Feature articles

Introduction

[specs-box]

Forest Pack Pro is a scattering plugin for 3ds Max developed by Itoo Software. It's designed to create environments made out of vegetation, from a little grass area to a forest with thousands of trees, and to mainly distribute, instance, and randomly modify certain objects' properties in a scene.

The plugin can be used with Mental Ray and V-Ray, proxy objects of both render engines are supported. You can also use the native instancing system of 3ds Max and the scanline renderer.

Let's see how the software performs.

Initial impressions and workflow

The interface is clean and efficient, small images are used to help quickly identify the various features available, and most parameters are similar to those used in 3ds Max and other 3d packages.

Forest Pack Pro can manage two kinds of objects: templates and custom objects. Templates are planes with textures, a kind of billboard, useful mostly for backgrounds trees. The second type, custom objects (i.e. custom meshes selected by the user), are what we will focus on this review. When you load an object Forest Pack Pro assigns it a unique ColorID (which you can choose to use or not), this is useful for dealing later with objects' distribution.

One way to define placement of objects in Forest Pack Pro is by drawing some splines in the viewport and using them to define areas for including or excluding placement of objects.

You can easily attach multiple splines in order to achieve a complex area shape. A falloff setting is available to help make your zone's limits look more natural. Areas are easy to setup, one just needs to pay attention to the direction of splines.

This feature could be improved: if you'd like to use a map instead of splines to define the area where meshes will be scattered, this could be done through the Distribution panel, but you'd lose the flexibility it offers to switch from groups, to scattered or dense maps in a click. So you would have to paint all the features in Photoshop. The possibility of using complementary, separate maps for areas and distribution would be a welcome addition.

Distribution is the feature that will help you scatter objects. There are many ways to achieve a desired result. Forest Pack Pro allows you to choose from a small but efficient library of bitmaps (that offer dense, very scattered or grouped distribution). You can also use your own map and some renderer-specific ones (I'm thinking of V-Ray's VrayDistanceTex, for example). After choosing from the Forest Pack Pro map library or your own let's see how this works.

A Density panel allows you to define the size of your map and the density in the distribution of objects. It works the same way as a simple UVMapping of your ground. The size of the map can be seen in your viewport.

Diversity and clusters are surely some of the most useful features in Forest Pack Pro. With them you can easily group objects that were imported using the Geometry panel (which were assigned a specific ColorID). For example if you want to create some autumn grass, some parts of the surface are going to be covered by short grass, some by tall plants and some parts will be left empty.

With Forest Pack Pro you don't need to use multiple objects to create this effect, using diversity and clusters makes it easy. In order to make things look natural, it's possible to apply roughness, blur and noise settings to the cluster map. All these changes can be seen in real-time in the viewport, a feature that is not available in some other scattering plugins.

Finally, a collisions feature will help you reduce the number of objects in your scene without losing surface covering. This is one of the basic features of a scattering plugin and certainly the best way to optimize render time and RAM usage.

Using the Surface panel, objects can be aligned to the surface's normals. Another interesting option is the altitude parameter. This is really efficient, it lets you limit object placement to a certain altitude range with just a few clicks. In my test scene I used this option. All four Forest Pack Pro objects are linked to the same surface, trees are only placed at the top, grass at the bottom, and in the middle there are some little bushes.

Adding variation

In order to help make things look natural and realistic, Forest Pack Pro offers the ability to randomize the scale, rotation and translation of each object. This is one of the most basic features expected from scattering plugins, the only thing that gives Forest Pack Pro a little edge is the mirroring option. If you want to, you can also use a map for determining scaling and rotation.

Variation can also be added through materials. There are two ways to do this: a simple gradient tool that'll modify existing maps, or the awesome Forest Map. This is a really great option, with this map you can mix a bitmap and a gradient (or a color), but the best part is that you can choose to do it per leaf, per tree, or both! With this option each leaf and tree will be different. You can also load multiple bitmaps and distribute them randomly.

Forest Pack Pro can handle animated objects and it allows adding a time offset to the animation in order to keep things non-linear.

The Tree Editor

When you use random distribution in a scene, you'll probably get something that looks good, but if you have a master plan to stick to, it'll need some tweaking. The Tree Editor is there to give you that flexibility. It allows you to deal with each object, so if you need to focus on some trees at a specific place you can. It will let you scale, rotate, delete and move one object, or groups of scattered objects.

This is a really nice option because it gives you a simple way to handle a single object with a few clicks and it doesn't kill the RAM optimization provided by Forest Pack Pro. Distribution settings won't be available after using this feature but viewport settings still will. Using the points cloud display in combination with this is a real pleasure.

Camera and Display options

The Camera panel offers two important features, one for planes and one for custom objects. Forest Pack Pro can rotate planes automatically, so that they are always facing the camera.

The second feature is "Limit to visibility". With it, trees are only created where the camera can see. There are also some clipping plane and density falloff options, these are very useful for reducing RAM usage and keeping render times as low as possible without losing density of scattered meshes.

A small problem with this function is that during a fly-through trees can disappear at the border of the frame. The pivot point is for Forest Pack Pro the point of interest of each mesh and when you're working with big trees you will need to use the expand parameter to resolve this. But it won’t always work, depending on the size of the object and the pivot point. I'd love to see this handled automatically by the plugin.

Camera options also offer the possibility to cast fake shadows from 2d objects.

The Display panel lets you choose the way you want to see scattered objects in the viewport. The points cloud display is a really nice way to see objects while working with clusters or with the tree editor. You can also limit the number of objects visible in the viewport, and use simple geometry replacements (plane, box, pyramid, etc.).

Something that could be improved regarding display options is adding a feature currently offered by Multiscatter (another widely used scattering plugin) that allows you to toggle scatter objects on/off through the utilities panel in 3ds Max, or select them without having to switch viewports.

Handling RAM issues

Forest Pack Pro allows you to administer RAM usage through its statistics option (General panel). This will help you resolve related issues that may surface. If there is a problem with a render, it will pop-up showing how many mesh instances Forest Pack Pro is managing, and it will also provide a brief description of how to resolve the issue.

Forest Pack Pro Library

The Library Browser is a tool to quickly add plants to your scene, and you can of course save your own models to the library. The latest release of Forest Pack Pro brings a number of free 3d models courtesy of Konstantin Kim's 3dmentor.ru. I've used these models for most of the scene I created for testing the plugin, I just added some trees from another collection. At Konstantin's site, there are also some free grass and flowers available.

The test scene

Now let's see how all the features discussed helped me create the test scene:

- There are two grass types. Modeling looks the same but rendering does not. I've used the Forest Color Map with a gradient but with solid colors inside.
- There are 6 kinds of flowers, these are actually three colors combined with two kinds of flowers. I've made this to show how clusters can be useful to group objects without having to use multiple Forest objects.
- For trees, I've used five tree models. Those all have the same Forest Color Map for leaves. I've used it to obtain a little variation on each tree. I really love this feature, it's fully integrated to my workflow with Slate. You can also use it to load multiple bitmaps for the diffuse channel, it looks like CGSource's MultiTexture map, a must use feature of Forest Pack Pro!

A word about licensing

There's something important to mention regarding licensing: you can use the same Forest Pack Pro license both at your job and at home. It's not written in the license of the plugin, but Itoo is flexible about this and authorizes users to activate licenses more than one time. Just make sure not to use them at the same time.

Conclusion

You've probably guessed by now that I've really come to like Forest Pack Pro. RAM management works very well, tools are easy to use and some of the features (Forest Map, camera visibility, clusters, altitude range, etc.) really help with the scattering process and allow you to give a very natural look to your scenes.

Thanks to its accessible interface, the plugin is very easy to adopt for CG studios. Of course, artists will have to dedicate some time to read the help file and watch some video tutorials in order to learn all the little tips and get the most out of the plugin.

Forest Pack Pro is one of the easiest and most powerful plugins available for creating realistic vegetation environments. It is useful not only for architectural visualization, but also for cinematics, commercials and other kinds of CG production that require the creation of complex environments.

For those who'd like to give it a try, a free Lite version can be downloaded from Itoo Software's website. You won't get all the features but you'll still get to test much of its functionality.

Introduction

[specs-box]

3D-Coat is a voxel sculpting software developed by Ukraine-based Pilgway, that aims to make it possible to sculpt without any topological constraints, while providing at the same time a comprehensive toolset for UV mapping, painting textures and retopologization.

The software supports CUDA acceleration provided by NVIDIA GPU cards. It can utilize 3DConnexion's space navigator controller and pressure sensitive tablets like Wacom. 3D-Coat is being developed as cross platform software for 32-bit and 64-bit Windows, Apple OSX and Linux, and currently supports 8 different languages, which might be a plus for non-English speakers.

The version reviewed is the third major installment of the software, with features that bring it closer to a bigger professional audience, steering away from the niche software corner.

Note: among the images accompanying this article, only the models that explicitly state so were created 100% with 3D-Coat. All models were made by Olli and Sami Sorjonen.

Initial Impressions and Workflow

When starting 3D-Coat the user is greeted with a gray user interface and a menu with various options, which helps a lot when taking your first steps in the application, as some of the panels and tools are full of options and their purpose may not be immediately obvious.

There are different workspace modes, called "rooms", for approaching the different tasks: Paint, Sculpt, Retopo, UV, Voxels and Render. The software is basically split into two logical parts: Paint, Sculpt and UV "rooms" (for working with meshes), and Voxels and Retopo "rooms" (for working with hires voxel models). In the case of the Retopo room, voxel models show up there as a reference mesh, which you can use as a model to build new geometry on. The Render room, finally, shows both meshes and voxel models.

The different workspace modes can be accessed via tabs located on the top part of the 3D-Coat window. Pressing a mode tab changes the UI with one mouse click, in similar fashion to Maya or Softimage.

Initially it's hard to figure out how one should do certain things. For example, if you wanted to import a hi-res mesh and start retopologizing it, it's not immediately obvious how to do it. When you click buttons to access different modes, the model shows up in certain rooms and not others depending on its type (voxel model or low-res polygonal mesh).

When you get over these UI oddities, however, the interface enables you to handle different tasks quickly.

User Interface

Overall the feel of 3D-Coat is pretty snappy and fast, all selections, viewport navigation, etc., are quick and responsive. Most of the UI is customizable, layouts can be saved and restored. Mouse and keyboard navigation can also be customized, in similar manner to Mudbox and ZBrush, so that you can adjust the viewport navigation and other buttons to mimic those of other apps you are using.

Presets exist for default views and there are tools to quickly frame the model in the viewport, in case you've lost focus on it. Light intensity and direction, as well as ambient light can be adjusted through viewport buttons. There's also a 2D overlay grid available for checking proportions. Another UI feature worth mentioning is the possibility of assigning shortcuts to camera views.

Introduction

[specs-box]

What is Frost other than cold crispy ground covering? Simply put it is Thinkbox Software's latest point meshing plugin for Max. Once an in-house production tool, it has been seriously updated, refined, and released to the world.

Frost meshes points or particles, and so does Max's BlobMesh compound object (particles anyway). So, the question is inevitable: why would you want another particle mesher? Because if you are serious about meshing particles, you do.

First off, Frost meshes particles in *many* different ways, and it does this fast and easy. Second, Frost is a very versatile tool. Third, it has the ability to accept virtually every one of Max's particle systems, mesh objects, spherical gizmos and various particle file formats via Thinkbox's Krakatoa PRT Loader. Fourth and certainly not least, it interfaces insanely well with Thinkbox's Krakatoa Particle Suite.

Unlike any of the other particle mesh plugins currently available for Max, Frost is the only one that is 92% multi-threaded (only the Custom Geometry Mesh Type is currently single-threaded). This means speed, and it scales really well, the more cores the faster the meshing. The faster the meshing the more iterations, and the more iterations the less time it takes you to reach your desired result.

Interface

At first glance seeing the interface is a bit overwhelming since there are so many more options than any other meta mesher. Considering all of its options, the user interface is laid out very well. Since it's a Max object it is tied to the Command Panel, with all of the rollouts open it can consume nearly full three columns.

The developers have done a good job on minimizing clutter by adding menus and submenus where appropriate, as well as good parameter grouping, making relevant changes easy without too much unneeded scrolling or even leaving the Frost interface.

