Crytek, the makers of Crysis, discuss how far gaming engines have come, and what is coming next.
We as gamers tend to take technical advancements in games for granted. We expect games to be bigger, faster, and shinier than the titles that came before them, and we tend to dismiss the massive amount of work that goes into the finished product. That is a sadly inevitable fate, especially in an entertainment field that pushes us to experience it on a visceral level.
But behind all those impressive explosions, realistic facial movements, and animations that make destroying things all the more satisfying, there’s a graphics engine powering that game. The best engines make the rounds and get licensed out to several other developers, serving as a blank canvas for game after game. Despite heavy modifications from the developers that license them, the fortunes of games themselves become heavily reliant on that initial development.
Of the developers that make their own engines, one of the most highly respected is without question Crytek, makers of the CryENGINE game engines. The first CryENGINE unleashed Far Cry on the world, an impressive game, to say the least. Not to be outdone, Crytek followed that up with CryENGINE 2, the game engine behind one of the most graphically intense and lauded games ever created, Crysis. The game was so technically intensive that it became a running joke as to whether or not a computer could run it at full specs. Even five years later, many new computers still have problems with it.
In 2009, Crytek announced it would release CryENGINE 3 in October of that year, and then later released a free-to-use engine called CryEngine 3 SDK. The original version of the engine was used on Crysis 2, and then later used to bring the original Crysis to consoles. It is also now being used to power the upcoming free-to-play game, Warface, as well as the hotly anticipated Crysis 3, due out in February of next year.
We spoke with Crytek’s Carl Jones, Global Business Development Director for CryENGINE and discussed the role of new game engines, what we can expect in the future, and the role of tech in video games.
Are you excited for the next, more powerful generation of consoles?
Carl Jones: Yes, we’re very excited about the next generation of consoles. Crytek has always made benchmark-setting, high-end PC games, in terms of the quality of our graphics. What we’re expecting to see from the next generation is a console that allows us to take off all the restrictions that come from relatively low-end current hardware, and even go beyond where we’ve been before on high-end PCs.
As graphics and systems become more sophisticated, does it become more difficult to develop new engines, or do the new tools at your disposal make it easier?
CJ: Generally, new systems mean new opportunities for us at Crytek. We hope, of course, that any new hardware advance actually makes it easier to develop new engine features! The sophistication of DirectX 11, for example, gives us the opportunity to do a lot more low-level work, as we get more control over the GPU. You can actually start doing general-purpose GPU systems, which enable you to write really cool, movie-quality simulations for hair, cloth, particles, etc. and have it run really fast on standard DirectX 11 hardware. There aren’t really tools that enable people to build engines, but certainly ensuring that the operating system is simple and that it’s easy to access the hardware makes it easier for us to do cool new things in updating our engine for more powerful systems.
How important are graphics in gaming?
CJ: Since graphics deliver what you see, it’s one of the core and critical components of any game, alongside audio and gameplay. Whether it’s more important than those other two is a matter of opinion. Obviously you can have a great, fun game without a huge amount of high-end graphics. But if you really want to immerse your player in a believable world and have them cross that barrier into really feeling the game and being a part of it, then the believability of your virtual world is key. And the way you do that is via better and better graphics. CryENGINE still leads the way in terms of the quality of our realism and believability of environments. And so for a large number of games, especially high-quality games, you want your player to be truly immersed in the environment, and at that point graphics are vital.
How far away are we from photorealistic graphics and faces?
CJ: Crytek achieved near perfect photo-realism way back with CryENGINE 2 and Crysis. I think where we are now is that you can actually go beyond photo-realism and get into the kind of dream worlds or CGI-style worlds where you’re not limited by realism. More importantly, you can let your creators’ imaginations go wild, and they can be sure that what they choose to design will look believable.
I think that’s the point we’re at now, where having been through photo-realism for environments, we now want to go beyond that and say, “Well, what if it doesn’t exist – can we make people believe it?” If you think of things like Avatar, for example, they’re not real but they’re certainly highly believable, perfect-looking environments. And that’s where I think games are moving to now. We’re always improving and there’s always more you can do with more power from the hardware in terms of the amount of stuff that you can show on screen at any one time. Achieving photo-realism and interactivity at the same time is always the challenge, and we’re seeing more interactive and larger environments with more and more stuff on screen that looks realistic.
With faces and characters, we’re getting closer. I think, again, to use the example of Avatar, we saw that photo-realistic or believable characters are now here in CGI and movies, and I think we’re really close to that in games as well. The issue that we often have in-game is the limited budget that we have to balance between everything that makes up the game. To make a believable-looking face, with believable movement, on top of a believable looking body – and have a whole game with hundreds of characters and lots of things happening at the same time, all completely under the players control – is still beyond the power we have in the hardware today.
Once you get there, how do you overcome the “uncanny valley?”
CJ: We are now seeing, on small scales, extremely believable photo-realistic character faces. We have them now in CryENGINE 3; the faces of the in-game characters are almost indistinguishable from a photo of a person. Of course, we’ve got a fair way to go in terms of the movement of a character’s face and making that feel real. But I think we’re not far off, and certainly you’ll see some cool stuff from Crytek in this area quite soon. The uncanny valley always gets deeper, the closer we get, but we’re definitely seeing some effects now and some CG characters that make it very difficult to tell they’re not real.
What is the next step in graphics engines going to be like?
CJ: I think you’ll see movie-quality graphics in a fully interactive environment, and I think we’ll be able to do a great deal more with physics simulation for cloth, destruction, fluids and hair. So the world that you’ll be playing in will become significantly more believable and interactive. Having a fully populated realistic and interactive city should soon be possible.
Unfortunately, to do that you need to create a lot of assets and the next real big step for all engines will be constantly improving the tools and the content creation process. So I think, while we’re going to see a leap in the short term in graphics quality, you might not actually see games being able to fully utilize that until the tools catch up and give the development team the ability to create so much content so quickly. People aren’t going to start just putting bigger and bigger teams onto video games in order to take advantage of the new effects. The tool pipeline needs to be improved generally for game engines to allow much quicker content creation to fill these worlds that we would now be able to create. We established the use of real-time editing with Sandbox years ago – all other engines now work in the same way – as instant iteration gives a lot of time back to the game creators. Now we need to look at other ways of removing downtime, increasing productivity through procedural tools and the like.
What is the trick to making a game look great? Is it the art design, the polygon count, the lighting, etc.?
CJ: It’s a combination of having the technology and the people to realize a well-directed vision. So you need talented art direction matched with the ability of the artists to create beautiful 3D art, and then have technology that can handle incredibly detailed interactive scenes. So really, it is a combination that mostly comes down to creative talent, from the artists, the art director and the guys making the engine and the technology.
Are there any technical things you would like to do, but technology just isn’t there yet? How far off are we?
CJ: We are currently working on things at Crytek that are based on future technology that we expect to arrive – they don’t necessarily exist yet! I can’t talk too much about that, but there are things that we already want to do, but we are waiting for the technology to catch up with our ideas and we’re pretty sure that it will happen within the next three years, maybe even sooner.
With CryENGINE 3 we are now achieving things on current-generation hardware, which we’re told shouldn’t be possible – these are features that other developers think require multiple graphics cards in a supercomputer. We’re managing this because of our research strategy, which is to create features that are ahead of their time, then optimize them to run on the lowest possible hardware specs. This strategy also means our engine is better prepared for AAA, next-generation console games – as we’ve been aiming CryENGINE 3 at the new generation since before 2010.
So, some of our research now will only become fully usable once future technology becomes available – but it’s going to be pretty sweet when it does!
