Sponsored By

Postcard from GDC Europe Mobile 2005: Case Study: Designing Games For 3G Networks

As part of the GDCE Mobile lecture series on the first day of Game Developers Conference Europe 2005, Matt Bellows, former founder of Wireless Gaming Review and now responsible for operations, sales and marketing at Boston-based mobile developer Floodgate Entertainment, presented a fascinating lecture on the higher-bandwidth 3G phone networks, and their potential for creating complex, near-console quality games.

Simon Carless, Blogger

August 30, 2005

10 Min Read
Game Developer logo in a gray background | Game Developer

Introduction

As part of the GDCE Mobile lecture series on the first day of Game Developers Conference Europe for 2005, Matt Bellows, former founder of the CNet-acquired Wireless Gaming Review and now responsible for operations, sales and marketing at Boston-based mobile developer Floodgate Entertainment, presented a fascinating lecture on the higher-bandwidth 3G phone networks, and their potential for creating complex, near-console quality games.

The most interesting part of Bellows' talk was his belief that, rather than insisting on the power of casual gaming for mobile titles, favoring simple titles such as card games or 2D puzzle titles, he and his company are increasingly aimed at the cutting edge of 3D mobile gaming.

While he was not necessarily rejecting the possibility that casual games are important to mobile phones, Bellows' decision to show Blizzard's MMO World Of Warcraft in his opening slide, commenting on the "intense, deep, rich play experience that people want to build", and further arguing: "The thing that's interesting to me about the inevitable progression of this industry is the entertainment, absorption and creative opportunities enabled by mobile phones", clearly placed his flag in the sand as an advocate of the advanced 3D graphics and in-depth gameplay shown by the most advanced mobile games.

3G For Go?

In his opening remarks, Bellows commented that the very best phone games have PlayStation-like graphical capabilities. But he believes that it will be the ability of swifter 3G networks, alongside 3D chipsets in new handsets, which will gradually enter the industry over the next few years, so that allowing lower ping times for playing multiplayer games, and equally importantly, faster download times for extra game content, will immeasurably help the sophistication of mobile gaming.

But first, some basic specifications. According to Bellows, 2G networks generally support up to 40kbps "dial-up"-style data rates, whereas 3G advertises 300Kbps to 2Mbps for its connection, and borderline '2.5G' would be at around 100Kbps. Basically, the point was made - lower latencies, of around 250 to 500 millisecond ping times, facilitate something much close to real-time multiplayer gaming, and can stretch out mobile multiplayer gaming beyond turn-based gaming.

As for technology drives in the mobile market, the improvement of mobile processors is clear - the newest phones have 150Mhz ARM9 class processors, fast DSPs, faster RAM, and higher bus speeds. Handsets have a short life cycle, anecdotally as short as 12 to 18 months, and vendors look for compelling features to differentiate new phone models.

In fact, Bellows points out, in some ways, it's not video games that have been the big mobile success story over the past couple of years - it's camera phones. These are one of the biggest drivers of mobile phone purchasing, but more sophisticated camera phones lead to more RAM, larger memory cards, hi-res LCDs, and eventually, GPUs in phones, as camera and video features require more processing power. This all helps the rise of the mobile game, as more sophisticated handsets mean better-looking games.

Facts, Darned Facts

Bellows then went on to state some facts about 3G phones. According to 3GToday.com, there are 187 million CDMA 3G subscribers today worldwide, from 111 operators, and 30.8 million WCDMA/EDGE 3G subs from 46 operators. Of these, by far the greatest amount are still in the relatively advanced Eastern markets, with a conglomeration of Korean operators alone totaling 34 million, KDDI's AU showing 19 million, and DoCoMo's FOMA having 13 million subscribers. In addition, perhaps optimistic survey are indicating that there may be 118 million 3G subscribers in China alone by 2008.

But it seems that, according to surveys, 3G phone subscribers buying more content than they would from 2.5G handsets. Most interestingly, Bellows points out, you can now build real client-server mobile games with the expanded bandwidth and speed on 3G networks, and you can put plenty of game content on the server, including new levels, textures, audio and so on. Bellows' company Floodgate Entertainment did a similar thing on 2G networks with its adaptation of BioWare's NeverWinter Nights, which had around 1.5MB of content that downloaded to the game while playing to allow a bigger player experience. On 3G, it's contended, you could possibly download tens of megabytes.

But, hang on, perhaps you're saying - what about the data costs of grabbing all this information over the network? Bellow referenced seeing some amazing online-enabled titles in Japan that were just not played that much because of costs for bandwidth. But the growing popularity of flat rate billing plans makes extended access for online-connected play sessions or extra online-downloaded content increasingly possible.

