VizWorld.com

While at SIGGRAPH, I had the opportunity to sit down and talk to Autodesk about some of their new offerings, the 20th Anniversary of 3dsMax, and the ‘rivalry’ between Maya and Max.  All in all, Autodesk doesn’t look to be shedding the title of ‘Behemoth of the Computer Graphics Industry’ anytime soon, and they’ve got lots on their plate to prove it.

First off, and central to many of their SIGGRAPH events, was the 20th Anniversary of 3dsMax.  Autodesk held a special lunch event with Tim Miller (owner of Blur Studios) and some Autodesk higher-ups to discuss the last 20 years of the product, and where they plan to go from here.  You can hear Tim Miller talk about 3dsMax in a special video at The Area, and his presentation at the event was similar. However, he showed some of the work he’s done over the last 20 years and his personal experiences with the product.  I personally loved hearing him talk about being briefly hired by a “major studio” prior to Blur, and quitting in frustration with how poorly the “Industry” tools on SGI workstations performed in comparison to 3dsMax on a wimpy little PC.  He concluded his talk with the amazing Star Wars: The Old Republic trailer that Blur created.

The other thing Autodesk was eager to talk about at SIGGRAPH was their new “Suites” products.  Similar to Adobe’s Suites, now you can buy collections of Autodesk products in a Suite which cuts the price (obviously) and adds guaranteed 1-click interoperability between applications via the FBX systems.  Their new Entertainment Creation Suites shown here, give you your choice of 3dsMax or Maya, along with SoftImage, Motion Builder, and MudBox.  Whether you choose 3dsMax or Maya, the price does not change and they guarantee the same 1-click interoperability between all of the apps.  That means you’ll be able to click a model inside 3dsMax or Maya and click 1 button to have it immediately transferred over to MudBox for refinement, and then click a button to send the results back to 3dsMax or Maya when you’re done.  It’s a whole new level of interoperability not previously available without extensive plugins and file transfers/import/export mundaneness.

This leads to the last point that you really had to dig to get to at SIGGRAPH:  Neither 3dsMax nor Maya are going away anytime soon.  Several people (myself included) had expected that Autodesk’s acquisition of Maya would mean the end of the product as Maya’s features merged into 3dsMax.  Not the case, as both products have continued on parallel and unique development paths since the acquisition, and Autodesk still has plans for many new features for both products.  The fact that you can get the suites with either case at no cost difference seems to really drive home the point that Autodesk simply sees them as two tools to accomplish the same task : Act as the Central Hub for all of their other products.  Model in MudBox, animate in Motion Builder, Render in Mental Ray, but link it all together in 3dsMax or Maya.  Both products now offer identical capabilities, but expose them in different ways based on their historical audience.

I asked some Autodesk engineers about this, and they echoed the sentiment.  If they tried to make 3dsMax more like Maya, users would probably leave for Cinema4d.  If they tried to make Maya more like 3dsmax, Users would probably leave for Houdini.  Leaving the two products along, but unifying them “under the hood” to reduce development time, lets each product play to it’s strengths and lets users remain comfortable in the environment they’ve spent the last 20 years working in.

So what’s in store for the products?  The main thing it seems we’ll be seeing is an extensively redesigned UI.  They admit that the current UI’s have gotten a bit cluttered with way too many rollouts.  They currently have a project underway to redesign the GUI (some of which has already happened in the last 2 versions of 3dsmax) to be more streamlined, and have such lofty goals as a 20s load-time (Wouldn’t that just be AWESOME?). They are also embracing new computer architectures and working to add more features for multicore/multiprocessor systems, taking advantage of those big quad-core/hex-core chips on the way.  But that’s not all, they’ve got a long list of things they’re working on, and I guarantee we will all love every bit of it.

If you were at SIGGRAPH & talked to Autodesk, what all did you see that interested you?

Graphics 3dsmax, autodesk, feature, maya, siggraph

One company I always enjoy talking to at SIGGRAPH is Lightworks, creators of an impressive suite of software you may have never heard of.  Many people are largely unaware of the Lightworks tools because they market them not to end-users, but rather to other businesses and application developers as a pre-built rendering solution.  The result is high-end photorealistic visuals (or not, depending on the application needs) in several applications, powered by Lightworks technology.

