Skip to main content

Digital Foundry vs. OnLive

At stake: the fundamentals of gaming as we know it.


For a new gameplay platform to be truly considered a generational leap over what is currently available, the requirement is simple: it just has to be better.

OnLive comes with some cute features that fit that requirement. Its use of the media video stream to spy on other gamers in the Arena is an example of the flexibility of the platform, and likewise with the Brag Clips, the use of video in such an imaginative fashion is something that PlayStation Network and Xbox Live can't compete with.

Other elements are also praise-worthy. The ability to play any game in the library almost instantly, and test the vast majority of them via playable demos that are actually 30-minute samplers of the whole game, is a superb idea. Similarly, the concept of hassle-free rentals via digital delivery is something that the gamers have been crying out for, and OnLive delivers. In this respect, the convenience factor is undeniable.

However, despite the incredible achievement in streaming gameplay with relatively low latency, the bottom line is that the gameplay experience is not better than what we already have - by and large it's tangibly worse. The varying quality of the graphics is questionable, and the lag is best described as "better than expected" - nowhere near the claims that have been made for the system, and still measurably inferior to current standards. It's just a question of how your personal perception level will interpret it as to whether it's a game-breaker or not.

In terms of buying games, the prices for new titles are too high and the selection of games is uninspiring. The notion of paying so much for what is measurably an inferior product compared to the physical disc means that OnLive simply cannot be taken seriously at this point in time - especially when you don't own the games you are buying. This is something the industry en masse needs to get its head around as digital delivery becomes more important: if you're going to take away basic ownership rights, and offer a lower-quality version of the game to boot, the price needs to diminish accordingly.

On the flipside of this, rental costs can be fairly reasonable: $7 to $9 for five days' play gives you plenty of time to wade through the average single-player campaign. A lot of people who buy their games, complete them and sell them on will be interested in this alternative option. However, the puzzling lack of consistency strikes again here too. Almost all games have playable demos (with the sole exception of Prince of Persia: The Forgotten Sands for some bizarre reason), but only some have rental options. Weirdly, DiRT 2 is only available as a demo. While there's a good argument that the publishers should retain control over how their games are deployed on the system, the bottom line is that when OnLive starts to charge for its subscription, the consumer should expect to have all titles available as rentals, especially bearing in mind the unappealing price points for the full games.

Of course, in terms of overall value for money, OnLive says you don't require any specific gaming hardware and you'll never need to upgrade, so this in itself sounds like you're making a massive saving, but the fact is that the PC required to run the system isn't insignificant. Unless you're running on a laptop, upgrading it with a relatively inexpensive graphics card will offer an overall gameplay experience that effortlessly beats OnLive on quality and response. Perhaps by the time expensive next-generation hardware is unleashed upon us OnLive's value proposition will increase accordingly, but until then the value just isn't there.

OnLive is like a box of chocolates. You never know what you're going to get. Video quality can be almost pristine in some cases (left), or really messy in others (right).

Almost 18 months ago now we came up with several good reasons why OnLive couldn't possibly work, at least in relation to the specs and claims being made by the company itself. Now we've been hands-on with the final product, the company needs to be congratulated on just how close it has got to sorting out the latency issues which were one of the key concerns. Out of controlled conditions, OnLive has managed to get within spitting distance of console response times and that's a clear technological achievement worthy of recognition.

However, even in this regard, lag doesn't meet OnLive's on-the-record promises, and elsewhere the system comes up short. The claims of 720p60 don't stack up compared to the reality of the service (unless you are describing the technical make-up of the transmitted video stream rather than actual game performance) and the quality of the image in challenging situations is poor and no match whatsoever for playing the same game locally. OnLive generally seems to be a system that can work well for certain games, but really isn't very well suited for others.

Our other key concern way back when was how the system would cope with extreme load from a mass of players all connecting to play the latest hot game. Right now this simply can't be tested as genuine, recent blockbusters are absent from the service and in the multiplayer titles there's a sense that OnLive isn't terribly popular at the moment - UTIII is the closest thing OnLive has to a Call of Duty or Halo. That extreme load we were concerned about simply hasn't materialised yet.

In the here and now, the question marks on performance and value combined with the lack of games mean that we can't really recommend the system, but you would hope that the content side of things would improve at least. From there it's just a case of whether the individual can get over the video and lag issues. Certainly, for UK gamers, the BT deal should provide a straight connection to the datacentres without having to traverse the internet - this gives the system a fighting chance of improving its latencies still further.

As for whether overall quality of Cloud gaming will improve, clearly it's early days right now. But it's something that forward-looking game companies like CryTek are investigating already - tying in the game renderer directly with the encoding process, based on the concept of "points of interest" maps which allocate bandwidth to areas of the game scene where it is needed the most. The potential there is quite startling, and we'll be looking at the possibilities in more depth in a later feature.

Read this next