Posted: 14-05-05 In: Editorials By: Munro2 |

‘The source engine gives us capabilities in four main areas, believable and realistic human beings, graphics that were previously impossible outside a Hollywood movie studio, an integrated material and physics system that create an unprecedented level of interactivity, and artificial intelligence that welds all of these things together into an experience gamers have never had before’

Gabe Newell, Valve Software E3 2003

Run through the forum listings of any of the many half life 2 fan sites, and invariably amongst the myriad of subjects, you’ll find members questioning the usefulness and validity of what Valve are bringing to the gaming world with the Source engine.

Like many others, when I first watched the E3 Trailer, my jaw dropped not only at the unrealised level of action I saw before me, but also at the world of potential game play possibilities that the source engine was offering up; both to developers and mod makers.

Now arguments rage back and forth regarding the pros and cons of particular game engines and their relative strengths and weaknesses. However, this article isn’t intended to be a glorification of one, or a defamation of another. I think that all of the forthcoming game engines have their own strengths and weaknesses. However, there are certain aspects of the Source’s engine capabilities that intrigue me.

The intention of this article is to objectively assess those capabilities and hint at the possibilities and applications they might have for future gaming, as well as the potential problems and pitfalls that they might also bring.

However, firstly it is important to separate the engine from the game:-

‘A little history’

In the great history book of gaming, the release of Half-Life stands out as one of those golden moments that marked a revolution of ideas and perception; both public and professional. As a product it laid down definitive precedents with respect to storyline, interaction and AI. As a game it set down clear market expectations.

Judging by the E3 footage, it is clear to see that Valve have considerably strengthened the core elements of game play that made Half-Life so much fun, whilst retaining the essence of what constitutes the original Half-Life experience. In order to begin to separate the engine from the game, we need to establish what constitutes the Half-Life experience. What made it so successful, and how can we summarise it?

‘Defining the Half-Life experience’

The key aspect that made Half-Life so compelling was immersion. Right from the beginning the game threw you into the world of Black Mesa. There was no FMV cut scene, or similar plot exposition. Instead you found yourself in Gordon’s shoes straight away, and the game delivered 99.9% of the story to you through this framework. There are very few static cut scenes throughout the game (and those that there are, are merely the briefest of flashes), almost all of the action and exposition of the plot takes place wholly within the confines of what you see, hear and witness as Gordon Freeman. Blink and you’ll miss it, just like real life. No game had really attempted to go as far as Valve did with Half-Life, to try and tell a story in this manner before – this was the key to the game’s success.

‘The limitations of the Half-Life experience’

The particular approach and manner of storytelling that made up Half-Life worked brilliantly for its subject and style. However, it is important to recognise that like all systems, it possesses limitations. For example, there is a clear distinction between the Half-Life and Deus Ex storytelling experiences: however, both games operate loosely within the FPS envelope and share genres to a certain degree.

In Half-Life the story is revealed to you through episodic encounters throughout the game, but the manner of delivery is always one where by the AI dictates events to you without deviation. As Gordon, you never ask questions, the AI simply responds to your presence.

In Deus Ex, the whole game revolves around you asking questions and investigation techniques, the nature of the game alters to a greater or lesser degree upon your choices, which impact whether the AI aids or attacks you accordingly. The dynamics of the game are largely player driven.

However, the nature of Deus Ex requires an entirely different game play methodology compared to Half-Life, because of its story telling approach. Deus Ex uses cut scenes and external views to explore almost all of its player AI interactions, subsequently its engagement leans more towards the plot rather than environment and action, which lead to a separation of immersion.

What I hope this illustrates is that the game engine should not be confused with the interface that it uses; the interface is merely the mechanism of interaction for the game play. The thing to be cautious of is that the more complicated you make the level of interaction in a game, the more you risk reducing the level of environmental immersion.

If we take away from Source the expectations of how we imagine Half-Life² to play, then we can begin to assess its features more critically. Because of the high level of integration that source brings, it is very hard to separate particular elements and discuss them without referencing others. So instead of talking directly about these features, I think it best to view them in relation to the impact they will have upon the two principal elements that make up gaming situations; the characters and environments. It’s a big subject to discuss, so I’m going to concentrate on characters for this part, and with the next section discuss environments.

