What We Learned About Cloud Gaming From The Stadia Launch

Posted by

Google Stadia shows a ton of promise for cloud gaming. We believe that the future of gaming will be powered by streaming. There are some lessons to learn from the launch about the importance of content and unique experiences that the underlying streaming technology enables, but the opportunity to deliver games directly to consumers through streaming is here and only going to grow into the future.

We’ve been building game streaming technology for nearly four years at Parsec. We believe in a future where consumers are freed from hardware and can access interactive content and software from any device. To enable that future, we’ve built ultra-low latency game streaming software that runs natively on most hardware at 60+ frames per second and up to 4K quality video. Consumers use our product to play games together, connect with a significant other across the ocean, keep the games going when they’re not at home, access their gaming library on any device no matter where they are, and increase productivity with 3D applications when they’re not sitting in front of their workstation. Across 10s of millions of gaming sessions, we’ve optimized our networking protocol and our video capture, encoding, decoding, and rendering software to create the most consistent gaming experience possible on thousands of network configurations and hardware. Fortunately, we’ve had years toiling on this problem in obscurity to learn some of the lessons about cloud gaming that Google Stadia stumbled on yesterday with their launch.

New Distribution Technologies Need Killer Use Cases

Game streaming is a new distribution technology/medium for delivering games. Throughout the history of entertainment, new distribution technologies have created enormous opportunities for the content that they support. But it is the combination of new types of content for that medium in addition to it that generally makes consumers excited. Looking back at the release of the radio, the television, the first gaming consoles, and the internet, a pattern emerges. At first the distribution technology is too expensive and offers nothing new compared to the distribution technology it replaces. Typically, due to a lack of content natively built for the medium, the content on the new medium is ported from the previous technology. Only once the content creators develop something specific to that technology (soap operas on the radio, richer visual sets on TV, replaceable cartridges for consoles, and social media/blogging on the internet) did consumers really adopt the new technology. Game streaming will almost certainly follow the same pattern. Right now, Stadia and xCloud are porting AAA games that run on your consoles to their new distribution technologies. Generally, the biggest value add for consumers is that you can play on any device and never invest in gaming-specific hardware.

Cloud gaming/game streaming will take off when new experiences that were never possible before are available there. When we say new experiences, we’re not talking about exclusive releases to the platform. We’re talking about truly new experiences that are currently impossible. At Parsec, we released our Arcade to make it so people can play offline games online with their friends. That’s a unique use of this technology and something that gamers actually want. Gaming is a form of entertainment, but it’s also a conduit for socializing with friends. We’re building that social experience for every game that hasn’t built online multiplayer directly. Separately, we’re also giving gamers access to their PC from any device. That way, they get the benefit of playing from any device without being dependent on an online cloud gaming service.

The problem that Stadia and other cloud gaming service providers face is that we’re compared to an exact alternative: playing games on a console or gaming PC. Playing on your local hardware will always be better for the exact same experience. For some people and many experiences, they won’t be able to tell the difference between local hardware and streaming (it’s amazing how far this technology has come). But as soon as there’s a blip, a consumer will blame it on the streaming service. There are ancillary benefits of access to more titles, taking your games with you, and no downloads, but those are pains that people are willing to live with if they get to have the best experience once they fire up their game.

In the future, we believe that publishers will create incredible cloud-only experiences with tie-ins to other media (like a movie) or take advantage of the infinite resources available in the cloud to create smarter NPCs or more immersive environments. Those will require years of investment and a game engine that is specifically built for cloud gaming. Once that happens, we believe the sky is the limit for cloud gaming.

Content Is King

Stadia launched with a library that just wasn’t that exciting. That’s because publishers are smart. They’re not going to hand over the keys to their kingdom for a low price. Publishers want to deliver their games directly to their customers. It’s a $24 billion opportunity for publishers to deliver their games directly to consumers through cloud gaming according to Morgan Stanley. The profit is a great motivator, but it’s also valuable beyond that in data collection. When a publisher delivers their game directly to a consumer, they own a lot more data about the behavior of that customer, and that data is very valuable. In video streaming, that’s one of the main drivers behind Disney+. Game publishers have studied what happened in the video streaming world when media companies made a massively bad bet licensing their content to Netflix. Game publishers see how that played out, and they will almost certainly skip the licensing phase of cloud gaming and go direct to consumer. The wildcard is xCloud. With access to a huge catalog via Game Pass, Microsoft is in a great position to offer the subscription cloud product, but even the head of Xbox, Phil Spencer, thinks that this is a few years away.

Despite all of that, in reality, a library of games actually isn’t as valuable to a gaming consumer as it is to the music, TV, or movie consumers. At the end of the day, gamers actually only buy and play a few games each year. There were 350 million people who bought a premium game in 2018, and they spent $17.9B on those games. That means each player is only spending $51 per year or about 1 AAA game per year. Now, those people are probably playing more games, including free-to-play games, but it’s not 100s of games because they’re playing each of those games for 100s of hours each. That’s one of the many reasons why publishers are drastically reducing the number of games they release while investing in games-as-a-service where they continuously update a title they released. For example, in 2008, EA, Activision, and Ubisoft released 98 games; in 2018 they released just 28, not including expansions. Game Pass has shown that engagement increases and people play more games with a subscription, but it’s not a significant number to change the economics of a publisher to the point where they’re willing to bundle their content with 100s of other games.