This year at SIGGRAPH, I took the time to talk to the Lightworks crew and see what new offerings they had.  Lightworks has been interested in the real-time raytracing market for quite some time, and I first found them when they were demonstrating their integration with the CausticOne card back at SIGGRAPH2009.  They were back again this year, again demonstrating the amazing CausticOne integration capabilities, but they had really taken it to the next level this year.

One technology they were demonstrating, but not being terribly vocal about, is a new system called “Lightworks Architect”.  The Lightworks Architect is, at it’s core, an interface layer between applications and underlying rendering technologies.  This means that using Lightworks Architect you can program your application using a single interface, and then swap out renderers underneath.  At their booth they were demonstrating various options using the CausticOne card, NVidia’s OptiX & SceniX, an Intel CPU-only renderer, and their own renderer.  This alone is impressive, but they also had support for OpenRL, meaning that they could interface with any OpenRL compatible renderer.  Unfortunately, there aren’t very many of those right now but with the involvement of Caustic hopefully that will change.

From speaking with their developers, adding new renderers is a breeze as well.  Merely writing a translation layer and providing a single DLL for the application was enough, and they were able to adapt NVidia’s OptiX raytracer in a mere matter of days, and now it’s available for all to use via Architect.  Architect isn’t a fully released product yet, more of an in-house tool, but hopefully this will become another product in the Lightworks Portfolio, enabling real-time raytracing on a variety of platforms.

In addition to Architect, Lightworks got a lot of buzz from their new ‘Cloud Rendering’ tool named Alto. The hype was a bit off on this one, as Alto is not a true ‘Cloud Rendering’ system, but rather a remote rendering & queue system, similar to Autodesk’s Backburner.  What Lightworks found was once they had Architect in place, it became trivial to write a renderer that simply bundled up the scene & handed it off to someone else across the network.  They constructed Alto to allow you to setup your own Lightworks-based Render Farm and use a cluster of computers to accelerate your renderings, or perhaps use a single extra-powerful machine by multiple people in an architectural or design house. Complete with queue management & control features, it looks like a nice resource sharing system for use by small studios.  Again, this is a technology that will be offered to developers for inclusion into their own products, so Lightworks-driven products may find themselves with some nice remote rendering capabilities in an upcoming update.

These types of advancements really showcase the power of embracing a solution like Lightworks, rather than “rolling your own”.  When a company embraces a solution like the Lightworks tools (or Luxion, or Fovia, or whatever other OEM/B2B provider you can think of), they wind up getting all of these new functions without dedicated their own software resources to creating them.  Application developers can focus on file-formats, GUI interfaces, and hardware interfaces to remote sensors or exotic input systems like multitouch, haptics, or tablets, and leave the Rendering to a team of experts offsite.  Then as new accelerated renderers come along, or support for clusters and remote resources, you get access to that technology without any effort on your part.

Where’s Lightworks going next?  Well I can’t spill those beans (yet), but suffice it to say they’re keeping busy.  Around the show you may have seen a few people demonstrating the Lightworks Artisan product, which is a popular tool for viewing models.  In their booth they were demonstrating an internal product, not yet available to the masses and not confirmed if it ever will be, along the lines of BunkSpeed Shot and Luxion KeyShot which would allow you to arrange a scene and model and then “Take a Picture” to initiate a full-resolution rendering.  This was one of their demonstration platforms for Architect, as they could then show how similar the renderings were between OpenRL, OptiX, and their internal renderer, therefore showing how well Architect translates between them all.

Architect and Alto both look like great products, and hopefully they’ll have more coming down the pipe soon!  If you made a stop by the LightWorks booth or presentations, what caught your eye?

Graphics feature, lightworks, siggraph

Update 8/5: By request from Khronos, I’ve added all the little ™ and ® ‘s.

Last week I posted a rather, well let’s just say “sensational”, article about the coincidental announcement by AMD/ATI of their new OpenGL|ES2.0 Driver for Desktops and Khronos’s announcement of the OpenGL® 4.1 spec which offers full backwards compatability with the OpenGL ES 2.0™ standard.  Most people wouldn’t care about OpenGL ES 2.0™ on the Desktop, as it’s the OpenGL Spec for Embedded Systems like set-top boxes and mobile phones, however the OpenGL ES 2.0™ spec is the foundation of the up-and-coming WebGL spec that promises plugin-free 3D graphics on the internet for all to enjoy.