Most float/value parameters have a customizable menu in which you can add/remove custom values. If you are familiar and comfortable with Krakatoa these are a welcome addition. Another Krakatoa carry-over is the Preset system. Both of the options combined make it quick to get decent base setup. You like it, you can Preset it.

Features

With Frost you can surface most anything that contains a vertex or particle. Your source options are geometry objects, all of Max's native particle systems including Particle Flow, Thinking Particles, any Krakatoa PRT object including custom streams via PRT Source object, Spherical gizmos and the list goes on to include Splines, NURBS surfaces, and of course other Frost objects.

Instead of using particle objects you have the option to simply use particle files. These file types include Thinkbox PRT, Realflow .bin, or even a Comma Separated Value (.csv) file full of point data. This rollout has most of the familiar base options of the PRT Loader. You can hold frames, re-time, frame offset, custom ranges, etc.

Frost has several different meshing types, currently five plus Vertex Cloud. Each Meshing type has its own parameter-specific rollout offering more refined control of the surfacing algorithm. You have Geometry, Union of Spheres, Metaballs, Zhu/Bridson, Anisotropic, and Vertex Cloud modes. All with their own unique attributes.

The Geometry meshing type has its own subset of shapes and controls for animation timing and orientation. The shapes include plane, sprite, tetrahedron, box, sphere, and custom

I have only used animation timing in Custom Geometry mode. This allows you to set whether your animation cycle for each Frost geometry instance begins at frame 0 or current time. The handy part is the Offsets which allow you to set a custom start time with the GeomTime channel.

Orientation provides some great options. You can align to picked nodes, such as a camera or other object, match a target object's orientation, grab it from a channel, vector, or specify it in the interface. Divergence options are also supplied for orientation variation. Materials can be straightforward or complex, you can drop a map on Frost itself or use any data channel options to color the mesh.

As with other mesh type plugins the Meshing Quality rollout, quoting Borislav "Bobo" Petrov, provides both Relative and Absolute mesh resolution controls - the Relative adapts to the largest particle size, the Absolute uses a constant mesh resolution in world units regardless of the particle radii.

This inherently solves issues related to random scale variation during an animated sequence. Both resolution controls have access to Vertex Refinement, an accuracy parameter that snugs up the mesh closer to its underlying particle.

By far one of the best features for any Thinkbox product, however, is the support. If something is broken, in most cases, it will be fixed or at the very least you’re offered a workable alternative solution. This has been my experience with Krakatoa support up to today, and I’m sure it will be no different with Frost.

Frost and Krakatoa

It is easy to see why Frost loves Krakatoa. With this combination you have the whole PRT/KCM workflow at your disposal. This opens a whole range of possibilities and it even works with the evaluation Version of Krakatoa - so you are truly not limited by lack of funding. All of those particle renders now have the option of being turned into a "hard" surface. You now have the ability to mesh FumeFX simulations directly with no intermediary such as PRT Loader or Particle Flow.

Krakatoa, as of version 1.6, has the new PRT Source. This allows import of PRT-compliant data streams from any software as long as someone builds the bridge. This PRT Source object is also accepted by Frost.

PRT Volumes give you the option to turn any geometry object that is compatible with a PRT Volume into particles. You can shell this object and grab appropriate data such as mapping, generate velocities, or manipulate it in many other ways, then add and pipe the data into Frost.

Conclusion

Surprisingly enough (or not), Frost was more like testing a 2.0 release than a beta or even a 1.0, it is, in this reviewer's opinion, a solid plugin. If you do find a repro crash (i.e. reproducible crash), support is right on it. The few minor ones that I found during beta were generally fixed within a day or two, sometimes faster. That is a feature within itself.

It meshes, it's quick, it's efficient memory-wise, it works with just about everything including Realflow .bins. I think they met nearly all the criteria you could want in a meshing plugin. All in all it's a very robust tool, especially at its price point.

It's going to be very interesting to see what people come up with using this plugin when it starts getting pushed to its limits. The only ceilings I have bumped against have been because of my hardware. It's also good to mention that I could still generate a nice dense mesh on my laptop (dual core 3.0 ghz, 4 gb ram) So you don't really need a monster machine to get a great result - but it doesn't hurt either!

By Dave Baker, Neil Blevins, Pablo Hadis
and Scott Kirvan

Foreword

In our aim to analyze the evolution of the CG Industry we've had the great pleasure to talk to David Stinnett, co-founder of the legendary Blur Studio. Blur is a privileged place to ask our questions: it's one of the longest running studios (as its "Blur.com" URL address attests), a central player in the CG business, consistently producing top-quality work, and a first-hand witness of the changes in the industry. Last but not least, the studio's distinctive no-nonsense attitude means we get real answers, from real people - no filtered replies, no marketing trash, no noise on the line.

For Max Underground this article is, in the year of its 10th anniversary, a way of going back to its roots. Early news items posted at MU point to Blur's website: this is because MU was conceived while its founder was working at Blur Studio. Neil Blevins, who is currently working at Pixar, has spent many years working at Blur, too. This translates into a Blur interview done by Blur insiders, you'll notice that the replies offer particular insights into the inner workings and culture of the studio. Asking questions about Blur, however, is asking questions about the industry itself.

Blur turned 15 years old last April, it was founded in 1995 by David Stinnett, Tim Miller and Duane Powell who wanted a place to work at without the bothersome daily interferences and bureaucracies of a management-oriented mindset. A place for artists, created by artists. This was possible in part due to the rise of desktop computers and 3D desktop software, and the history of Blur and its founders is tied to the history of 3D Studio and 3DS Max.

Without further ado we present you an interview with David Stinnett on the past, present and future of the CG Industry. Enjoy!

The Interview

Tell us a little about your makeup effect days. You worked on Tremors, and puppeteered the Crypt Keeper from Tales Of The Crypt, correct?

Yea, I started in '88 doing makeup effects and animatronics (years earlier if you count personal projects and student films). After doing misc stuff on some ultra low budget films and tv shows, I landed my first big gig on Tremors. After that I worked for various MFX shops working on the 2nd and 3rd Chucky films, the Bill and Ted sequel, Tales from the Crypt, Terminator 2, Honey I shrunk the kids sequel (yes... did lots of sequels), and various other films. On Tales from the Crypt, it took 6 people to operate the Crypt keeper, and I was one of those for a few seasons. I also did animatronics and prosthetic makeup on the show for a couple seasons.

What made you decide to pursue digital effects instead of continuing to do practical effects?

I had always had a fascination with CG, starting I guess with Tron, but I also remember getting all the "Mind's eye" Laserdiscs and just thinking they looked cool. I got my first computer, a gateway PC around '91, but that just started as a game machine. I then remember seeing Autodesk Animator Pro, a 2d animation package, and thinking "I can do that myself? Awesome!". Not long after that I was at a local computer fair (with table after table of hardware and software for sale), and saw a demo reel playing for 3D Studio v2, and I remember my jaw dropping. I tracked down a local reseller (yes, ALL my software was purchased legally... There was no "internet" per say back then... just dial in bulletin boards, and Compuserve)... anyway, I bought 3D Studio V1 with a free upgrade to V2 when it came out. I then spent my nights messing around with that. I guess there had been 3d software on the Amiga before 3DS, but I didn't know that until long after starting with 3DS... If I had known earlier, I would have likely gone the Amiga route... and interestingly enough, had I done that, I wouldn't have got my job at Sony, and would not have been part of Blur (I'm sure Blur would still be here, it just would have started a little differently).

In some ways, learning digital effects must have felt like starting over again from scratch, after already being established in a career path. Did the learning curve worry you?

It was certainly starting from scratch, but it didn't worry me. I learned the Makeup FX on my own, through trial and error, and many years and late nights in my garage. With computers, it was the same. I just spent lots of time and late nights learning it by doing. This included the hardware as well as software. Come to think of it, I guess a lot of those late nights were spent playing Doom, but that's still working on the computer, right?

Talk a little bit about your time at Sony. Was it your first digital gig after doing make up effects?

I was active on the Compuserve 3DS section, and when I posted the first pics I was comfortable posting, I got a call from Gary Yost, the Developer of 3DS. He was amazed that I was able to get near photo real images of organic subjects from his software. That led to me getting on his beta testing team for 3DS 3, 4, and the first plugins, IPAS (remember those?). In late '94, if I have my timeline correct, while still doing physical FX, I got a call from Sony, saying that Gary Yost recommended me for a job there. I went in for an interview, and got the job. I started when my FX gig finished a couple weeks later. I had to take a pay cut, but since I was starting over, that was fine.

Do you remember the first day you met Tim Miller?

The exact day? No, but he seemed like one of the only guys there who was really into it. We would geek out about the latest software, plugins, etc. It wasn't just a job to him, he really loved doing it, which was something we had in common.

Blur is more known for doing amazing game cinematics and working on fully-animated shorts/features, but the studio produces visual effects for feature films too. You tend to supervise many of these film jobs - you recently completed several pick up shots for Avatar. What draws you to the film projects?

I've been drawn to VFX since I was a kid. Star Wars was the start of that (that's the 1977 Star Wars, kids, not the new ones :-) I started out doing models and miniatures in my garage, then transitioned to Makeup FX. But I've always had a love of VFX. Unfortunately the nature of VFX for a shop of our size means its very difficult to make money, or break even, doing the typical vfx that we bid (tends to be the low budget stuff). The game cinematics aren't big money makers either but we do ok, and it keeps us going. The budgets are better for large format ride type stuff and commercials. However, The game cinematics are really more related to what we ultimately want to be doing... our own CG features. They are short stories, all CG, and we generally get a lot of creative freedom with the concept, writing, editing, animation, sound and music. They are really just small movies. However, I still like the VFX work when we get it. On Avatar we did all the space shots, and would have done more on it if we had the people, but we were packed at the time and had to turn down some stuff they asked us to do.

How many projects are generally in production at Blur at the same time?

Obviously it varies, but it's not unusual to have 10 projects of various sizes and in various stages in house at one time.

What size was Blur at its smallest and largest? What major changes had to happen to allow for such growth? How has the atmosphere of the company changed / stayed the same due to these changes?

We started with the 3 people and 4 computers in '95, and within a few months we moved into a bigger building and hired a small handful of guys. My memory is a little fuzzy on this, but I think after a year or so we were at 12 or so people, and it stayed there for a couple more years. For a while we grew to 36 or so and stayed there for a while. We've been just shy of a hundred for a few years now. When we started out the guys we hired were going to be able to do everything... animation, modeling, lighting, comping, and we were going to pay everyone the same salary... simple, easy to manage, everyone knew what everyone else made, etc. As we got bigger and the projects more complex, we realized our system wasn't going to work. As much as we wanted to avoid separate "departments", we found we had to hire people who were really good at doing certain things, and the 'jack of all trades' way of thinking slowly fell to the wayside. With it went "one salary for all", not necessarily because of the particular job but as we had people who were new (talented, but green), those who had been with us for a while, and those coming in with much more experience than others, the pay had to reflect that.

Also, the management of the artists became much more time consuming. As much as we wanted to keep internal politics and personalities from the daily workings of the studio, the fact is that artists are a fickle bunch, and that stuff did have an effect on how we ran the studio (it took much more time to keep everyone happy).

Two nice benefits at Blur are the job stability (you rarely hire a ton of people for a project and then lay them all off afterwards), and you provide decent health benefits to everyone in the studio. Why did Blur decide that these two benefits are really important when a lot of other companies don't offer them?

We hire based on talent, and once here, we want to keep that talent. Working in production is not easy, even when you love what you are doing, it's hard work. So we try our best to treat our guys (and gals) well, via health benefits, cool work, etc.

What is your view of the current state of the CG field? Do you feel it is as exciting as it was 10-20 years ago, or has innovation basically slowed down these past few years?

The 90's were definitely a big time for CG development. I remember there was new stuff, new programs, plugins, new ways of doing things coming out all the time. When a new version of the software came out (3ds/Max especially), we would all gather around, giddy with excitement playing with the cool new features (yes, we were geeks). Nowadays there is not as much to drool over in new releases, as far as brand new innovative things. That is not to say things have gone stagnant. Advances in rendering, image based lighting, hardware acceleration are all exciting, especially as the hardware gets faster and faster.

How have the needs and requirements of clients evolved since the first years of Blur (and what changes has Blur made in order to meet them)?

