PS4 declassified: How Sony used its PS3 mistakes to build the ultimate developer's console
The PlayStation 3 was a troubled machine, launching at a price that gave gamers fits, and using technologies that gave developers headaches. This is how Sony avoided making the same mistakes twice.
It's the fall of 2006 and Sony has a problem. The
That placed the system out of the reach of many gamers, but it was a cost necessitated by complex, custom internals -- not the least of which being the physical inclusion of an Emotion Engine CPU, the brain that powered the
Cost, though, is something that can be fixed, and Sony solved that problem. The economic implications of volume production are well understood, and video game components are high-volume things. Sony would streamline and refactor the PS3 on multiple occasions, ultimately ripping out that PS2 backward compatibility, and bringing the cost of the higher-end PS3 down to $499 the July after launch, then introducing a new, midrange $399 model that November. It was an effective $200 price cut in just one year.
But there was a bigger problem that could not be so easily fixed, a fundamental restriction that would hinder the success of the system throughout its life. The PlayStation 3 was, quite simply, too damned hard to code for. It was built around that custom Cell processor, a chip that promised (and, in some cases, delivered) supercomputer-like processing power through a combination of multiple, specialized cores. It's a unique architecture that never really caught on. Shaun Himmerick, a Midway producer who worked on titles for both the Xbox 360 and PS3, called development for Sony's console "a huge pain in the ass." Valve's Gabe Newell famously said the PlayStation 3 was a "waste of everyone's time."
Hyperbole, of course, and the PlayStation 3 would prove its worth by providing the gaming world with some truly iconic experiences, like Journey and The Last of Us. Still, Mark Cerny, lead system architect for the new PlayStation 4, pulls few punches in describing the challenges faced by the outgoing system. While he sugar-coated the response to the $599 price a bit, tactfully calling it "not ideal," he is otherwise very candid about the difficulties of developing for the PS3.
Cerny and I met in a demo room full of gaming journalists and hangers-on, all eagerly getting their first proper taste of the PlayStation 4, the product that he has spent nearly six years bringing to market. He's something of a legend among serious game aficionados, getting his start working for Atari in 1982 when he was just 17 years old. A year later he would design Marble Madness, one of the most iconic games of the early '80s. He's more recently had a long tenure working with, and ultimately for, Sony, having a hand in powerhouse titles like Jack and Daxter, Ratchet and Clank, and Crash Bandicoot.
Cerny's dark, reddish-brown hair and expressive face make him look far younger than his age of 49. A soft-spoken voice and reserved demeanor seem somewhat at odds with his position as the guiding force of the single most important product in Sony's vast portfolio. His honest and straightforward analysis of the challenges faced by the PlayStation 3 is a refreshing departure from the bombastic declarations of superiority that accompanied 2006's PS3 launch. It makes his well-grounded optimism about the future success of the PlayStation 4 feel all the more infectious.
"PlayStation 3 is very powerful. It's a supercomputer on a chip, that's the Cell processor. But, at the same time, to unlock that power you have to spend a lot of time. People have now spent eight years learning the depths of that architecture, how to make games look beautiful and make these rich interactive worlds. You can see that in The Last of Us. It did take a lot of effort. I wanted to be sure that the next time around that it was just a lot easier for developers to make the games. It was very important to me that they could be focused on their creative vision and not learning the minutiae of the hardware."
Ease of development was a primary goal in the initial design discussions for what would become the PS4. Work began in early 2008. That's a stark contrast to the goal of the PS3, seemingly maximum performance at any cost. In the pursuit of power, even developer support, crucial tools that allow for the creation of great games, was left by the wayside. Says Cerny: "With PS3 we didn't really start creating developer tools until the hardware was done, and it turns out that was too late, and so it was pretty tough making the launch titles."
Michiel van der Leeuw is technical director at Guerrilla Games, developer of the eagerly anticipated launch title Killzone: Shadow Fall. He remembers the early teething pains of Sony's previous next-gen effort: "Early games didn't use SPUs at all. [The processing units that made the Cell processor so powerful.] A lot of games had frame rate issues. It took a while to get the hang of it. There was a very big learning curve to it."
And how much easier is the new system? "A lot. We have created a launch title in two and a half years. It's got a campaign that's longer than ever. The world is about four or five times bigger...even with all the struggles of being the first one out the door, with the operating system being in development. I think that's a testament to how easy it is."
Of course, van der Leeuw had something of a privileged position as being a developer on what's widely considered to be the PS4's premier title of the moment. But, his elevated status truly began many years ago. Before Cerny's team got too far down the design process for the console, his team gathered opinions from developers, asking them what their ideal PS3 successor would look like.
"I went out and spoke to 30-something development teams during development and got their feedback on what they wanted," Cerny told me. "This was 2008 to 2009. I couldn't admit that this was about the PlayStation 4. I concealed this as a questionnaire about what the future of video game hardware would look like. But they knew what I was talking about."
Van der Leeuw was one of those 30, and yes, he knew exactly what Cerny was talking about. "It started in 2009, I believe, that I got the first e-mail from Mark Cerny, that he wanted to discuss something about the future and was very cryptic back then. It was very early days, roughly sketching what the future of consoles would look like and what are the biggest problems. I think I was the only one disclosed at this moment and it stayed like that for about a year. I would have secret conference calls in the corners of our office in the middle of the night, because of course the rest of the people were in Tokyo or in the United States." Did he know it was PS4? "Yeah."
