Digital Foundry's 2010: Part 2

Invert the polarities!

Feature by Richard Leadbetter Technology Editor, Digital Foundry

Published on 1 Jan 2011

Welcome to part two of Digital Foundry's 2010, a retrospective article where technology editor Richard Leadbetter discusses the major stories of the year, adds some personal thoughts on the games and hardware unveiled and reveals some behind-the-scenes information on the making of Digital Foundry's articles.

Part one of this feature, posted last week, covered off January to June, and this concluding chapter picks up the story right from where we left off. In addition to musing over the news stories, talking about the Face-Offs and discussing the articles, we've also included our most popular articles of month, decided upon by the volume of readers they attracted. Yes, a "performance analysis" of our own work, for a change.

Indeed, while we have the opportunity, let's abuse the stats database still further, with a top 10 Digital Foundry articles rundown encompassing the entire year. It's an intriguing look into how the overall editorial mix works out, and it's nice to see that some of our most ambitious articles got the readership they deserved.

The internet being what it is, the results are in a state of flux though. DF features tend to remain well-read from one month to the next, so if we re-ran this analysis a couple of months down the line, the positions of these stories would probably shift about a little, with the more recent pieces such as the GT5 Tech Analysis moving within striking distance of the number one slot.

Now, let's return to the story of Digital Foundry's 2010, kicking off with how the genesis of the company brought about one of the most successful pieces we posted this year...

July

At the beginning of 2004, I set-up Digital Foundry Ltd, which began life mostly creating video showcases and pack-in DVDs for the games industry. The launch of the Xbox 360 in 2005 radically altered the business – with the lack of gaming capture hardware available, the company went out and made some. After that, rather than make videos for other people, they bought the kit from us to do it for themselves.

The roots of the business in creating concept videos lives on, and occasionally I like to post a showcase that demonstrates this. A foundry, after all, makes things and I like DF to contribute something new and unique every once in a while somewhat removed from the usual menu of analysis, news and platform comparisons. Red Dead Redemption: World in Motion proved to be the video hit of the year for the site, and the corresponding blog entry pulled in an enormous range of new visitors.

Red Dead Redemption: World in Motion. Yes, we really did capture every shot in real-time! One game 'day' is around 45 minutes in total. It's a very long process as we typically capture around 50 shots in all and then use just the best.

The hardware side of the Digital Foundry lives on, but we're running out of gaming HD standards to support with our TrueHD kit - thanks to the wonder of digital handshaking, it instantly adapts to the signal being fed to it. We had pretty much complete support for PlayStation 3D running in our tools in less than a day, allowing us to begin the first of our PlayStation 3D performance analyses, a process we subsequently refined and improved for our later Gran Turismo 5 tech magnum opus.

The launch of OnLive the previous month also allowed us to get our first hands-on test of the revolutionary, "disruptive" service while we were in the USA. We posted a detailed latency measurement ahead of a full DF analysis the following day. The result boiled down to a system that worked, but offered variable response (150ms upwards) and image quality compared to conventional systems. We included frame-rate tests of several games in the piece, which proved to be something of a challenge.

FPS analysis is all about counting dupe frames, but OnLive picture quality is so degraded compared to conventional video that our tools could only isolate around 60 per cent of the identical frames. This meant that I spent a day meticulously going through each video, checking every frame and adjusting the readouts when a dupe was discovered but not recognised as such by the software. I think the results were worth it: the OnLive piece was fair - acknowledging the genuine technical achievements (a 150ms response is faster than some local games) - and it was comprehensive. We were able to accurately assess the system based on established methodology, and could finally put the many PR claims to the test.

I've still yet to try Gaikai, but I did get a look at it briefly during E3 - I saw it running on PC, but intriguingly saw an Xbox 360 prototype too. In both cases, image quality looked significantly improved over OnLive, but fast action games still had issues - a sneak peek at Split/Second: Velocity looked disappointing compared to the flawless World of Warcraft. And in that situation, the server was actually in the room with us as I chatted with the Gaikai staff, making latency evaluation a bit of a non-starter.