A prime example of this is IdeaWorks 3D and EA Mobile's Need For Speed Underground 2, which has recently rolled out in North America as part of Verizon's V-Cast service. This is a 3 or 3 and a half megabyte download which streams a gigantic amount of data, including car and scenery data, from the network while playing, and even has MP3s downloading from the network and playing in the middle of the game. Clearly, this would not be possible or sensible if the user were paying per kilobyte of data.

Swashbuckling For Dummies

So, as has been noted, latency for multiplayer mobile titles should drop with 3G, but it just won't be Quake deathmatch in near future - and indeed, 3G networks still are far from dominant. With that, Bellows introduced a multiplayer game that his company is working on, Swashbuckler, which works by far the best on 3G networks, but also works reasonably on 2 and 2.5G networks, and is not simply turn-based, as the vast majority of the titles are. Swashbuckler uses a thin client approach, and each client is designed to handle somewhere in the order of 9k/sec for download, and 3k/second upload for responsive real-time play, but can throttle back somewhat on 2G networks.

Swashbuckler practices instance-based multiplayer, with each instance of the game handling 32-64 players - but a single physical PC server running to receive the client instructions can handle hundreds of instances. There are also periodic updates of new content to be downloaded to phone.

But what about the game? Since the title is an online pirate ship combat title, even with latencies of 250ms to 1000ms, the cannon attacks that the game uses have inherent game design latency - the projectiles are not necessarily fast, and there's time for the ships to reload. This built-in latency in the game does not fix some problems with ships 'warping' and popping to other locations on slower 2G networks, but with 3G, at around half of that latency, the game works excellently.

More GPU, Less Latency

Finally, Bellows went on to talk about the rise of the 3D GPU-enabled phone, which although only incremental, is likely to make a major difference to the visual fidelity of mobile games, and ties in with 3G's lower latency to produce an increasingly console-like feel. Again, Bellows made the point that casual gamers are less likely to be enthralled by 3D, but nonetheless, 3D phones with GPUs may, by his estimation, reach a penetration of 20 million handsets by mid 2006 and 35 million by the end 2006,

Of those 3D GPU-enabled phones, the major players are Nvidia and J2ME, used in phones from Motorola, Sony Ericsson, and Samsung, and ATI and Brew, used for handsets including LG and Samsung. Popular 3D chips for mobile includes the ATI Imageon 2300, Nvidia AR10 core, and the Intel 2700. As Bellows explains, 3D chipsets provide 30,000 to 150,000 polygons per sec, which works out to 1,500 to 7,500 polygon scenes at 20fps, not unreasonable. The chips also allow features such as MPEG4 playback, Z-buffer usage, texture mapping, anti-aliasing, and even, on the high end, pixel shaders.

But, Bellows concludes, outside of Korea and Japan, wide adoption of GPU-enabled 3D content in phones is still at least a year off. In addition, 3D game development requires specialized programming and game design skills which closer approximate console development. Nonetheless, there are already options for 3D mobile gaming standards, and, according to Bellows, OpenGL ES 1.1 seems to have an early lead, with JSR184 and Direct3D options as well.

So, although 3D titles are increasingly being commissioned by publishers who want to differentiate themselves from the mass of 2D games in the market, Bellows points out that development costs for 3D mobile phone games can increase by two to five times for technology, art, and design. Even just switching to a higher resolution can be expensive, as FloodGate's own experience with NeverWinter Nights for smartphones was that it took almost double the art months to produce art for an enhanced-resolution Brew version of the game.

Probably the outstanding question from viewing Bellows' talk was - as phones get ever more complex and 3D and higher-bandwidth becomes ever-easier, is the ever more-expensive cost of developing complex titles going to mean diminishing financial returns? Is there, perhaps, going to be a split into low-cost 'casual' titles and high-cost 'hardcore' titles for mobile games? The question is still very open, but Matthew Bellows seems to clearly believe that cell phones should tend to the World Of WarCraft experience, and he may well be able to thrive flying his gryphon and doing it.

______________________________________________________

Read more about:

Features

About the Author

Simon Carless

Blogger

Simon Carless is the founder of the GameDiscoverCo agency and creator of the popular GameDiscoverCo game discoverability newsletter. He consults with a number of PC/console publishers and developers, and was previously most known for his role helping to shape the Independent Games Festival and Game Developers Conference for many years.

He is also an investor and advisor to UK indie game publisher No More Robots (Descenders, Hypnospace Outlaw), a previous publisher and editor-in-chief at both Gamasutra and Game Developer magazine, and sits on the board of the Video Game History Foundation.

Daily news, dev blogs, and stories from Game Developer straight to your inbox

You May Also Like