Currently in the WebGL space, you have about 4 options:

• Download a Plugin.  This is basically what Google started out with, you download a plugin that offers a translation layer between the actual hardware support and the WebGL spec.
• Download an OpenGL ES 2.0™ Driver. This is what AMD/ATI Announced just prior to SIGGRAPH.
• Download an OpenGL® 4.1 Driver. This is what NVidia announced during SIGGRAPH.
• Download a new browser. WebGL is currently supported in some fashion or another in the latest dev releases of Chrome, Safari, and FireFox.

However, each of these has problems.  The whole point of WebGL is to be “plugin-less”, so the first one is out.  This leaves the other three, however, they have hidden issues as well.

Before I continue, many people ask “Why do I need a driver?”, and it’s a valid question.  Right now, without any OpenGL ES 2.0™ driver, you can go download development versions of Safari, Chrome, and FireFox and get WebGL.  It all works just fine, but you’ll notice it works a bit differently in each one.  This is what the driver is for: consistency. With working drivers in place, the visuals will be identical across browsers and hardware, because the rendering is all handled in the Driver, not the Browser.  Currently, browsers have a built-in translation layer that turns the JS-based OpenGL ES 2.0™ commands and turns them into regular OpenGL commands, and some do a better job than others.

Initially, as in the article, I was a bit harsh on ATI for releasing a dedicated OpenGL ES 2.0™ driver and favored NVidia’s announcement of an upcoming OpenGL® 4.1 driver that would encompass the same results.  This way, the user only has 1 driver to manage that in future systems will be installed by default, so the user literally has to do nothing.  Unfortunately, NVidia’s OpenGL4.1 support is limited to only the latest revisions of hardware:

You will need any one of the following Fermi based GPU to get access to the OpenGL® 4.1 and GLSL 4.10 functionality:

• Quadro Plex 7000, Quadro 6000, Quadro 5000, Quadro 5000M, Quadro 4000
• GeForce GTX 480, GeForce GTX 470, GeForce GTX 465, GeForce GTX 460

This means all those people with little GeForceM cards in their laptops are out of luck, as well as anyone with the GTX285 or earlier.  No doubt these cards have the horsepower to handle WebGL, but they’re currently unable to get the drivers necessary.

ATI’s solution is a bit less elegant, but by offering a dedicated driver they open it to all of their hardware, not just the latest and greatest.  Unfortunately, it does return us a bit to the previous world of loading plugins, except you’re loading a system-level driver instead.  However, this opens the world to all those old ATI Rage chipsets in laptops and FirePro’s in the wild, covering the full gamut of users.

In the end, I’m sure NVidia will offer a driver for WebGL to older hardware, but there’s no news on when that will be.   If NVidia lags too far behind, we could find ourselves in a “VRML Situation”, where individual browsers begin to support various extensions in attempts to best utilize the hardware, leading to inconsistencies and incompatibilities we already see with HTML & CSS across browsers.  Hopefully, with a good standards organization in place like Khronos, which VRML didn’t have, we’ll find consistent drivers coming to all platforms soon.

Hardware, Science amd, ati, feature, khronos, nvidia, opengl, webgl

So, you’ve just finished the week/month long task of modeling your huge environment.  You’ve got beautifully detailed architecture looking over expansive plains, beautiful HDRi lighting and environment maps, perhaps some of the most beautiful architectural landscapes ever created.  But it’s completely lifeless.  There are no people, no cars, no signs of life.

The “simplest” solution is to simply add in some vehicles.  Car models are a dime a dozen these days, but animating a car to look lifelike is a nightmare.  Believe me, I’ve tried it.  First you need the drive path, that’s easy enough.  Then you need the wheels to spin, the car to properly lean around corners, proper acceleration and braking maneuvers.  A simple “drive by” can take days to animate realistically.  Just imagine the pain of trying to animate something more complex, like a video game intro complete with screeching tires and power-slides.  Vehicular animation is one of those things that is easy to do poorly, but painstakingly difficult to do well.

Enter Craft Animations and their amazing tools for 3dsMax, Maya, SoftImage, and Cinema4D.  Available as a plugin for the system of choice, they turn vehicular animation into something as simple as a video game, literally. Their product is something so naive and intuitive that anyone can pick it up and become a vehicle-animating expert in mere seconds.  Simply load in one of their “rigs” (4-wheeler, 2-wheeler, airplane, helicopter, etc) and attach it to your desired model (Match up the Wheels, the Axles, the Chassis, etc).  Then, attach any game input device you like (USB Gamepad, joystick, or just use Keyboard and Mouse) and click “Record”.  Then Drive.  That’s it.

As you drive, all the action is processed internally by their own game physics engine which allows for realistic drifting, gravity effects, inertia, and more.  You simply drive around the scene as it records all the details of what’s going on as a keyframe animation, suitable for editing in your own animation studio if you so choose (or for binding additional actors and objects to for later ragdoll or impact physics).  If you mess it up, so what. Just click “Erase” and “Record” again, and try again.  You can rewind to the last “good” point and record from there if you so desire.

The end result is that you can create realistic vehicle animations in mere seconds that would take you days or weeks otherwise.  Don’t like the first one? Do another, it only takes another 30s.  With the realistic physics, you may even find effects and tricks that you hadn’t even considered (extra draft on a helicopter during a turn, power-sliding around a corner at high-speeds) that lend additional realism to your animation with no extra effort.  But the vehicles aren’t all, Craft offers several other things you can add to your package to increase the realism even further:

• Trailer Addons – turn your truck into an 18-wheeler complete with a full payload, or add new realism to that oil tanker before it flips and rolls down the highway.
• Crawler Rigs – Realistic tread & crawler simulations make tanks and ATV’s a breeze
• Missile Addons – realistically fire and track missiles from your helicopters and tanks
• Fully Rigged Models – Need to add a vehicle in a hurry?  Then use the new Pre-Rigged Models feature and import models from agencies like Dosch Design and PerspectX that are fully rigged and ready to go, no configuration necessary.
• And many more.

I had a chance to take a helicopter for a spin, and was amazed.  Helicopters are difficult to animate because of the incredible complexity in the flight physics.  There is no forward thrust, only vertical lift so that the helicopter must lean forward to achieve forward momentum.  The unique ability to hover, move directly laterally, and spin in place all create difficulties in realistically animating, but I had a helicopter flying between (and crashing into) some tall skyscrapers in mere seconds.  The potential is amazing, and the simplicity of the entire tool simply has to be experienced to be believed.  A demo version ships with 3dsMax2011, and I highly recommend you try it out.

Whether you work in gaming and build cinematics for high-end games, or work in architecture and simply need to add some generic traffic to a scene, it doesn’t get much simpler (and more FUN!) than the Craft Animation tools.

Graphics craft director studio, feature, siggraph

Here’s some photos of SIGGRAPH2010 taken during my stay.  I’m leaving now, but it’s been a great week.  A wealth of technical content (Computational Photography, Image Statistics, many in-depth sessions on Avatar) and a nice exhibition floor.  The exhibition floor, however, was surprisingly small this year, I would say even smaller than last year’s New Orleans Siggraph.  However, nobody was “missing”, it was just that several companies combined their booth presence into more focused and better presentations in existing booths.

Nonetheless, I’m flying today so don’t expect too many updates.  Enjoy the pictures, and keep checking back for more SIGGRAPH information over the course of the next few days as I review my notes.

Graphics feature, siggraph

Since the Fermi chipset came to reality in the GTX480 card, high end graphics professionals have had a difficult choice.  They can take the latest and greatest technology, the GeForce GTX480, or go back a generation for a QuadroFX 4800 (or similar).   Several of the newer CUDA applications for professionals, like the Adobe CS5 applications, only works on Quadro cards.  The added performance of the Fermi chipset, tho, makes it worth it in certain applications.  It’s been a difficult decision time for Quadro fans.

Today, NVidia has announced that this is a difficult decision no more.  They’ve just announced the newest generation of their popular Quadro cards using the newly designed ‘Fermi’ technology and combines it with the rest of the Quadro ecosystem to bring to bear a card that truly is top of its class.  The core of the new offering is three new Quadro Cards:

• Quadro 4000 – 2GB, 256 cores
• Quadro 5000 – 2.5GB, 352 cores
• Quadro 6000 – 6GB, 448 Cores

Yep that’s right.  The Quadro 6000 sets a new record for video memory with 6GB available on a single card.  In addition to these three, there will be a new Quadro 5000m design, which packs the regular Quadro 5000 into a mobile form factor suitable for high-end laptops without compromising functionality.  Also, if you really need power, you can get the new QuadroPlex system which contains 2 Quadro6000′s for truly amazing power.

NVidia gave me one of the Quadro5000′s to review, and you can read my results below.
Read more…

Hardware feature, fermi, nvidia

This afternoon I took in a panel discussion entitled “Future Directions of Graphics Research”.  I had expected a panel of experts going into blue-sky visions of research so mind-bending it would leave us all raving lunatics, but instead I found something much different.

The Computer Graphics industry is suffering from its own success.  Recent smashes like Avatar, Toy Story 3, and others have several people thinking that computer graphics is ‘done’, there’s no more research to do.  The technology has matured to the point where we can not only realistically create digital actors, but completely make up alien planets and worlds.  What is possibly left to do?

Such thinking is beginning to impact researchers and academics financially, as government grants are becoming more and more scarce.  Shrinking government budgets doesn’t help things, so the academic community has decided to come together and write up a lengthy report to the National Science Foundation (NSF) detailing areas that still need research.

Read more…

Science feature, siggraph

Today began the first day of SIGGRAPH 2010 here in sunny Los Angeles.  The schedule is a bit different this year than previous years, and I thought I’ld share some of the changes.

Perhaps the most visible change is in the Schedule.  Sunday has traditionally been nothing but a registration day, but this year contains tutorials, panel discussions, and talks all afternoon.  To offset this, they’ve completely eliminated the final day of the Conference (Friday), so the Conference runs now from Sunday-Thursday.  Friday has always been a pretty dead day at SIGGRAPH, and this is a clever way to rearrange the schedule around it.

Also this year there is no special ticket for the Electronic Theater.  Showing are every night starting Tomorrow, but the times vary a bit each night (Monday it’s 6-8, but Tuesday it’s 7-9).

From rumors I’ve heard around the conference, it sounds like this year is another unusually low attendance year.  Attendance is down, and the exhibition hall is a bit small as well, compared to previous years.  However, this isn’t impacting the volume of press releases coming out of the event, as I’m currently being inundated with several each day of the show.

Today attendance was sparse, but it’s a Sunday so it’s a bit early to say.  Keep checking back for regular updates from the floor!

Graphics, Science feature, siggraph

A short while back I found myself at Club Auto Sport, an interesting tech incubator for automotive companies, attending a mixer hosted by NVidia.  While you may not think NVidia and Cars have much in common (Aside from the Tegra/Audi collaboration), one of the companies there finds NVidia products a key component of their product.

Jack Ulstad runs a business there called ‘Trinity Racing Concepts’ that designs high-end reconfiguration racing simulators.   Jack and his team have build a highly flexible reconfigurable frame, called the ‘Revolution Frame’, that can quickly and easily be converted into everything from NASCAR to Formula 1, providing a wide range of simulations.  They combine their frame with a 3-screen stereoscopic display powered by NVidia 3D Vision technology, force-feedback steering wheel, and custom force-feedback seat.

Read on after the break.

Read more…

Hardware feature, nvidia, simulator

With Toy Story 3 in theaters and continuing to set records for animated films and Pixar productions, this week’s recommended resource is a celebration of some of Pixar’s greatest work in the Short Films area: ‘The Art of Pixar Shorts“.

While Pixar Animation Studios was creating beloved feature-length films such as Monsters Inc. Ratatouille and WALLïE it was simultaneously testing animation and storytelling techniques in dozens of memorable short films. Andre and Wally B proved that computer animation was possible; Tin Toy laid the groundwork for what would become Toy Story; and Mike’s New Car exposed Pixar’s finely tuned funny bone. In The Art of Pixar Short Films animation expert and short film devotee Amid Amidi shines a spotlight on these and many more memorable vignettes from the Pixar archive. Essays and interviews illuminate more than 250 full-color pastels pencil sketches storyboards and final rendered frames that were the foundation of Pixar’s creative process.

This book and many others is available in the VizWorld Store.

Graphics feature, pixar, resource of the week