‘Believable and realistic human beings’

 

‘I think the part of it that's going to really shock people is how they feel about the other people in the game: the emotional reactions they have to the characters in the game. And at the end of it they are going to look back and say: I didn't think I could have that experience with a computer character.’

Gabe Newell, Gamespot interview 2003

‘A lot of it was just listening to what our fans told us. A lot of it was just them saying, "Wow. We really liked Barney." And we didn't expect that at first. Barney and the scientist were just supposed to be these "throw away" characters and weren't really supposed to be anything too interesting but then, all of a sudden, people were really responding to them, they really loved them and we thought that was great. And then we said, "Okay, this is something that we have to pull ahead on. We have to do more with this.’

Josh Weier, Valve Software. www.sharkeyextreme.com interview 2003

With respect to characterisation in mind, the biggest element that Source brings to the game is the advanced AI facial and Lip-sync technologies it contains. The level of expressive realism displayed by Alyx and the G-Man sets a new precedent for game players’ expectations; it also opens up a gateway of opportunities for mod makers and developers to explore more diverse themes within the world of gaming.

This technology is a core feature, and is one that is set to become as fundamental an aspect of future story driven games development as physics has become in recent years. Certainly under close scrutiny no-one is going argue that either Alyx or the G-Man are realistic in appearance, however as simulacrum there is no denying that the illusion of humanity they exude is uncanny. As rendering technology improves, greater levels of realism will only enhance this feature.

A lot of the decriers of the facial/lip-sync tech cite its superfluous nature in the frenetic world of multi-player FPS gaming. However, just as oranges are not the only fruit, deathmatch is not the only game style. The immediate gains of this technology are going to be seen in the realms of adventure gaming, adding greater levels of emotional depth, complexity and immersion to the genre; from action orientated games such Deus Ex or Knights of the Old Republic, through to more puzzle orientated games like The Longest Journey or Grim Fandango.

Much like the silent screen stars of old, present gaming icons have tended towards the visual aspect of their representation. Perhaps such technology will bring us game icons that we acknowledge more for their actions and demeanour than simply for their appearance.

The matter of how far and how widely such technology can be used is always going to be tied to the power of our imagination versus the hardware limitations we are subjected to. In many ways it is easier to consider the pitfalls and problems that new technology introduces in order to perceive its potential.

The facial/lip-sync technology gives developers and the mod community the opportunity to bring to life genuinely well rounded and believable game characters for the first time. Naturally, this means that much greater levels of development need to be brought to bear in the creation of such characters, needless to say increased development in one area has to have an impact elsewhere, either in time, staffing, cost and ambitions. It is therefore going to be very important to gauge the status of a character’s role within a storyline and round them out accordingly. Your major characters will be very major, but your minor characters will be very, very minor.

It is clear to see that the facial/lip-sync tech works well in the real time first person environment of Half-Life², where the AI relays information to the player without any real player interaction. However, if you are developing a more interactive game, then issues naturally arise regarding a dialog interface. The simple solution is to fall back to a cut scene dialogue position as used in games like Deus Ex: Invisible War or Knights of the Old Republic, which is a known formula, but can impact upon immersion and in some ways negates the advantage of the technology as a result.

The more challenging approach is to devise an ‘in-frame’ response system, which maintains immersion and offers up dialogue options. However, such a system naturally relies upon swift responses by the player in order not to break the atmosphere; it also relies upon the AI not being distracted by other activities around it, as it will be occurring in real time. A compromise would be to introduce time freezes when dialog options appear; however, this, much like the cut scene dialogue option is bound to affect player immersion.

Similar questions also arise when you give thought to how to handle more complicated inventory management and simple RPG tasks such as reading a book etc. and yet still maintain a sense of immersion within an ‘in-frame’ perspective. Small as these aspects are, they all add to the atmosphere that makes or breaks a game’s success. However, interactions like this begin to fall into the realms of the environmental; a subject that I want to look at in the next part.