Back then the PS4 was known as NGH, or Next Generation Hardware, and part of Cerny's job was corralling the opinions of dozens of developers, encouraging them to speak their minds. Then began the daunting task of satisfying them all. "I asked them: how many CPUs do you feel comfortable working with, because some parallelization is necessary in software on consoles. I heard answers of anything from one CPU up to 1,000, but the consensus answer was four or eight, which is why the PlayStation 4 has eight. They asked for unified memory as our number one request, and we have unified memory. They want a hard drive in every console. And, within our constraint of making an affordable console, we tried to fulfill all of those requests."
The team at Guerrilla Games seems to feel content that the result meets their needs. Managing Director Herman Hulst told me, "We pretty much got what we wanted" -- including a headphone jack in the controller.
Of course, people don't always know what they want until you give it to them, and Cerny's team felt compelled to go beyond the basic demands for more power, more memory, and more storage. Cerny is particularly optimistic about the console's asynchronous fine-grained compute capabilities. This is, basically, using the PS4's Radeon GPU for nongraphical tasking, like more atmospheric audio or realistic physics. "This vision for how that almost two teraflops of power could be used to enhance the interactivity of the world was something that we had, to use the hardware to bring the game much closer to reality," says Cerny.
And then there are the social aspects of the console, enabling gamers to quickly and easily capture and share video of their greatest digital achievements. Online video sharing was just getting into its stride when PS4 design work began, and integrating this was something his team wanted from the beginning. Cerny worked to build bridges across the notoriously siloed Sony enterprise to make this functionality come together. "It was definitely part of the core concept. Once we had the hardware worked out we formed a multidisciplinary team to work on the user experience. That included people across all of Sony Computer Entertainment, and in fact people from other parts of Sony were involved."
Cerny believes that this social element, the ease of sharing footage and highlights from games, will increase the discoverability of niche indie games. These are the sorts of titles that offer refreshing experiences but often get lost amid the big-money marketing budgets of the major publishers. Secret Ponchos is one of those games, a sort of Spaghetti Western beat 'em up that drew crowds at the most recent PAX East. It also caught the eyes of two very important gamers: Brian Silva and Nick Suttner from Sony Computer Entertainment of America's Developer Relations team. Yousef Mapara, president and creative director for developer Switchblade Monkeys, had no idea that he was making the pitch of his life on the show floor.
"Usually you pitch to executives on their terms and you go and they're in suits and using BlackBerrys. Sony was the opposite. They were on the floor disguised as regular gamers, you didn't know they were Sony, and they were just checking out which games they liked....They introduced themselves and said, 'We like your game. How do we get it on PS4?' Before that we didn't think that was a possibility. I actually told them, 'I don't think we can do that, because we're so small. There are too many hurdles.' Brian asked, 'Well, what are your hurdles? We'll help you through them.' I think that kind of proactive approach is going to give them a really impressive indie portfolio."
Indeed, while Microsoft seemed to be ignoring the indie scene when it launched the
"The PS3 was very powerful, but you needed to put in a lot of custom work to get a lot of that power. PS4 just out of the box has more standardized power that is just easier to tap into and is less difficult for programmers....There are processes in that they're making it very easy. If you don't know what to do you can call someone up. There's an account manager that will help you, there are forums where you can post stuff. You get a lot of support from Sony."
Guerilla's Van Der Leeuw is equally impressed by the improvements on the developer support side. "The hardware is much simpler. The unified memory is really important. The developer support is much better. The operating system is much less restrictive, it does a lot more stuff for you. There's a lot of stuff you don't have to worry about anymore. But also, the TRC, the Technical Requirements Check, the testing that every game needs to perform to, that was a big set of requirements you have to fulfill, and that's been simplified."
Of course, there is a dark side to all this: the dreaded port. By making a system easy to target, easy for developers to get up to speed quickly, the barrier of entry becomes shorter. When you're talking about a talented indie developer with a great idea, making things easy is of course a very good thing. However, all too often that same benefit is exploited by bigger, revenue-driven publishers trying to shovel the same game out to as many platforms as possible. When that happens, the game is typically designed for the lowest common denominator, such that a restriction of one gaming platform winds up being felt on all of them.
This is already happening, with big-budget shooters like Battlefield and Call of Duty rushing to hit as many platforms as possible, rather than focusing on being amazing on one. Focus, Van Der Leeuw says, is what enabled his team to do more. "We were able to create a much bigger, more involved world because we were not so limited by the hardware. That's one thing where you'll see about next-generation titles. They'll get richer and bigger as soon as they shed the previous generation, because the cross-generation titles need to work within the limits of the old generation. But they'll soon be gone."
Of course, Guerrilla didn't have much choice: it's owned by Sony. Yousef Mapara's Switchblade Monkeys is not, but by electing to focus on the PS4, his company was given quite the golden ticket, including free dev kits and lots of support. "They brought us to E3 and put us on stage. A small studio could never afford to do that kind of stuff. So just helping us with exposure has been amazing."
These are the sorts of measures required to woo a developer to focus on one platform or another, but given the rapidly escalating cost of development for a major title, winning that kind of exclusive focus (without buying the development studio outright) is getting ever-more difficult. With both next-gen systems offering very similar functionality and internals, there's less reason than ever for a publisher to want to focus on one platform over another.
That said, if developers can more easily tap into and make full use of the power of one video game console than the other, games on that system will look better and will run better. Better games usually mean greater sales, and that's ultimately the name of the game. In that regard, Mark Cerny is already confident. "We have the most preorders of any console we've ever released, so I'd say the reaction to this has been quite good."
As Cerny looks over the room of gamers clutching DualShock 4 controllers, many getting their first taste of the PlayStation 4's first titles, I ask him whether he's ready for launch. "It's been six years. Yes, I'm ready."