Simple... each year they want it to look better than anything else, and they want more of it... for the same price as what they paid last year. Unfortunately, that is the way it is. As the software, hardware, and our abilities get better and faster, we can put out work that looks better and better. That doesn't mean the client is willing or even able to pay more for it. Hopefully the fact that we can work faster year to year means we can pump better quality work through our pipeline in the same amount of time as we did previously. But the bottom line is that we need to push the envelope, not only for the clients but for ourselves. That is why they come back year after year.

What are the most interesting developments that you see coming up for the future of CG? (in the next 5-10 years)

It looks like hardware rendering may finally be taking off, which should be interesting. Being able to get quicker feedback without the need to constantly be doing test renders should help to speed things up, especially in scene assembly, but also in modeling/texturing.

How has the profession of CG artists evolved from the early days of 3D animation?

I can only answer this from the perspective of Blur, but as stated earlier we tended to be more all around people, now we have dedicated modelers, animators, riggers, cloth/hair people, etc. The one thing that we still have not split up is lighting/rendering/compositing. We simply refer to those steps as "Scene Assembly", and our scene assemblers take all the models, animation, etc., in raw form, put them together into a shot and see it through to final. We also, overall, have much better realtime (or close to it) feedback of what we are doing, due to the advances in hardware and software.

Any features developed for Max that were a direct result of Blur's feedback?

I'm sure there were, I can't remember any features specifically, but way back we had a very close relationship with Gary Yost and the Yost group (the original developers of 3ds and Max), and were very active in the development of the software. Unfortunately, I can no longer say we are, as with schedules the way they are we just can't risk using alpha or early beta software in production (at least not the key programs like Max). We submit our wishlists as always, but we are not as intimately involved as we were in the past.

Blur recently turned 15 years old, and got there producing top-notch art and enjoying worldwide recognition. What pending debts do you feel the studio has?

We turned 15 in April. I'll just have to come to terms with the fact that we likely have a good portion of our guys who were only between 5 and 10 years old when we started the studio ;-) Our main goal of Blur has always been to do our own features. While we have done many other projects, a lot that we are very proud of, our own movies are still the end goal. It's been a long time coming, but as we have learned, it's not a quick process to get one to the point of having the money in pocket to actually start production. We are however, closer than we have ever been. We are in development of a few feature projects, including "The Goon" but the one we are closest on is "Heavy Metal", which we are developing in house with David Fincher. There are a number of talented directors who have expressed a strong interest in supporting the film and directing segments.

What are your hobbies/activities? What do you enjoy doing when you're not working?

Building models... real ones, not CG :-). I was into racing karts for years, but it's been a while since we've been out... nothing like racing a kart capable of 100mph on the edge with your butt 1 inch off the ground... literally. Lately my spare time is taken up by RC Helicopters. I admit it. I like my toys... what can I say...

If you could move Blur from Los Angeles to any city in the world, what would it be? Or is LA the place to be?

I'm sure there are plenty of cities I'd like better than LA, but for doing what we do, LA is really the best place to be.

What's the story with the Whaleboy name?

That is really Duane Powell's fault (one of Blur's founding partners). That was his nickname for me. Why? The first CG image I ever did (that I publicly released) was of a whale, done in 3D Studio DOS v2 (or was it 3?). I did a 30 second animation of it soon after, and it made it on the 3dstudio reel. The whale image is what caught the eye of Gary Yost, and got me on the 3DS beta group. Gary got me in touch with Sony, which led to Blur... all from that little CG whale.

What is the current favorite video game played at the studio?

Believe it or not, it's still Quake 3. We still have our daily 45 min game, although I'm ashamed to admit it's been ages since I've played. We've tried other games, and a small group of us did Battlefield for a while, but the simplistic fun of Q3 has lasted the test of time. But I do have very fond memories of working (well, kinda...) until 3am playing Doom2, Descent, and of course Duke Nuk'em. Ah, those were the days...

Many thanks to David Stinnett for sharing with us his experiences from years of working at Blur, and his views on the evolution of the CG industry.

Introduction

On this occasion we have the pleasure to present you with a view of the evolution, current state and future of the CG industry from the perspective of an expert in the camera tracking and visual recognition field. Russ Andersson, founder of Andersson Technologies LLC and developer of the SynthEyes camera tracking (matchmoving) software, shares his insights from years of working in the CG industry. Enjoy!

The Interview

What first attracted you to writing graphics software, as opposed to another field of computing?

My background originally is in robotics and computer vision for robotics. I segued into some video game work that presaged the Wii & Natal that appeared at EPCOT, and from there into camera tracking work. Robotics has been stuck a bit because the mechanisms haven’t kept up with advances in computing; the film/tv post industry is able to push ahead to new technologies without being hampered the same way. I do miss the visceral impact of bugs in my robotics software, though: crunching metal and flying pieces of plastic.

What features in camera tracking software development have been demanded by the CG industry?

The most recent has been stereo, and with the commercial success of Avatar, I expect to see even more stereo projects. SynthEyes has been used for stereo projects for some years now by advanced users, even though there weren’t specific stereo features in it. I’d originally anticipated adding stereo features about two years ago, but when I asked around, it seemed that there wasn’t a strong readiness at that time. A year later, when I went ahead and added specialized stereo capabilities, those features went instantly from my lab into "secret projects" worldwide, which turned out to be Avatar and perhaps some things we haven’t seen yet. The industry was ready and we’ve seen the audience was ready.

What changes do you see camera tracking has brought to the film/VFX production process?

I think it has allowed directors to plow ahead to reach their artistic vision during shooting, with less and less regard for the details of how that vision will ultimately be achieved in post-production. That can be a full-employment plan for post-production, but the added flexibility while shooting offers the opportunity to achieve a better end product.

What changes have you seen in camera-tracked shots through the years, how have they evolved? Which shots or uses of camera tracking surprised you?

The rise of "reality" and mock-documentary shooting styles, such as for Battlestar Galactica, have meant that there are plenty of shots that are quite wild, not just the more traditional crane and dolly shots. There are more challenging long shots, and of course the resolution requirements have gone up.

The most unusual uses I’ve seen have been for image analysis of Antarctic ice boring images (tracking air bubbles in the ice), and for the analysis of colonoscopy images. So it’s the same idea in much different locales!

What is your view of the current state of the CG field? Do you feel it is as exciting as it was 10-20 years ago, or has innovation basically slowed down these past few years?

It’s only at the simplest level that you might think innovation is slowing. Sure, it’s not like when you saw smooth shading or antialiasing for the first time. But there are still a lot of things happening, a lot of people working in the area, and feature sets continue to explode. The steady increase in computer power, currently being driven by multi-core technology, means that many more things will become practical — speed is important not only for end users, but for developers as well, to be able to identify and develop commercially viable technologies. Taking advantage of all the cores is more difficult, but I’ve been doing that for many years now; multi-processor systems have always been necessary in robotics.

What interesting developments do you see coming up in the future for CG in general (including camera tracking and related fields)?

Yes, you can expect to see more information extracted from images, and done so more intelligently over time. Many things are possible as toy research projects; it’s always a case of whether or not you can make a reliable and controllable artist-driven tool.

How has the rise of the Net in the 90s changed software development?

Certainly there’s more availability of quality open-source utility packages for tasks such as reading and writing JPEGs etc, which allow developers to concentrate on more significant work. Unfortunately we have not seen similar efforts for state-of-the-art video formats, and as a result the existing formats are less useful and reliable than they could be.

Your attitude towards software development helps support independent filmmaking. Have you ever made/participated/been tempted to make a complete film that includes matchmoving and VFX?

I am a small business, and I do like to see my work allow small independent shops take on bigger and more sophisticated projects than they could otherwise, whether they are the filmmakers directly or small one-man effects houses. It seems like every time a local car dealership owner goes to the movie and sees a sophisticated new effect, they go back home and ask their local guy to do the same, but for a couple hundred dollars, not the couple hundred million of the original.

I’ve learned to leave the film-making to the experts. I have a bunch of my own equipment for shooting demo/tutorial material, but I do that on a very focused basis. In any case, it’s backwards to specifically try to make a movie to show some particular match-moving —- the match-moving and effects need to be a result of the story, not a cause, or it will be too contrived.

Anything you’d like to mention to readers that we haven’t asked you about?

One thing I am asked frequently by people who are new to match-moving is "Tell me which buttons to push" to do their entire shot — typically one I haven’t even seen. It’s like "Tell me which buttons to push to make Toy Story!" You can’t hide behind technology — SynthEyes is an artist-driven tool, not a magic wand.

The first thing that people need to develop, as with any artistic endeavor, is the ART OF LOOKING at a shot carefully, training their eyes and brains to understand what they are looking at. Whether you’re sketching, or doing global illumination, or setting up a sophisticated skin shader, the art of looking carefully must come first and drive the process. In match-moving, people send "moving camera" shots that are nodal pans or have small zooms all the time. Maybe it was shot on a dolly, but at the end they ran out of track and used a zoom for the last few seconds. It’s important for the tracking artist to have the skill and film-making experience to be able to look at a shot and understand how it was filmed and what that might mean for the tracking. Effects technicians must remember that they are part of the artistic process too!

We would like to thank Russ Andersson for sharing with us his views on the evolution of the CG industry. We would also like to thank Brandon Davis, Adam Watkins, Borislav Petrov, Ben Grossman and Arvid Niklasson, as well as Prime Focus, The Syndicate, Mainframe studios and Raz PR for their help in the preparation of this article.

Related Links:

Andersson Technologies' website - home of Syntheyes Matchmoving resources Matchmoving at Wikipedia History of camera tracking at FXGuide History of computer vision at Graz University of Technology History of CG at Wikipedia A Critical History of CG and Animation at Ohio State University

Liked the article? We would like to hear your comments.

"Shutter Island" images are (c) 2010 Paramount Pictures, provided courtesy of The Syndicate. "Pony Tale" image is (c) 2008 Actionhorse Films, provided courtesy of Arvid Niklasson. "Avatar" images are (c) 2009 Twentieth Century Fox, provided courtesy of Prime Focus. All images are the property of their respective owners, no content in this article may be used without the express written permission of its owners. Text, editing and layout by Pablo Hadis.
(c) 2010 MaxUnderground.

By Dave Baker, Neil Blevins, Pablo Hadis
and Scott Kirvan

Foreword

The Max Temple is one of the images most commonly associated with 3D Studio Max. Besides its intent as a promotional tool for the launch of the software it carries a strong message regarding CG art, one that reflects the philosophy of the Yost Group, and is inherently tied to 3D Studio's development. In this article we tread the dangerous steps of high-end versus low-end CG, SGIs vs. PCs, blatant prejudices in the art world, and the completely pointless but never-ending argument of which software or computer makes one a better artist.

For this tricky, sensitive and volatile task we are joined by José María De Espona, author of the Max Temple image and co-founder of renowned Spanish company REM-Infográfica, who, rather than walking with careful steps, rolls right over the subject and shares the experiences that led him to leave the SGI world, the role 3D Studio played in the democratization of 3D art tools and how this chain of events led him to the creation of the Max Temple image. Due to its characteristics, visceral message and wide circulation, the Max Temple could possibly be catalogued as the first art manifesto of the CG age. Hold on tight for a most fun roller-coaster ride that includes a providential SIGGRAPH encounter, the demise of "high-end" workstations, secret cults, melting computers, a second world war fascist unit and much more.

Leaving the SGI world: the difficulties of a religious conversion

We were among the first ones to move from a professional SGI platform, which cost aprox. 30,000 euros, with professional software (add another 30,000 euros), to a “toy”, a 486 which cost around 1,000 euros. The move included using a new software that at the time wasn’t even considered a choice in the professional field: 3DStudio. Truth is back then (1991?) I myself used to look with certain disregard at 3D software running on PCs, such as Lightwave or Strata. I had never used a PC before, only huge animation machines, such as the Symbolics, Tektronix, HP 9000, Prime and of course the SGI.

My conversion took place when I installed 3DStudio on my father’s home PC. I wanted to teach my father how to use the 3D software so that he could use it for his university courses in Madrid (he’s a naval engineer and was back then Professor Emeritus at the Naval Construction program). I had to learn the ins and outs of the application, and I found the original 3DStudio modeler to be very interesting but extraordinarily stiff. I was completely surprised when I started to use the renderer, though. It was unbelievably fast in comparison with TDI’s renderer [Ed. Note: TDI refers to Thomson Digital Image's Explore - see Notes], and it was above all very flexible and reliable when it came to adjusting lights with depth maps. Back then, the only way of producing self-shadowing objects was to utilize overly slow ray tracing or laboriously assign by hand a hierarchy of shadows among the different objects in a scene. This was not the case with 3DStudio. Its philosophy for handling lights was very similar to Renderman’s, but it was much more affordable, fast, and had a user-friendly texturing UI.