Slowed-down video gives us the bandwidth necessary to produce a good representation of the quality level OnLive offers. Somewhat variable, it has to be said - sometimes pristine, often terrible. Picture quality overall can change radically just from one frame to the next.

I'm still not sure what to make of these streaming video gameplay systems. It all seems to be about fooling the player into thinking he's playing a local experience, but any side-by-side testing would leave you under no illusions of the quality level on offer.

Video compression by its very nature is best suited to movies and scenes where not much visual information changes from one frame to the next. Only some videogames conform to that ideal. Most do not. There is no magic video codec that could handle the visuals of something like Super Stardust HD, and even a standard driving game is going to be significantly degraded over the local experience. I think the future of OnLive can only really be as a sub-par alternative. Priced correctly, or for those raised in the YouTube generation where video artifacting is taken as the standard, maybe that won't be a problem. But everyone looks for value and right now OnLive doesn't offer it.

The free market is conspiring against OnLive too. Consoles and indeed games are now so cheap that a paid-for service like OnLive simply doesn't make any kind of sense... The sampler mechanic of Gaikai is undeniably intriguing, however.

Other Digital Foundry articles: Guerrilla Games made bold claims about controller response in Killzone 3 - something we backed up with hard input lag measurements a few months later. Sony's Rick Marks paid us a visit to demo the startling potential of PlayStation Move, while the Criterion tech team sat down for a chat about the brilliant work they were carrying out for Need for Speed: Hot Pursuit.

Face-Offs: The last multi-game round-up for several months was posted, covering Singularity, Transformers: War on Cybertron, LEGO Harry Potter: Years 1-4, Tiger Woods PGA Tour 2011 and Afterburner Climax.

Digital Foundry's Most Read:

August

We've heard about sub-HD games and how lower-resolution framebuffers are often upscaled to native 720p, but LucasArts' Dmitry Andreev came up with something far more exciting: a frame-rate upscaler. Demonstrating a 60FPS demo of Star Wars: The Force Unleashed II, Andreev had successfully manage to upscale temporal, rather than video, resolution. Not only that, but he posted compelling video evidence showing his techniques in action.

Similar to the insane amount of interest in our anti-aliasing article in January, it was pleasantly surprising to see this article, and indeed the follow-up Tech Interview, garner so many readers. It's fair to say that a lot of the Digital Foundry output, particularly the interviews, can be hard work for the average gamer, but it's clear that the appetite for the information is there, particularly when the innovations we talk about are demonstrated so brilliantly.

With 3D gaming having once again been showcased extensively by Sony during E3, Digital Foundry took a look at Xbox 360's claims of being 3D-ready. Without the major backing of the platform holder it remains unlikely that we will ever see a coordinated attempt to make the Microsoft console have anywhere near the same kind of support as the PS3, even though games like Call of Duty: Black Ops actually run smoother and look better on 360 compared to the PS3 game.

In our article on Xbox 360's stereo 3D credentials, we covered the work of TriOviz in providing a 3D image without the performance issues seen in many of the other 3D console titles. This vid demonstrates that while TriOviz might not offer full stereoscopy, frame-rates and resolution remain much the same as the 2D game.

Story of the month was the news that PlayStation 3 had been cracked by the mysterious PSJailbreak team, with our full analysis being one of the most popular articles of the month. I surmised that it would be a short-lived cash-in, and while the promised updates from the PSJailbreak team turned out sporadic and not particularly impressive, I did not expect an open-source alternative to spring up so quickly, which effectively kept the piracy bandwagon rolling all the way up to the release of Need for Speed: Hot Pursuit and Gran Turismo 5 three months later.

Other Digital Foundry articles:Digital Foundry revealed more inside information on Kinect, its capabilities and the challenges facing developers in our Case for Kinect piece, and mulled over rumour-mongering about a PlayStation Phone. We also broke a genuine developer exclusive, with the news that DICE collaborated with Criterion on Need for Speed: Hot Pursuit.

Face-Offs: Just the one in August, a triple-format dissection of the disappointing, not to mention migraine-inducing Kane & Lynch 2: Dog Days.

Digital Foundry's Most Read: