In depth: Autodesk Project Skyline
Autodesk looks to build the future of game development with Project Skyline, first shown at GDC 2011
The GDC demo of Project Skyline shows the game trace view, where individual animation clips can be seen on the timeline
Without a vision, the people perish. It’s a situation that can even affect digital media creation software companies like Autodesk; or perhaps that should be, especially affect companies like Autodesk.
As a giant of the industry, the company has been acquiring major rivals such as Alias and Softimage over the last few years, resulting in a more tempered product management approach to the major 3D applications.
But the big fish of the pond is running out of water to swim in. Where does Autodesk’s vision lie?
Game development tools
The market for game development tools is huge, and while many developers provide game engines, physics, networking and artificial intelligence, Autodesk, has almost no experience in real-time technology.
This is one reason why Autodesk has been buying up middleware companies such as Kynogon and Scaleform. (See The Scaleform Factor)
With the character running in the game window, a new animation is added to the Project Skyline blend tree
Yet Autodesk is bridging the gap between content creation and runtime engines that is the main prize, offering massive potential both in terms of productivity gains for game makers as well as commercially for the tools supplier. Ah. Now there’s a vision.
In this context, it’s significant that Mathieu Mazerolle, the project manager of Autodesk’s Project Skyline initiative, has come to the company, following stints at game publisher Ubisoft and special effects house Digital Domain.
Just as Skyline, which was first publicly shown in an animation demo at the Game Developer Conference, is an attempt to bridge that gap, so his career has been spent working across it.
“Based on my experience working on Assassin’s Creed, the artist makes these lovely character animation clips, which need to be assembled into blend trees and combined with physics and IK,” says Mazerolle, explaining
how animation works in the traditional game development process.
The finger of blame “By the time the programmer has stitched everything together and the artist gets a chance to see the integration in the game engine…” He pauses. “This is where we start what I call ‘pointing the finger of blame’.”
Not only can this phenomenon be detrimental to team morale, it’s also a key obstacle to the final quality – not to mention production efficiency – and hence a drain in terms of overall cost and timeframe.
“The challenge is providing transparency in terms of how different elements work together,” Mazerolle says.
“The artist doesn’t have visibility of how the programmer is
ing his animations, and the programmer doesn’t have visibility of the artistic intention.
With Skyline, we want to bridge this situation so the artist can see changes instantaneously, but also to create feedback that enables the game engine to talk back to our tools.
Demonstrating the power of real-time interaction, this screen shows the ability to modify a source animation in the left-hand window while the game is live; the results immediately show in the right-hand window
Being able to see what’s going on within the runtime environment, the interaction with elements such as game logic and AI that the content creation tool doesn’t create, is the key innovation for Skyline. It’s a really powerful type of workflow.”
As with many technological innovations, it’s this point of intersection and interaction between two previously separate areas that could change everything; and a sweet spot Autodesk feels it’s uniquely placed to solve.
While the GDC demo only dealt with animation, Project Skyline’s ultimate goal is to handle all elements of the development process.
“This is an issue that people in the games industry are familiar with on a day-to-day level, but they don’t have a lot of time to solve it,” Mazerolle suggests.
“But it is logical for Autodesk. If we were lazy, we would go on making tools for artists and not worry about the programmers and level designers, but the reality is we care about game development as a holistic process, and one that involves people who don’t [only] create polygons.”
The underlying framework to Skyline is a node-based visual programming environment, codename Amino.
It’s influenced by Softimage’s ICE (Interactive Creative Environment) technology, although that influence is described as being more of a brain trust than actual code.
“We were able to retain a lot of the workflow and user interface, and recoup a lot of work at the conceptual level, which is why we’re been able to move really quickly on this project,” Mazerolle says of ICE.
“The lessons you learn building a system such as ICE are lessons you learn about how people work with visual programming in general.”
That Vision thing
Project Skyline hopes to replace the iterative asset creation process with a closed loop system
It was a similar situation in terms of choosing a game engine to integrate into. Autodesk could have built its own, which would have been easier in terms of creating dedicated hooks, but instead it took German provider Trinigy’s cross-platform Vision engine.
The point to prove was that the technology would work with any off-the-shelf engine. “It was a harder road,” says Mazerolle. “We used Vision as a black box solution, without any modifications, which wasn’t always easy.
“But the reason we held ourselves to that decision was to test our approach. Over 80 per cent of our customers use custom-built engines, so you have to ensure you provide the right pieces, and that you don’t have too many wires sticking out.”
Running Skyline in Maya was another pragmatic choice to get up and running quickly. “We needed a UI, viewports and tools that we were familiar with. Maya is the most open platform. It’s not an anti-Max thing,” he says.
Yet for all the excitement following the GDC unveiling, Mazerolle is keen to point out that Project Skyline is in its earliest stages.
“The level of excitement took me by surprise, and when people get excited, the first thing they want to know is ‘When can I have it?’.”
“We’re not at that stage yet,” he admits. “We want to make sure we’re doing the right thing first. Our product development process is very deliberate. We take it in small stages and make it transparent, and get people to kick the tyres, but we’re currently at step one.”
Indeed, Mazerolle expects plenty of changes, perhaps even fundamental ones, before Project Skyline loses its prototype status and, presumably, its codename.
“You shouldn’t have any sacred cows when you’re trying to build a good architecture,” Mazerolle says. “The ultimate test is always the customer.
“No matter how well you build the foundations – and we’ve built strong foundations – I think once this gets into the hands of users, we’re bound to change our assumptions. We’re prepared for that.”
The road ahead
Still, at this stage of the process, it’s not the technology that’s the important thing: it’s that vision thing.
“Part of showing Skyline was to fire people’s imagination,” Mazerolle enthuses. “That’s what’s really exciting for me.
“Speculation about where it will be going after animation is good in terms of our next step, although it might be years off. It’s good to stir the beehive sometimes, and scaling to deal with bigger problems is where Skyline becomes more than a little project.
“If we’re on the right track, it’s something Autodesk can have faith in.”
A run down of Autodesk’s Scaleform acquisition is provided on the next page
on Thursday, June 9th, 2011 at 12:53 pm under Analysis, Features.
You can subscribe to comments.
You can leave a comment, or trackback from your own site.
Tags: Autodesk, game authoring, game development, Games, Maya, Project Skyline, scaleform