When it comes to CG, I’ve always had a tendency to judge the final result more than the procedure used to obtain it, and it was a fact that 3DStudio’s renderer was much more efficient for the kind of images I wanted to make. From then on our solution was to model on SGI/TDI and render on 3DStudio.

It was not however an easy task to convince my partners to adopt 3DStudio, and even harder for my clients to accept the fact that something could be produced with household PCs. We had additional problems, such as the need to keep PC boxes always open and put small fans over the CPUs, because we had literally melted some 486 PCs. Every time clients walked into our office, they would stand there with their mouths wide open looking at the atrocious spectacle of all the disassembled PCs, with their chassis and metal guts in open air, accompanied by the loud noise of the fans. I have to say we lost some clients due to this, and also because rumor spread among the competition that we used home PCs for producing TV ads. It sounds completely ridiculous these days, but back then many people in advertising were convinced that the quality of the 3D renders was determined by the “quality” of the machine on which they were made.

The final push though came from Autodesk’s SDK. Javier Reyes, who was then partner at the company, had developed a program for TDI that simulated “metaballs” or “blobs”, which were quite trendy at the moment due to the animations of Japanese artist Yohichiro Kawaguchi. After chasing him for quite a few days, I finally managed to convince him to take a look at the 3DStudio SDK, and consider the possibility of porting the metaballs application as a 3DStudio plugin. The decisive factors for convincing my partner, however, were coming up with the hypothetical name of MetaReyes for the software, and the promise that this platform, 3DStudio, would end up dominating the whole semi-pro market, which back then we imagined would end up being hundreds of thousands of users.

A SIGGRAPH encounter

By the summer of 1994 we already had a working 1.0 version of the plugin, so I made some “flyers” using images from organic models of the statues of Abbu Simbel that I had made as part of ID spots for Spain's second public TV cultural channel, “La 2” from RTVE. With such meager baggage we arrived to that year’s SIGGRAPH, which took place in Orlando, Florida. At the hotel I made a short animation of some drops of mercury that blended together, we put the flyers into two bags and Javier and I went to the convention center. It was Monday, one day before the exhibition floor opened, but there were already various courses and seminars taking place.

Without asking for permission, we stood up at one of the alleys and started handing out “flyers” to anyone who was going in or coming out of the convention center, with variable results. Some would return them, others would stand there frozen looking at them, considering the possibility of using metaballs on a common PC, when the technology was not even available commercially on SGIs. After a couple of hours of handing out flyers, two guards from the convention center showed up, grabbed our arms, and taking for granted we were foreign exhibitors that had gone out of control, instead of kicking us out on the street, they opened the doors and left us INSIDE the convention center’s main exhibition hall, where exhibitors were still building their booths. So we looked at each other laughing and said, well, this is destiny!

We quickly looked for Autodesk’s booth, which was still in the process of being built. Among Autodesk personnel and those in charge of putting the stand together, there was a guy with a beard who was constantly giving orders to everyone. I told Javier that the guy must be Gary Yost, that we had nothing to lose, and that if the stars’ alignment had brought us into this situation we had to seize the chance. Javier, who spoke much better English than I, talked to him. He was in fact Gary Yost. Before he could react, I put in his hands some images from renderings I had done for RTVE and advertising companies: his reaction was instantaneous. He stood there studying the images. After a brief silence, he turned to me and said he wanted to buy them. He asked how much we wanted for them. The idea sounded completely surrealistic to me, among other things because I thought only my clients could do such a thing. So I told him they didn’t cost anything, that he could do whatever he wanted with them. Gary started giving orders around telling people to take down the images from the booth’s walls and replacing them with ours. Afterward, he kindly lent us one of the PCs from the booth so that we could show him our 3DStudio plugins, he loved MetaReyes.

From then on our relationship with Gary was constant and fluid, and during the three days of exhibition he changed the demo schedules to allow us to do our demos to the public and hand out our flyers promoting the plugins at will. I still remember telling Gary about the bleak future I thought awaited those many exhibitors who were selling software or hardware for big workstations. He replied to me, pointing at both sides of the aisle while turning around "all of this will be vanished in a couple of years...". At that time, to the eyes of the professional market we must’ve looked like a couple of psychopaths, but it was clear that Gary already had 3DStudio Max in mind.

In the coming years our relationship with Gary grew closer, we became beta testers and they listened to our suggestions and observations. We in turn sent them everything we produced with 3DStudio. Autodesk’s software catalogs and corporate videos included much of our work, and we even got word that the entrance hall at Autodesk’s premises in San Francisco had big posters of our work hanging on the walls. In 1995 Autodesk held a contest called “Planet Studio”, and I won in two categories (best rendering and TV advertisement). I still have the trophies on my bookcase.

The Max Temple

A few years later, Gary got in contact with us with a surprising, totally unexpected request. In autumn 1995 Autodesk was already preparing a mass press campaign to announce the official launch of 3DStudio Max, whose first version had been previously presented at that year’s SIGGRAPH. Autodesk had prepared a conservative campaign, consisting of a photo of a skeleton sunk in the sand, holding in his hand the box of the first version of Max. It was clear Gary wasn't too happy with the idea. He logically wanted an image that was created 100% with 3dsmax, which would explicitly show the capabilities of the new software: its powerful modeler, volumetric lights, interactive particle systems, etc., that is, a compendium of the features that made Max something radically new. So he sent me an email asking for an image that followed this concept, telling me I had total freedom and could approach it the way I liked it the most.

Having suffered through previous years due to our fanatic, apparently absurd, intuitive, sudden jump from the professional SGI segment to the PCs and 3DStudio, it almost immediately came to mind the idea of something pseudo-religious, a metaphor of our absolute belief in what Gary and his team were doing. A Temple. I had the clear impression that the image had to be a departure from what was being produced at other 3D companies at that time. I chose an ultra-baroque style, which resulted in calls from friends and fellow designers asking if I had gone completely mad. I chose to use surrealist elements, Masonic symbols, from constructors’ guilds, from a secret initiation ritual of a secret sect, subversive and unknown to the professional production segment, which would show the capabilities of the software to create some kind of “Gestalt”, a complete, all-around tool that would enable any creator to devise any world at will at the speed of thought. My intention was to show in one image the fulfilled dream of my youth’s ambitions, the one that pushed me to leave engineering and submerge into the world of computer graphics. I myself saw Max as the final fulfillment of all that, the ability to create without limitations, without economical or technical restrictions.

Gary approved the idea immediately, but he was far from a passive agent in the creative process. He was very surprised that I included a rhinoceros on the background of the scene, but Gary is a cultured man and understood the oniric, dream-like concept of the image. He did insist on certain details, such as putting one of the cherubs kneeling at the edge of the hole on the floor, looking at the cosmos the mysterious building was floating on. I thought it was an excellent idea and did it that way.

Regarding the angels, why include cherubs instead of birds or flies or anything else that could fill the space of those volumetric lights? Well, it wasn’t an easy or arbitrary choice. In the high-tech scene we worked in, it was on the edge of a “kitsch” or “bizarre” aesthetic. Nonetheless I did not renounce to those little angels due to their metaphorical character. It was a message directed to those CG users who had gotten into the field with the sole intention of creating images. It was a simple symbol, a direct call to the purity of spirit of those that preferred the evidence of facts instead of the prejudices that distort people’s behavior. To me, it meant that they should choose based on rendering results, to trust what they had in front of their eyes, rather than any preconceived notions of the purported advantages of the ultra-costly professional workstations. To provide a final touch, in the foreground, I put a golden angel (another metaphor) looking straight into the reader’s eyes, extending his hand and saying “hey, wake up, what are you doing? This is possible and it’s happening now. You’re losing time and money, make the jump and switch from that God damned platform!”

Looking at things from today’s perspective it’s interesting to note that nothing in that image was accidental or superfluous; everything was in line with a goal, with the intention of conveying a specific meaning. There were other changes that did not come from Gary or me. The first versions included a glowing X, positioned right over the “demiurge“, the total creator located at the back of the room. That X was part of Autodesk’s initial strategy for launching the software, it was MaX’s X, and I seem to recall that at a certain moment the software was about to be called that way, X. However, some European distributors commented that the X symbol could be interpreted politically due to it being linked with a suicide commando of fascist divers that existed during world war two and were known as X-MAS, the tenth MAS. I’m not sure if other rumors regarding the X being too closely associated to US politician Malcolm X were true or totally speculative. But it’s clear at this point that this was a long creative process and as such it was easy for everyone to add all kinds of opinions.

Again, nothing was gratuitous in this process, except for my own work. I refused to accept any kind of payment for the creation of the image, except for the requisite of including my name and that of my company wherever the image was published. All printed images included the text “Image created by Jose Maria de Espona. Triplefactor. Madrid, Spain”. Autodesk kept insisting on paying me, though, so I asked Gary to give us a couple of 3DStudio Max licenses for life, a verbal agreement that regrettably was linked to the destiny of Triplefactor, a company my old partners from Telefónica and Planeta blew up in 1999. But there was another payment from Gary that I still have and keep with great affection, and that’s hanging right now over the desk at my office: a while after 3DStudio Max’s launch they sent me a huge framed cibachrome print of the “Max Temple” image.

Muchas gracias a José María De Espona for giving us such a great account of the making of of the Max Temple image, including his religious conversion from the SGI world. And also for all the trouble he took in digging up the images from old archives and generously lending them for these articles. We would also like to thank Jamie Clay for his help.

Part I: Tom Hudson interview
Part II: Gary Yost interview
Part III: De Espona on the Max Temple image

By Dave Baker, Neil Blevins, Pablo Hadis
and Scott Kirvan

Foreword

In this double feature we have the great pleasure to present you with the history of 3D Studio as related by the head and leader of the Yost Group: Gary Yost himself. Gary shares his experiences of more than a decade of 3D Studio and 3DS Max development, along with many unique facts and anecdotes gained from his personal perspective as captain of the ship. In addition, he also takes a look at the evolution, current state and future of the CG Industry. For this occasion we are accompanied by the esteemed collaboration of Spanish artist José María De Espona, who had a prominent role in the creation of art that was used by Autodesk for 3D Studio's promotion. José María shares the stories behind the production of some of the images that appear in this article, chosen by Gary Yost, which many 3D Studio / Max users will quickly recognize. Join us once again for another refreshing dive into the waters of 3D history.

The Interview

- What first attracted you to graphics software development, as opposed to another field of computing?

I had always been interested in photography and making short films when I was a teenager -- at that time I thought that my future was in video or film production.  When the Atari 800 was released in 1980 I saw that perhaps microcomputers would be able to support the creation of “interesting” graphics.  The fact that the 800 had multiple co-processors to handle different graphics and sound functions seemed to indicate that micros were headed towards a fun place to play.  That really sparked my imagination and led me to believe that possibly my interests in imaging could converge with computer tech in the future.  It really didn’t take too long for that to happen… I founded Antic Software in 1984 and then when Tom Hudson showed me his little 3D prototype for the AtariST at Comdex in November 1985, a blindingly bright lightbulb popped on above my head.  Really!  It was an epiphany.  We spent the next 2-1/2 years making the AtariST do things the Tramiel family never dreamed of.

- What made you decide to take Eric Lyons’s offer, jump ship from a comfortable position at Antic, and embark on the development of 3D Studio? What convinced you that Autodesk would be a good company to help bring that project to fruition?

First let me say that developing for the AtariST was a very limited game.  Our market was small and the lack of expandability was really crippling.  The display supported only 4bit color and I’d been hearing rumors of IBM launching the 8bit VGA display.  Plus, on the IBM there were 3rd party 24bit framebuffers like the AT&T Vista and Targa cards.  And of course the IBM was a commodity market, which had additional benefits of size and expandability.  So when David Kalish and Eric Lyons at Autodesk offered us a development contract after Comdex ’87 it really just seemed the next logical step.  Autodesk was having quite a bit of success with AutoCAD by that time and we felt comfortable with their laid-back experimental philosophy… very non-corporate with dogs in the hallways, weekly beer busts where ideas were shared, and a general good feeling.  Plus it was just across the GG Bridge from SF where I was living at the time, in a beautiful part of Sausalito.  It was a great situation.

- When did Dan Silva join the original 3DS development group? What unexpected tasks did you find yourselves fulfilling as the project grew over time?

This is an interesting story.  Originally the 3DS engineers were going to be Tom Hudson and Jim Kent.  The thinking was that after creating Autodesk Animator, Jim would start working on the keyframer while Animator was in maintenance mode.  But you know how software becomes all-consuming and of course our Animator users wanted more features, more resolution, etc and Jim needed to get to work on Animator Pro, which ended up being quite an amazing app.  (Aside: Jim later became a hero in the scientific world when he wrote the software framework which gave the Human Genome Project the ability to finish sequencing the genome 3 days before Celera could patent it.)

Tom had started work on the framework and modeler/lofter modules in 1988 and by summer of 1989 things were starting to look really interesting.  I had been asking folks at Autodesk if they knew of any engineers who were interested in 3D animation for the keyframer position, and in August I was approached by Walt Spivak, director of product marketing, about his neighbor Dan Silva.  Dan had developed DPaint for EA and literally over his backyard fence had been expressing his boredom to Walt.  I called Dan and asked him if he might be interested in working with a little team on an interesting new graphics project for Autodesk, but I did not tell him it was specifically in the area of 3D keyframing.  (I wanted to find out what he was interested in first, because my theory of putting development teams together was that the developer must have a passion for the task.  That would potentiate much higher synergies and productivity than if the developer is arbitrarily given his marching orders.)

So, Dan came into a meeting room with us in September 1989 and we asked him what he wanted to do.  He mentioned two things… one was a resolution-independent paint program and the other was a 3D keyframer.  Bingo.  Dan and I signed his contract at about 3pm on October 17, 1989… just 2 hours before the magnitude 7.1 earthquake that struck the SF Bay Area that day.  I was in Sausalito at the Autodesk offices copying his contract on their Xerox machines during the earthquake.  Those buildings were built on landfill, so you can imagine how much they moved.  It was quite a day.

Regarding unexpected tasks…  After 3DS DOS r1 shipped on Halloween 1990 we started getting lots of user feedback.  It was obvious to all of us that there was gold in that feedback and I built systems to collect and organize it for future versions.  Of course many of the feature requests were “unexpected”, but we took every request seriously and that became the hallmark of our future versions… user-driven features.

- What was the Yost group's dynamic like? What was a typical day of work like?

The dynamic was… dynamic.  We had taken on a pretty hefty responsibility coming from a small Atari software publishing operation and then being tasked to write the “AutoCAD of 3D animation”.  You can imagine there was a fair amount of pressure associated with that.  So we worked really crazy long hours.  A typical workweek from 1988-1998 was 90-100hrs, and for the 3 months leading up to a release 110+hr weeks were de rigueur.  Since we all worked out of our own homes and were in long-term relationships, we had enough support to make that possible, barely.  We only took time off after a release shipped, so you can imagine what those 18-month release cycles were like.

A typical workday was just sitting in front of our PCs… for me mostly working on specs (2 days a week with Jack Powell), communicating and running tests.  From the outside it would’ve appeared pretty fairly calm, but internally the pace was frenetic.

- Autodesk co-founder John Walker states in his Information Letter 14 (April 1991): “The very existence of the Multimedia group is an admission of the neglect for Animator after its hugely successful initial launch. If Autodesk had, in 1983, treated AutoCAD the way it treated Animator and 3D Studio after their introduction, Autodesk would not exist today. (…) Don’t touch the product. Just move the price down to where the early adopters can afford it.” How was 3D Studio perceived within Autodesk and the company’s Autocad-oriented development strategy? What was the Yost Group’s relationship with John Walker like?

I’m not sure what I’ve done with my copy of Walker’s historical book “The Autodesk File” [see Notes], but that’s a really incisive look at what was going on back then.  I recommend it to anyone who wants to learn about Autodesk’s early days.  3D Studio was originally viewed internally as little more than a research project to see if it was possible to achieve workstation-level 3D graphics on a PC.  Once 3DS/DOSr1 shipped and we started getting so much great feedback, the people at Autodesk began to see that a new franchise might be possible.  For quite awhile we were the poor stepchild of the company but we slowly gained more and more respect from the CAD division.  Eventually, once Autodesk acquired discreet logic in the late 90s, Max and the other DCC products began to be seen as a full-fledged franchise in their own right.

Our relationship with John Walker was friendly, and he was a great supporter of what we were doing.  That said, he was stepping out of the day-to-day business of the company by the early 90s and on his way to new ventures.  Most of my favorite memories of John were his show-and-tell moments at those legendary Friday afternoon beer busts.  He’d bring in all sorts of new tech to share with us, including superconducting materials (in supercooled containers) and stuff like that.  John is an extremely smart person… like nobody I’ve known before or since.

- Tom mentions that the name 3D Studio was suggested by Autodesk marketing, along with other possibilities such as “Yost”. Do you recall other possible names that were considered?

Tom’s memory is correct… there was a very surreal meeting in which I had to shoot down the idea of using my own name as a brand for the Autodesk line of DCC products.  The name 3D Studio came about because Tom, Jack Powell and I had previously created a package called Cyber Studio for the AtariST, which included CAD-3D as the primary component, along with some other plugins and 3D clip art.  3D Studio is just a concatenation of the second parts of CAD-3D and Cyber Studio.  3D Studio!

- John Walker recommends in April 1991 to “move the product to Windows as soon as technically feasible”. When and why did you decide to embark on the development of 3D Studio MAX?

Oh boy.  We had a good thing going with 3DS DOS.  Lots of IPAS plugins, great enthusiastic user base and excellent sales.  But there were lots of problems with that architecture and we all knew that we wouldn’t be able to keep our DOS users happy once Windows was capable of supporting large memory via 32bit addressing.  The separate working modules were a huge problem, along with lack of procedural editing and totally modal IPAS plug-ins.  These things drove us (and our customers) crazy by the time we got to DOSr3, and as soon as we found out from IBM that WinNT 3.1 was going to be released summer 1993 we made the decision to re-architect.  The first step was hiring Don Brittain away from Wavefront, where he was chief scientist.  Don was unhappy with Wavefront’s decision not to move to the PC platform and he’d been talking with me about that.  In fact, Don was on the verge of accepting a job at Microsoft and moving from Santa Barbara to Seattle when I came to him in January 1993 and asked him to join our team.  Of course when he found that he could work from his home instead of moving that was a great benefit, and he started working on Max prototypes in February.

1993 was a big year for 3DS/DOS too.  We added hundreds of new features while Don was in the skunkworks with the Windows version.  He had some pretty exciting prototypes working by the end of that year and we were all itching to move over and get working.  But there were still some critical features we needed to get into the DOS version before we spent 2 years on the new platform… particularly inverse kinematics.  Remember, we only had 2 developers on DOS and 1 on Max at the time… no extra resources were available.  Fortunately I’d been working with Rolf Berteig on some interesting IPAS plugins (in the area of L-System plantlife) and he was such a joy to work with that I began talking to him about doing a more ambitious IPAS IK plugin.  He jumped right in and that became the heart of 3DS/DOSr4 (well, maybe we should’ve called it 3.5).  It was obvious that Rolf was an intrepid software engineer… he had no qualms going anywhere or doing anything.  During that project we made him an offer to become a full partner and work on Max with us as soon as he was finished.  Once DOSr4 shipped in mid ’94 we had four brilliant and completely unstoppable engineers to focus on Max.

Don had a very solid prototype by that time and we had written a specification for this new app that we code-named Jaguar.  In September we began development in earnest, but we ran into snags.  Our MIS infrastructure was very shaky… we were using dial-up 56Kbaud modems into a server at Don’s house.  The code base was going to get quite large and already in the beginning we began to run into problems with dropped phonelines during source control system updates.  Don was spending all of his time just trying to manage source database corruption problems during October, and by early November we were almost ready to give up.  (For the want of a nail…) 

Luckily Gus Grubba, who we’d been working with on Video Post and other aspects of 3DS/DOS and was a great friend, mentioned that a new form of data-communications called ISDN was just coming online.  We all checked with our local telecoms and found that it was available for all of us just that month!  It took a few weeks to get everybody up and running with ISDN but by January 1, 1995 we were flying, rock solid.  Now things really started to move, and the rest is pretty much history.  It only took us 6 months to bring Max up enough to show (what looked like a nearly finished product) at Siggraph LA 1995.  The demo totally blew away the Microsoft/Softimage folks, who had just merged and begun work on XSI.  And the people from Alias simply stood around our booth, mouth agape.  That was my favorite Siggraph ever.  We shipped Max in April 1996, so in all there were only about 15 months of actual development time in r1… quite a feat for 4 engineers with a completely new architecture.

How did Character Studio authors Michael Girard and Susan Amkraut join development?

Michael Girard and Susan Amkraut were an incredibly important part of all this.  They had met with our product manager in very early 1995 and were then introduced to me.  We showed them what we were doing and they showed us some early Biped tests.  There was a lot of excitement in the room that day and we all saw how interesting it would be for Biped to prove the flexibility of our APIs.  Plus of course Biped was a brand-new way of animating – something that our customers had been only dreaming about… procedural character animation.  Another revelation along the path.  The dancing baby animation actually first manifested as a dancing monkey in our 1995 Siggraph demo, and that totally floored the crowd.  Not only had we launched a new procedural architecture, but there was Biped… a revolutionary way of animating.  It was almost too much for some people.

- When and how did the name 3D Studio Max come up?

This was a tough one… we were obsoleting our DOS product with this totally new (and basically incompatible) Windows technology, but we wanted to maintain the 3D Studio brand.  In a 1995 meeting with Anna Mellilo (our marketing manager at Autodesk), we were struggling with how to do this.  On a break, she started talking about her young son Max.  I’d been chewing on the name problem in the corner and heard her say “Max” just as I was thinking “3D Studio”.  So thanks to Anna (and her son), we got our name.  (Although it was ironic that Max did not run on Macs.)

- When did you come up with the concept of integrating the different 3DS modules into a unified modeling/animation environment?

We’d been looking carefully at various workstation-level 3D animation systems all the way back to the 1980s, and were especially impressed by systems from Symbolics and Side Effects.  They provided a level of integration and procedural control that was just so powerful.  At the same time, we wanted to create something that was easier to use.  It was obvious that an integrated environment was what our customers wanted, so that was our paramount design goal.

- What key elements were taken into account when designing Max’s modifier stack? Did you think about pushing proceduralism further?

We certainly were committed to a procedural approach, but even more important was our need to make it robust.  Our biggest design goal was to make sure the stack was going to work quickly and reliably in production.  That inspired things like lazy stack evaluation to make it fast by not doing anything that wasn’t necessary at the time.  (One of the 8 patents that we were awarded for Max was in this area.  You can google it with the search term “3ds max lazy evaluation”… it’s interesting reading and was totally novel at the time.)

The capabilities of the modifer stack improved over time – especially when the development team at Autodesk added the Stack View for r4.  That was an outrageously great enhancement to the stack and I’m sure everyone is grateful for it.

- IPAS, plugins and scripting capabilities have been critical to 3D Studio/Max’s success. Were you involved (or would've liked to be involved) in the development of any particular Max plugins? Which plugins/scripts developed for 3DS/Max have surprised you?

Max plugins?  No.  I primed the IPAS plugin pump with my catalog of add-ons to 3DS/DOS because we needed to show people what could be done with a plugin interface.  But Max was so all-consuming that I had no additional bandwidth to put into Max plugins.  Of course I’ve been astounded at all the creative plugins that have been developed over the years, but that was expected… it was why we put all that work into the Max SDK.

- Why did you decide to leave the Max team?

I had been operating under the “deferred-life plan” for about 10 years by 1998, and there were a lot of other things I wanted to do.  Photography was my first passion and I wanted to re-explore my relationship to that, plus my family needed to get out of San Francisco and find a house in a more rural area surrounded by nature.  But the biggest thing that my wife and I wanted was to adopt our second daughter, Ruby (which we did, from China, in 2001).  That has been the biggest joy of my life.

