are correct: to render a Rhino model in Modo, you’ll want to create meshes first by using the “mesh” command. Once meshes are created in your 3DM file, Modo will read the native 3DM file. This is actually a huge pipeline advantage for Rhino > Modo, especially if you’re using T-Splines for Rhino.
In answer to your question about accuracy, the meshed model is as accurate as you tell it to be. You want to create a mesh that is light enough to be useable, but dense enough to come across as visually smooth. I have settings that I use, but everyone has their own threshold for that kind of thing.
Ultimately the answer is yes: the phone you build in Rhino will look 100% the same in Modo as it does in Rhino… usually better.
Adam
Which program would be the best for map modeling (for games)?
Hi Deividas:
I’m afraid your question falls outside my expertise. I know that all of the major polygonal packages are used in the gaming industry, and my (limited!) understanding is that modo and SoftImage are particularly well adapted to the task.
Best of luck!
Adam
Ok,
Thanks for your response.