Even if a consumer wanted a huge library of games, they don’t need cloud gaming for that. Playstation Now, Xbox Game Pass, Uplay+, EA Access, and more provide access to a library of games on your own hardware at home.

Over Promising To A Skeptical Audience Hurts The Image Of Cloud Gaming

Cloud gaming has a lot of potential, but we have to be realistic with the earliest adopters. The experience is going to have issues in certain situations, and we should be clear with gamers that this will happen. The gaming community is rightfully skeptical of new technologies. Telling reporters that Stadia was achieving negative latency, had all sorts of features that it didn’t launch with, and used AI to predict your next move in the game really wasn’t the right approach for reaching gamers. Cloud gaming has failed in the past — promising that this time would be better than your console at home created unrealistic expectations and disappointment. With the current version of cloud gaming where a customer is just playing a AAA game in the cloud, you have to be realistic with them and say that this isn’t going to be as good as that Xbox at your house. They’re willing to forgive you for that if you enable something new that they couldn’t achieve without your technology. For instance, the share play feature where someone could jump into a game where a streamer left off on YouTube was a killer feature if implemented correctly.

The Technology Is Close And It Will Only Get Better

With that being said, Google Stadia has shown the world that game streaming is close to being ready for prime time, especially if you’re on a great network and have the right hardware. We’re close to being ready for everyone to stream games because our internet infrastructure is getting better. A significant hurdle for game streaming to be viable has always been an infrastructure problem. In the United States, access to high speed bandwidth is increasing across the board. Since 2013, the number of households with access to a 250mbps connection has increased from 0% to 60% in 2017. This has almost certainly only increased in the interim period. With 5G services being released to bring higher bandwidth connections to your home via cellular networks, more competition among ISPs and telecom companies will likely generate a significant increase in access to these higher speed connections at a lower price (competition is good for consumers).

FCC Broadband Survey

For reference, a 60 FPS 1080p video using the h.264 codec requires about 30mbps of consistent bandwidth to avoid video quality degrading. Using the newer h.265 codec for the same video requires about 15mbps. That means most of the United States has a good enough internet connection (as long as it’s consistent) to support game streaming today.

On top of the bandwidth requirements, there’s a hardware requirement too. In order to maintain a low latency video stream, both the cloud GPU and the consumer’s device need a specialized piece of hardware to encode and or decode the game stream using hardware acceleration on h.264 and h.265. h.264 hardware accelerated decoding became standard on most devices in 2012 and 2013 while h.265 hardware accelerated decoding became standard on Nvidia GPUs and iPhones in 2015 and Intel integrated GPUs in 2016. With hardware refresh cycles happening about once every 5 years for computers and once every 3 years for mobile devices, almost every device in circulation in the United States and Europe will have a specialized chip to decode h.265 in 2020 or 2021. This is very important to the success of game streaming because hardware accelerated decoding and encoding reduce latency and power consumption for the video stream significantly.

This Is Really Difficult Technology To Get Right

We’ve been working on low latency game streaming technology for four years. Granted we’re not Google, but this is our mission. We can tell you from experience that this is not an easy technology to build. Every gamer has different preferences and every piece of hardware and networking setup is unique. It’s impressive that Google rolled out Stadia to the masses all at once, and it performed as well as it has in the first couple of days. There are many tricks and software optimizations to consider when building game streaming technology.

For example, we believe that building your own networking protocol is essential — that’s why we invested nearly a year in that endeavor and continue to invest in it with every release. Additionally, every platform and piece of hardware requires low level optimizations that squeeze the most out of their performance. That’s why we built native apps from the ground up optimizing on each piece of hardware. And because each network and hardware scenario is unique, you also have to make trade offs on compromising video quality, frame rates, latency, and frame time optimizations. When people sense “lag” in their gaming, it can ruin the experience of escaping into the world of the game. Lag comes in many different forms, however. Reaction time is just one noticeable form of latency. The other source of lag is actually less about the reaction time and more about the timing of frames. We believe that people can see slight differences in the time between frames arriving more than they can sense a few extra milliseconds of latency between a button press and the visual appearing on the screen. That’s why we’ve always been extremely focused on delivering frames as close to a 16.67ms interval at 60 FPS as possible.

Parsec is far more likely to deliver a frame at 16.67ms than the competitor we tested against

In Conclusion, Game Streaming Is Here

We believe that game streaming will change how people play games. Our mission has been to free consumers from hardware and increase access to the world’s best content and software. We’re doing that through our game streaming technology and powering unique experiences for people to play games together and access their gaming hardware from anywhere. We’ve also opened up our technology to other developers, enabling anyone to build streaming into their application and increase access to experiences they create. Cloud gaming, like Stadia, is the future. We think it’ll take a different form when it truly takes off in the market, but it’s inevitable that games and content will be streamed from hardware in the cloud. If you’re working on games or software that can benefit from an ultra low latency, high frame rate streaming solution, we’d love to hear from you.

Leave a Reply