- What have you been up to since leaving the Max team? Please tell us a little about your work at Mental Images.

After starting up version 2 of our family I began to have some very interesting conversations with Rolf Herken at mental images about how it may be possible to add a 3D backbone to the internet to make interactive 3D possible on a global scale.  His vision led to the development of RealityServer.  I was involved with that on a consulting basis until we began to get serious about the idea of a device-independent rendering language that would solve the myriad of problems with shaders that arose because they were platform-specific.  This was a really meaty problem, and Rolf Berteig had been doing some research in this area.  I came on board officially, brought RolfB into the mental images fold about four years ago and the results have been tremendous.  Max now ships with MetaSL language support along with the mental mill editor, and that’s tremendously satisfying.  MetaSL is now being used by major Hollywood studios and will be integrated into many other DCC and CAD apps in the near future.  You can read more about mental mill, MetaSL and RealityServer on the mental images site.

Beyond that I’m raising a beautiful daughter and love the work I’m doing with my personal imaging work.  I’m going to start teaching photography to middle-school students next year, which is really exciting for me.

- What is your current perception of the CG industry? How has the industry changed since you started developing 3D Studio, and what role do you think 3DS had in those changes?

Hmmm….  Consolidation.  When I started it was a highly fragmented industry with hundreds of different small software entrepreneurs developing lots of creative tools, healthily competing with one another.  Now of course the Big Three are all owned by Autodesk, and that’s certainly changed the competitive dynamic.  Also, when I began doing this many of these tools were very expensive… in the tens of thousands of dollars.  We began the trend of cost-reducing professional tools, but that got out of hand for awhile – sort of a commoditization of the market.  Customers started to expect an awful lot of functionality for a very low price… lower than the costs of actually developing this stuff.  And that put a lot of great engineers out of work.  Many have since moved on to other industries.  So there’s been a brain drain due to this low-cost price expectation and I feel partly responsible for that.  Those high prices in the old days weren’t sustainable though, and it was one of the reasons consolidation happened.  Perhaps this was all part of the process.  Who knows?

- Why do you think 3D Studio / Max has become the widest selling 3D software for almost 20 years?

It’s the same reason why the Swiss Army Knife is still so popular.  We wanted to make a tool that could be used by anybody to do anything.  It was a lofty goal but the market proved that we succeeded.  And I attribute that success in great part to the stellar design and engineering work of Tom Hudson, Dan Silva, Don Brittain and Rolf Berteig.  Plus we had a real human being on our team to vet all of our designs… Jack Powell.  Without Jack our work would’ve ended up as a bunch of arcane tools, usable only by true mutants.  There are lots of other people who we’re indebted to as well.  Mark Meier, Gus Grubba, Jamie Clay, Martin Doudoroff, Kyle Peacock, and many great people at Autodesk (and elsewhere) who took the ball and ran with it in such creative ways.  Max is a huge team effort and never would’ve been possible without this gestalt of imagination that blossomed at that particular place and time.

- Do you feel that the graphics field is as exciting as it was 10-20 years ago, or has innovation basically slowed down these past few years?

Sure, it’s always going to be exciting.  The goal in 3D computer graphics is to pretty much play God within the silicon.  Nature is an awesomely complex system and the attempt to simulate the universe  is a never-ending quest.  You can see what James Cameron has done with Avatar as a great example of a good signpost along the road.  The mocap is very close to perfect, and the rendering is getting better and better.  (Actually the potential is there now for the rendering to be purely photorealistic in such a film, instead of the cartoony quality that we see now.  It would just take much more time and be even more expensive.  That’s coming soon.)

I always used to say at our big yearly user events at Siggraph that any specific release of 3DS was a finger pointing at the moon.  And who knows?  One day we might actually walk on that moon.  We’ll see.

- What are the most interesting developments that you see coming up for the future of CG?

For me, the most interesting work is in this area of making all DCC and CAD apps work together in the field of rendering.  A shader you create on any platform is renderable by every MetaSL-compliant hardware and software renderer on the planet, and it’s future-proofed so that it’s renderable for all time.  That’s my particular niche right now, and I love it.  And on a larger front, the concept of making interactive 3D a ubiquitous part of the Internet experience is a super challenge that we are making great strides towards.

Of course there are a lot of other developments coming up, but most of them are going to be evolutionary updates to existing technologies instead of radically new ideas like MetaSL and RealityServer.

- You mentioned choosing the name Cyber Studio for Antic’s software because the company was creating tools for “graphiker pilots to voyage into their imaginations”.  How has the profession of CG artists evolved from the early days of 3D Studio to these days?

The word Cyber was originally derived from the Greek root Kubernetes, which means "helmsman" or "ruler".  I chose to use Cyber for our tools just after reading William Gibson's seminal work "Neuromancer", which launched the concept of Cyberspace.  Our dream was to help make that possible.  Little did we know at the time how challenging that would be.  So, how have things evolved for CG artists?  The tools are certainly better and more complex.  This has led to more and more specialization in the field.  I used to envision these tools as being so general purpose that anyone would be able to design and build characters and sets, then animate and light them and we’d all be Renaissance-level creators.  That jack-of-all trades dream is still only available to a relatively small group of very creative and special individuals.  I have always had the utmost respect for them, and their work always inspired me to go farther creating better tools.

- You participated and saw the birth of many socially-oriented CG activities and trends. How has the net changed 3D related social activities and CG software development?

How?  In pretty much every way conceivable.  The Internet has made it possible to share ideas, which creates a synergy that nourishes creativity in unpredictable ways.  The early versions of the net made it possible for us to use Compuserve as a place for our customers to meet and tell us what they wanted. That was key to our success.  Now it’s possible to share everything… the net is the fertilizer for the global organism that we’re constantly creating with our minds.

- What one piece of advice would you give the team currently working on Max in Montreal?

Keep listening to customers but always focus on reliability.

- Any interesting or funny anecdote that you recall from 3DS / Max development?

The night before submitting our Jaguar specification to Autodesk in 1994, I was completely exhausted.  I had to go to Kinkos to make a bunch of copies of it, and on the way I took my wife out to dinner at a Chinese restaurant in SF called Ton Kiang.  After dinner, they brought us fortune cookies and my fortune read “Your future plans are going to succeed.”  I took that over to Kinkos, taped it to the back page of the spec, and the rest is history.

- Have you ever felt tempted to become the head of the Max development group a second time (sort of like Steve Jobs' return to Apple)? If so, what would be some new directions you'd like to bring the software? Would you develop it as cross-platform software, for ex?

No. Been there. Ken Pimentel, Shane Griffith and their team at Autodesk have done a fantastic job continuing to grow Max. That has allowed me to happily move on to new projects while feeling that our "baby" is well taken care of.

We made the decision not to go cross-platform because of our small size.  Autodesk themselves had originally been highly cross-platform with AutoCAD, but it caused so many headaches that they shut that down to focus on Windows.  It’s the old story about spreading yourself too thin.  It’s extremely challenging to support multiple operating systems.  That said, we’re multi-platform at mental images and although it adds many dimensions of complexity, the tools now exist to be able to manage it.  It’s a necessary evil for us in the component software business.

- Any particular images or short films created with Max that you enjoy especially?

I’m not too current on all the work out there, but as Tom noted in his piece, I have to bow to Alex Roman for his masterpiece “The Third and the Seventh.”  That’s exactly the type of work that would’ve spurred us on to much greater heights in the past.  We fed off that kind of outrageously amazing stuff.  Without it, we never would’ve had the energy to push as hard as we did.  My Dad was an architect, and he would’ve been absolutely floored by seeing what Alex did with Max.

- Anything you’d like to mention to readers that we haven’t asked you about?

Hmmm…. Just that I want to thank all the Max users out there for being such good sports.  You pushed us when we needed it, and you also cut us some slack when we had to move on.  I’m grateful for all of it.

I had a tradition back in the old days of writing a haiku poem to accompany each of the weekly releases during our beta periods.  It was a fun way to end a hectic week and it brought a little sense of peace into our crazy lives.  While working on this interview for you I came up with this one to sum up my feelings on this winter day.

Twenty years later.

The rain still falls. The sun shines.

We remember friends.

We would like to thank Gary Yost for sharing with us his experiences from years of working on 3D Studio development, for all the effort he dedicated to this article, for his kindness and generosity. Also for the extra effort he took in finding, processing and recovering the images from the Yost Group and making the best out of them so we could present them to the public. We would also like to thank Tom Hudson, Michael Girard, Susan Amkraut and Martin Doudoroff. A special thanks goes to José María De Espona, for all his goodwill and help in the preparation of this article.

Part I: Tom Hudson interview
Part II: Gary Yost interview
Part III: De Espona on the Max Temple image

By Dave Baker, Neil Blevins, Pablo Hadis
and Scott Kirvan

Foreword

3D Studio was released to the public in October 1990. The 3D arena at the time was populated by a diverse range of software running on various platforms: LightWave 3D, Sculpt-Animate 4D, Imagine, Topas, Electric Image, Softimage|3D, Wavefront Advanced Visualizer and Alias PowerAnimator, among others. Back then, the words PC and multimedia were not used in the same sentence, and few people would've dared to place their graphics software bets on PCs. The first version of 3D Studio was created by four developers, and its low cost, versatile functionality, and expandability via a plugin architecture would make it the most widely used 3D software in the world.

We had the great pleasure of interviewing Tom Hudson, who has been one of the main developers of 3D Studio, as well as the author of the software's precursors, Solid States and CAD-3D. In this interview, done some months away from the software's upcoming 20th anniversary, Tom shares fascinating details on the development of Autodesk 3D Studio and 3DS Max, as well as his views on the current state and future of the CG Industry. Be sure not to miss the extra info written by Tom that can be found by clicking on some of the images, and which sheds light on many of 3D Studio's iconic illustrations. We hope you enjoy this dive into the waters of 3D history.

Pre-History - the Antic Software connection

The history of 3D Studio's development began several years before its official launch in October 1990. Its early origins can be traced to Solid States, a 3D wireframe renderer that ran on the Atari 800 and was published by ANALOG Computing magazine, and CAD-3D, the next incarnation of the software, which was published by Antic Software. Antic's software arm was being managed by Gary Yost, and thus started the collaboration with Tom Hudson in 3D graphics software development. Gary, Tom and Jack Powell were the three initial members of the Yost Group, and Dan Silva (who created Deluxe Paint for the Amiga) joined them later. The four of them worked on the development of 3DS release 1.

This part of history is told in great detail and with excellent documentation by Martin Doudoroff, an early user of CAD-3D who collaborated with the Yost Group on many projects. We strongly encourage you to read it.

The Interview

- What first attracted you to writing graphics software, as opposed to another field of computing?

It's kind of a strange, indirect path, but the best I can explain it, it all started with a TV show called Space:1999 back in 1975. I have always been a science fiction fan, as far back as I can remember, and when I saw the special effects in Space:1999 back in high school, two friends (one of them Jon Bell, the author of such books as 3ds Max Killer Tips and 3D Studio Max f/x) and I started making 8mm films, concentrating on the special effects (multiple exposures, mattes, miniature work etc.) It was a lot of fun, because you'd have to figure out how to make something miniature look real, or work out how to make a good-looking planet, and so on.

Then came Battlestar: Galactica in 1978, and they had these awesome vector-graphic displays all over the place that were displaying beautiful wireframe graphics of spaceships. I bought my first computer, a Compucolor II, shortly after that, dreaming of creating cool graphics and making animated movies. I taught myself the math involved and got stuff working but the Compucolor II just wasn't up to the task of doing smooth animation like I wanted.

The next several years involved working on Atari machines (first the 400/800 series, then the 68000-based ST), all the time looking at movies like Tron for inspiration. I wanted to make my own movies using 3D graphics.

Once I got involved with Gary Yost, who is a true visionary, the ball really got rolling and we were on the track to creating software that would do what I wanted, and we kept getting better and better hardware that moved us toward that goal.

- One thing that made Max special was the fact it was sort of halfway between Houdini's procedural workflow, and Alias's "Follow my command, and don't ever let me go back to change my mind" sort of mindset. What made you guys choose that sort of mindset at the beginning, or was that something that just sort of evolved naturally?

Without a doubt, working with the software personally led us to that! Because I wanted software that I could make movies with, everything we put into the software in those early days was something I wanted or that made the software better. Every day, after finishing work on the software, I'd sit down and use the new features I wrote to create some sort of animation. By actually working with the software and experiencing the good and bad features, I was able to really tune it to be useful. Jack Powell, who was part of the original Yost Group with me and Gary, was adamant that we have consistent user interface from one part of the software to another.

Anyway, the original 3D Studio for DOS was sorely lacking in undo capability. It was maddening at times, and it's something we should have done from the start. I think the big thing about it back then was memory usage to save the prior state, and we were trying to watch every byte - So we had a workflow where you'd simply use a Hold/Restore command to manually handle undo operations.

When Dan Silva joined the team, he immediately noticed that once you did something like loft a shape into an object, that object's form was fixed and if you needed to adjust something, you had to go back, adjust the loft shapes and execute the loft again. Dan's idea, which became the core idea in Max, was to have each object have procedural elements that could be tweaked after the object was created. It was a key element, and along with the undo system we developed, made the system a million times better than the old DOS version!

So to answer your question, it was something that evolved as we used the product and seemed like a natural addition, made more feasible as computer processor speeds increased and more memory became available.

- Why did you choose not to go fully procedural, like for example a complete schematic editor to control all portions of your scene?

Well, early in the development of Max, we were dealing with some pretty tricky concepts in terms of what would be efficient with processor time for evaluation, and as I recall we were a little nervous about going down any "blind alleys" that would turn out to cause problems for the viability of the application. I think we pushed the envelope about as much as we dared!

- How much autonomy from Autodesk did the original 3DS DOS and Max team have? Was Autodesk heavily involved down to the specific features, or was it mostly a publisher kind of relationship where you guys provided the finished product and Autodesk sold it?

This was the really sweet part. Gary, Jack and I created a spec for what was to become 3D Studio DOS and took it to Autodesk's Eric Lyons. The spec was very ambitious, built on all the work we had done on the Atari ST platform with our "Cyber Series", but added things like a keyframer and rendering features like bump mapping, which I had been itching to do but the Atari graphics hardware was too crude to use it. We were totally in the driver's seat with the project, and the cool part was that after we had made our presentation to Eric, he took me aside and said "Tom, can you guys really pull this off?" When I told him we could absolutely do it, he was totally stoked. Everybody was so excited to take our work to the next level, it was a great time.

We almost never got dictates from Autodesk about things we had to include. They wanted us to include support for ADI, the Autodesk Device Interface, but it was more of a request than anything forceful. We added it so that we could have tablet support and such, and as it turned out, it was lucky we did so, because when I added the ADI code, I realized we could build on that mechanism to create the "plug-in" concept we added later!

Most of the features we added were driven by customer feedback and by my desire to make a product that I could do my animation work with. Some examples:

--- A couple of weeks before we rolled out 3D Studio DOS R1, I drew up a quick storyboard for a little film I called Cornerstone, which told the story of a building cornerstone who dreamed of going off to the beach, but he couldn't leave because the building he was part of would fall down. I decided to whip up the film using my two 386/20 computers in my office. Jon Bell created animated texture maps for the cornerstone's eyes using timing sheets I gave him, and I modeled up the assets and did the animation. Because Dan Silva had written the keyframer, it was just about my first serious animation using that part of the package. As you might imagine, with only two programmers on the project, it was a crunch just getting the thing finished on time!

One thing I wanted in Cornerstone was some nice dissolves from one scene to another in a couple places, and 3DS wasn't set up for that, so I snuck in some code that operated from the command line and used the bitmap handlers to do simple compositing, fades and such as a batch process. It was actually an undocumented feature of the release version of 3DSr1! For 3DSr2, we built an interface for it and created the Video Post module.

Sadly, my crowning achievement in the keyframing of Cornerstone, where the cornerstone starts to break free of the building, was fouled up when several frames of the sequence were corrupted on the media I sent to Autodesk for mastering. As he initially breaks out, his eyes squint several times with his effort, then he quickly slides out, pushing several pieces of trash on the sidewalk out of the way nicely. In the final piece as it ended up on the video, he starts to squint, freezes, then suddenly pops out. I was heartbroken when I saw it - Nobody caught it when they did the mastering because they didn't know what I had intended. The original 3DS files and frames were lost later on - I had them backed up on some magneto-optical media, cutting edge at the time, and when I went to retrieve them several years later, they were unreadable.

--- Another piece I wanted to do was a little short set to "Blondie's" Europa, called Assault on the Sun. It was to feature Freddy and Joey, characters developed by me and my high school friend Steve Dailey in 1975, and Klanky the Robot, Joey's home-made robot friend, flying to the sun in an Apollo capsule to retrieve a bucket of sun. I came up with the idea shortly after finishing up one of the 3DS releases, when I was at a SIGGRAPH conference, and while planning the piece I realized I would need a nice motion-blurred starfield for the background as the Apollo capsule and lunar module flew toward the sun. This was where I came up with the plug-in concept that became the original "IPAS" plug-in interface - I'd simply create a starfield module that would get information on the camera view and send the resulting starfield bitmap to 3DS for compositing! As I mentioned above, I built on the mechanism used by the ADI devices to accomplish this. Once I had this mechanism in place, we came up with several places where it could be used. I never made the short, but we finished IPAS!

--- Probably the best example of my needs driving the development of 3DS was when I got a job doing some rendering for Paramount's Star Trek: Deep Space 9. Their physical special effects models, which were used for the exterior of the space station, wouldn't be ready for months, but they needed footage of these elements to place on video monitors in the station's security office for live-action filming. I had worked with Paramount's Mike Okuda on physical models before, and he contacted me to see if I could create virtual, 3D graphics versions of the station exterior for these shots. I said I could, and got to work.

This was done on relatively slow machines by today's standards, and I had a lot of work to do in a short time - Moving starfields, composited behind several shots of the station, with people moving in the windows. I had three computers in my office at this time, and rudimentary Ethernet networking. Copying the files around from one machine to another and wrangling all the data for scenes that all rendered at different rates was maddening. Then it hit me - The idea for integrated, distributed networking in 3DS. I whipped up a rudimentary networking setup where one machine could act as a master, handing out work to the other machines - And I got the job done on time despite the massive number of frames required. We polished up this network rendering mechanism for the next 3DS release!

- How did the name 3D Studio surface? Who came up with it? What about 3D Studio Max?

I had a couple of names in mind for the original product - "Paragon" was one, but I don't think that made much of an impression on anyone. I do remember the meeting where we (Gary Yost, probably Jack Powell and I) met with some people who were from Autodesk's marketing department and they pitched some names to us. I only remember one other than 3D Studio: "Yost"

Gary was stunned when they pitched this name - They were talking about it becoming a verb, almost like "Google" is today! I remember them saying that they could see people talking about going to "Yost a scene" or "go Yost that model". I think we were all amazed by this; none of us were happy with it, least of all Gary.

Then they suggested "3D Studio" which we all agreed was a good direction.

After the meeting, Gary told me he was sure the "Yost" name was a smokescreen, something that they knew we'd all hate so much that the name they pitched next (the one they really wanted us to go for) would seem so much better that we'd all agree on it. I guess it worked.

3DS Max was just a natural extension of the brand, so that was a simple one. Its code name was "Jaguar" because it was built on top of our rendering engine, which we code-named "V8".

- What was the Yost group's dynamic like? Was it a benevolent dictatorship? A republic? A democracy?

It was probably most like an equal partnership. Everybody got to contribute their input to the feature list and final product. Typically, when we were getting ready to start on a new release, Gary would work up a feature list, based on what we were hearing from users and what the members of the Yost Group felt was important. He'd run that by us and we'd all look it over, and normally the list was reasonable - We'd break it down into tasks and these generally fell into areas that were obviously suited to one programmer or another, and they'd go on our personal task list. Sometimes things might come up on the list that probably weren't doable in the time we had, so those would get pushed to the bottom of the priority list or dropped altogether.

- What was a typical day of work at the Yost group like? What was the office like?

This is funny, because there was no Yost Group office, per se. Gary, Jack and Dan were in the San Francisco area, but lived miles apart and I'm not even sure how often they got together. In the early years, I was in Shawnee, Kansas and worked out of my house. In 1993 I moved to Wisconsin. Later when we started on Max, we added Don Brittain (Santa Barbara, California) and Rolf Berteig (Seattle, Washington) to the team.

Periodically, I'd fly out to San Francisco for meetings at Gary's home/office and work with the guys on design stuff for a few days, but I'd head home to do most of the work. Everything was done via phone calls and email through CompuServe in the early days, then email and ISDN service later on.

Because I was working with people two time zones away, I shifted my hours so that I started work at around 10AM and worked late into the night. It was nice because there were no "official" working hours - I worked when I wanted to and as late as I needed to.

- How important was user feedback to Max's design at the very initial stages? At the later stages where you were flushing out the details of a design? Did you ever have to radically rethink a feature after getting user feedback?

User feedback was incredibly important, and as I mentioned earlier, Jack Powell, who handled all of our manual writing and such in the early days, really kept a close eye on our consistency between various parts of the program. With Dan and me working on separate parts of the program, it was common for us to do similar things in two different ways in the user interface, and Jack would see these right away and get us to use one common interface.

We also went through several iterations of the main interface before we settled on the final one, again thanks to Jack sitting down and actually working with the software in order to get a feel for the workflow. I remember the "command column" section of 3DS's interface being completely reworked several times before we felt it was ready.

Jack was our original "user" and was invaluable in getting the thing uniform from the start. He actually created some of the animations on our original demo reel, which I thought was pretty cool.

- Did you participate much in the Autodesk Compuserve forums or internet webboard?

I was in there a bit, but not as much as I could have been, simply because I was so busy coding all the time! I didn't have a lot of spare time. IIRC, Gary spent the most time talking to users and getting the feedback on features.

- Why wasn't Max developed as a cross-platform application? Was the idea considered? Why did it end up being tied to Windows?

Good question. In fact, as we were at the big decision point at the end of our Atari ST era, we were looking at going to the Mac because of its then-superior graphics capability over the PC. Gary had sent me a package with an Apple order form and a list of Apple computers and other hardware, and we were looking it over to see what we would need to start development!

Then the VGA card for the PC was announced and Gary could see that this was the way for us to go - By the time 3DS was shipping, he said, every PC sold would have a VGA card, which would give us a graphics solution with 256 colors. And he was right. There was also the ATVista graphics card for the PC, which we would be able to support for truecolor output (what I really wanted!)

Going with the PC gave us the potential for a HUGE market, which was important for a niche product like ours. It worked out well.

As for why we didn't go for cross-platform, it didn't occur to me much at the time, mostly because I was up to my eyeballs in work just getting the software working on the PC! It would have been nice to build the brand on the Mac as well, but we would have had to build a larger development team, and to tell you the truth, when it was just Dan and me coding, we were very efficient and almost never stepped on each other's code. The more coders you get, you start running into a "too many cooks" effect and diminishing returns. I also don't remember any discussion of this from Autodesk, interestingly, and that's where I would have expected a push in this direction to come from.

- Why wasn't Max's Video Post module further developed?

I'm not sure other than the fact that at a certain point, you started to get other packages out there that were very good at compositing, and there wasn't much user demand for it from us. I'm sure Gary would have pushed for it if users were asking for it.

In a way, Video Post is kind of a relic of the old way of doing things in a "batch" manner rather than an interactive way that is more possible now that we have faster processors. I still like it, though, because I wrote the darn thing and know how to use it, and usually when I need something done I need it now - Not days later, after figuring out how to do it in Combustion!

- Why did you decide to leave the Max team?

Whew. The bottom line, and most honest answer, is probably "burnout".

In the early days of 3DS, Dan Silva and I were coding machines. I loved it - Code all day, play around with the software you wrote at night, see what you needed to create the cool stuff you wanted to do, make a list and do it again the next day. It was awesome to make the kind of images that were possible as our software got better and better. For me, the success of doing this kind of work was its own motivation.

There was only one problem with this scenario - With just Dan and me working on the software, when someone wanted a new feature we were the ones on the hook to write it. Then I came up with the IPAS plug-in feature and others were able to create the add-ons, theoretically giving me the time to do what I wanted, which was to make movies with the thing.

I say theoretically because as life goes on, there are other things that start competing for your free time. I got married and we took on an old house that needed a lot of work, and being a hands-on kind of guy, I started putting time into that as well, and my personal animation work got put on the back burner.

A big part of the problem for me was that 3DS was always just one or two features short of what I needed to create what I was envisioning, so I kept working on it, thinking that as soon as it had this feature or that feature, I'd be ready to do what I wanted with it.

In the meantime, I was still coding on it (at this point, 3DS Max) and in addition to the four Yost Group programmers, we had a number of Autodesk programmers working on it. Max is an extremely complex piece of software and there would be times where I'd have a bug show up on my list and I had no idea where it had come from. I'd be on a tight development schedule and would divert my time to locating a bug, and sometimes it would take days to track it down because of the complexity of the code. In one case, I remember a series of bugs that cropped up for all the Yost Group programmers and they were all due to some work one of the Autodesk guys had done to some critical code without checking with Rolf (who had designed a lot of the internal pipeline) first - It broke all sorts of things, and we had to scramble to fix it, all the while falling farther behind on the other work we had to do. It was a huge source of stress.

I think that for me, the thing about programming that makes me love it so much is taking nothing more than some simple instructions and building something really beautiful out of it, or a cartoon that people can laugh at, or whatever - At a certain point in the development of Max, some obstacles came up that took a lot of the fun out of it and built up some serious pressure on me that I didn't need.

- What have you been up to since leaving the Max team?

In 2001, I ran for a seat on the city council in Port Washington, Wisconsin and won - I've been serving as an alderman (now Council President) since then.

I'm still working on Max plug-ins and little by little have been doing work on my Klanky the Robot animation. Trying to do this stuff by yourself is tricky - I have the majority of the first short complete but still need to do the audio and some funny Pixar-style "Outtakes". I found out there's a reason movies like "The Incredibles" have so many names in the credits!

In the past year, I got involved doing some incredibly fun development for the iPhone OS. I got a chance to work with a programmer I worked with back in the Atari days, who I had also hired to write the resource editor we needed for 3DSr1! It made me realize again just how much I love programming.

- When will we be able to see your Klanky shorts?

Well, I have a deadline for renewing the "Klanky the Robot" trademark that is coming up in May 2010, so I want to have the first short completed before then. I'll probably make it available as a downloadable film on the iTunes App Store.

- What production work have you done with 3D Studio / Max?

Jon Bell and I collaborated on some work for Terminator 2: Judgment Day - Some of the graphics seen on the computer screens at the Cyberdyne Systems Corporation. Jon Bell is actually in one of those scenes as the camera moves past some guys working at their computers.

Some of the work we created for Terminator 2 wasn't used on that film, and we were surprised to see it turn up in True Lies a couple of years later - Again, some graphics on computer screens.

Jon and I also worked on Honey, I Blew Up the Kid - Again, some computer screen graphics for the Sterling Corporation and Rick Moranis' shrinking/enlarging machine.

I did something like 12 different shots for DS9, as I mentioned.

Also, I did some advertising graphics for a producer in Kansas City and for Channel 41 in Kansas City, around 1992. Nothing earth-shattering, but one project was kind of fun - A fly-through of a huge model of an outdoor shopping center where the edges of the buildings are outlined in Christmas lights - I modeled up the scene then wrote a custom IPAS plugin that added the little multicolored lights to the appropriate edges of the buildings. That saved a lot of time!

- You participated and saw the birth of many socially-oriented CG activities and trends: CG art competitions, sharing of scripts, creation and distribution of 3D clip art, user forums, etc. How has the net changed 3D related social activities and CG software development?

There is really no way to adequately describe what the Internet has done to help move CG forward! I think back to the days when we were starting the Cyber Studio work on the Atari machines, and there were very few readily-available sources of information. I still have my Principles of Interactive Computer Graphics book on my shelf, which was the foundation of much of my early work. I remember running into problems that had no obvious solution, and trying to figure out how to solve them sometimes took weeks of research. Today, I'd simply Google them and have a solution in seconds!

One problem I ran into back then was a really simple one that had to do with perspective correction on textures, and when I found someone who had the solution, he wanted hundreds of dollars to tell me what it was! I went off and found the solution myself, but today people are sharing a lot more than that for free on the Net. That free exchange of ideas is invaluable. And it helps all of us.

- Have you been to SIGGRAPH recently, and what's your theory on the fact attendance keeps on going down?

I went to the Boston SIGGRAPH a few years ago, which was a lot of fun. The only reason I can think of that would be reducing attendance is that so much of the information I used to go to SIGGRAPH to get is now available on the Net! Sure, there are some cutting-edge papers and things that you'll get at SIGGRAPH, but for the most part, many people can find what they need without having to attend a conference like that. I always loved going to the Electronic Theater presentation there, which was always inspirational, but it seems like now that you have YouTube and other ways of watching videos online, even that has lost a bit of its drawing power.

- Do you feel that the graphics field is as exciting as it was 10-20 years ago, or has innovation basically slowed down these past few years?

It's absolutely as exciting, if not more so! Recently, a friend pointed me to a short film called The Third & The Seventh, a photorealistic CG masterpiece done with Max which was basically created by one person, kind of like my Cornerstone was, 20 years ago. In forums where the work was posted, there are skeptics who refuse to believe it's CGI!

The fact that this person was able to create something so absolutely photoreal in many places is incredible, and it is exactly what I had dreamed of years ago when starting on this path.

Back in the early days of 3DS, we were trying to get our code efficient enough to generate reasonably-realistic images in a reasonable amount of time, so we took all sorts of shortcuts and used various cheats (many gained from attending SIGGRAPH!) to get there. The machines we were using had 32MB of memory, which we thought was heaven itself, but compared to machines today, it's nothing.

The evolution of hardware continues to deliver faster and better ways for artists to work with 3D, and in a kind of endless cycle, this enables the software to get better at delivering photorealistic results, because the speed-enhancing cheats that compromised image quality are no longer necessary. Take a look at The Third & The Seventh and you'll see what I mean - The illumination of the scenes and realism of some of the shots is breathtaking.

I think the synergy of hardware, software and artistic talent is where the real excitement is today, and I don't see it slowing down at all.

- Gary Yost mentioned choosing the name Cyber Studio for Antics software because the company was creating tools for "graphiker pilots to voyage into their imaginations". How has the profession of CG artists evolved from the early days of 3D Studio to these days?

That's so hard to nail down! From my vantage point, which is going to be biased, I see something interesting, because a number of the CG artists I know (Andy Murdock, Michael Spaw, Darrell Anderson, for example) seem to have started out in more traditional media but when they got into CG, they started working with scripting and some actually started writing C++ code to push the tools to where they needed them to be! In the early days, artists were dependent on the coders (as with when Dan and I were coding 3DS DOS) to do everything - They were at the mercy of the software providers.

As soon as scripting languages, shader languages and plugin SDKs were made available, the really innovative and motivated artists jumped right on them because they saw them as an extension of their toolset. It's the term "renaissance man" brought up to the 21^st century - I'd easily see someone like Leonardo da Vinci coding away in MaxScript!

This is more exciting than ever for me - I remember Darrel Anderson, who worked with us back on the Atari software, showing us some of the awesome models he built with our simple scripting tools - Animated, rippling manta ray shapes, things I would never have thought of doing. It's why you want real, honest-to-god artists working alongside the technical geeks. It makes us look better than we are.

I say it all the time: I look forward to seeing what you create with our software - And that is absolutely true. From day one of 3DS, I've been thrilled to see what creative stuff our users come up with. The more tools we put in their hands, the more they surprise me.

- You worked on many areas of 3D development including rendering, stereoscopic imaging, scripting languages, APIs/plugins, and more. What current advances in these areas do you find interesting? What interesting developments in CG do you see coming up in the future?

Well, the most recent example, probably an obvious one, is the film Avatar. Back in the Atari days, we introduced the stereo vision add-on for the Cyber Studio software, and we learned a lot about stereo vision and how to make it work. When we were working with the technology, we learned a lot about what to do and what not to do. We were a bit ahead of our time with the technology, but I remember doing some testing modeling with the 3D glasses on and it was awesome to work with a mesh in 3D. Very natural.

Back then, if a movie was made for stereo vision, you invariably walked out of it with a headache because the technology was crude and inconsistently set up from one theater to another. Plus, those "gimmick" films loved to push the stereo effect too far, with stuff coming out of the screen and causing massive eye strain.

James Cameron obviously "gets it", because Avatar was simply a movie that happened to be shot in stereo. No over-the-top "coming at you" stuff that would cause eyestrain - It's all very restrained, and the new projection technology has eliminated the variables of setup problems at individual theaters.

The last time stereo moves became popular, it proved to just be a short-lived fad, but we'll see where this goes. In a world where movie theaters need a product that will draw in viewers who often have their own big-screen HDTV at home, maybe the stereo films will have more staying power. I know of a number of stereo CG film projects coming up; and films like Avatar will hopefully push the envelope a bit more so that we get more and better results.

But beyond the relatively simple technical challenges of stereo presentation, the big deal about Avatar was that the synthetic characters have done a great job of crossing the "uncanny valley" and becoming believable. That's a huge accomplishment, and it's set a new standard for CG films..

Before its release, I had my doubts about Avatar's appeal to the general public. But it's now clear that a film that's largely CG can rise above simply being a cartoon.

- What one piece of advice would you give the team currently working on Max in Montreal?

To always look toward making the software as flexible and robust as possible. Right now it's largely there, because plugins can do just about anything. Max is the ultimate expression of what we started to do back in the IPAS days with 3DS DOS - Open the platform up so that it can be infinitely expanded and improved by others.

- We've heard rumors of a Greeble Plus with way more control over panel types and such. Any truth to those rumors?

Yes! I have a new Greeble version that I've been puttering with for several years now - Much more extensible and flexible - The problem is getting focused on it long enough to finish it up. The original was written as a favor to Jon Bell, who needed something extra for his first 3DS Max book, and it was pretty limited, though people seem to enjoy it. Naturally, after playing around with it, I saw a number of places where it could be improved, and started toying around with some changes that would make it better at things like creating cityscapes and such - Better texturing controls, etc. One of these days I'll set aside some time and finish up the initial release.

- Any interesting or funny anecdote that you recall from 3DS / Max development?

Just an example of the "by the seat of our pants" attitude in the early days of 3DS:

When we started working on the 3DS project, it was just me coding for the Shaper, Lofter, 3D Editor, Materials Editor and Renderer modules - Then we brought in Dan Silva to create the Keyframer. Initially, all of these were separate application programs that were chained together because the memory constraints of DOS were so limiting. I hated it - It was really slow and clumsy and each module had to save out the data it was working on in order to hand it off to the next module.

On top of that, we had the problem of a lot of duplicated code between the modules for the GUI and such, and anytime we changed the GUI in one module, we'd have to make sure the same change was made in every other module that had the same functionality - It was going to be a maintenance nightmare.

Shortly before we were due to release 3DSr1, we found out that Autodesk was using a product called Phar Lap DOS Extender, an operating system that was specifically designed to overcome the limitations of standard DOS. Using Phar Lap would allow us to create one application executable with all the modules contained within, making the program faster and vastly easier to maintain. We were just a couple of months away from release, but we made a fairly high-risk decision to take the plunge and convert to Phar Lap.

I flew out to San Francisco and spent over a week at Dan Silva's home office and we ripped the program apart and put it back together under Phar Lap. Any number of things could have gone wrong in this major conversion, but we pulled it off! Other software at the time was hobbled by a "separate module" architecture but having 3DS run as an all-in-one solution, it really made for a smoothly operating package that felt like an integrated whole.

Looking back on it, nobody in their right mind would attempt such a radical change so late in a product's development cycle!

- Anything you'd like to mention to readers that we haven't asked you about?

Just that I'm always happy and excited to see what people are creating with 3DS Max. 3D Artists rock.

We would like to thank Tom Hudson for his great generosity and invaluable help in preparing this article. He has gone far beyond what we asked for, joining us in our research, and adopting it as his very own. The quality of the article is the result of his work and enthusiasm. We would also like to thank José María De Espona, Miguel Gómez (Dimensión 2.5), Jorge Seva (Alex Roman) and Richard Rosenman for their help and generosity. A special thanks goes to Martin Doudoroff, without whose excellent documentary work on the early history of 3D Studio much of this article would’ve never existed. In Yost we trust.

Part I: Tom Hudson interview
Part II: Gary Yost interview
Part III: De Espona on the Max Temple image