Search

Monday, November 07, 2005

Valve's Source Engine: What Went Wrong?

This month a year ago Half-Life 2 was made available over steam. Gamers were chomping at the bit dieing to get to play this highly publicized follow up to the ground breaking 1997 Half-Life title. Even though there was a botched release plan of one year due to a hacker leaking an early build, the gaming public pre-ordered in droves. I know I manage to get 8 licenses over Steam just to play over a LAN.

Somewhere between the aborted release and the final Steam release there were bad signs. The leak had shown that the Source engine AI was not anywhere as intelligent as the showing at E3 2003 had illustrated. Then there were the numerous bugs within the game itself. Despite its admirable "low spec" requirements, the Source Engine displayed some serious problems in its sound properties. The game offered sound stutters and glitchy 3D performance. That's not to say that the visuals in this game did not live up to expectations. They did. Everything from the water to the lighting in the Source Engine gets top notches. Even though there are games that match it, few offer so much on lesser hardware.

Valve even added an unexpected last minute surprise. They announced and released a Half-Life 2 Deathmatch component. So players were getting quite a bit for their money and even use the Source SDK to push their experience toward gaming Nirvana.

That Monkey Wrench Sound

As mentioned earlier, there was a profound problem that affected quite a few users. It didn't take much time before people using the Source Engines in game utilities to see how much the sound element was taxing their systems. For whatever reason, it appears that Valve had forsaken using audio hardware in favor of software rendering. Simply put, that means that despite the Source Engines ingenious use of 3D accelerators, the audio factor more than consumes enough CPU cycles to render multiplayer game play annoying.

It started with Half-Life Deathmatch and it's carried over with the release of Day of Defeat: Source. It's clear at this stage that Valve cannot engineer a solution to the problems. With three distinct titles under one engine displaying the same performance, one doesn't have to speculate why no other developer (except Troika) has optioned to license the Source Engine. Also of note are the different Mods that were scheduled to port to the engine before HL2's release. Many have since abandoned it.

With the release of Day of Defeat: Source, the changes that Valve made to game play to bandage the engine are clear. The problems were already apparent in Counterstrike: Source. There are clearly hit registration accompanied by poor Net code issues. In a firefight, people simply can't hit a target standing five feet in front of them. So with the audio demanding such high resources, Valve opted to include an unrealistic "Cone of Fire" to automatic weapons. This cone is designed to arbitrarily set your shots wide of the target. One can only imagine this was designed in to offset the complaints of poor hit registration already confounding players in CS:S. Needless to say, DOD:S is suffering the same backlash CS:S has fallen to. Fewer people playing the new release in favor of the older less buggy version. Although there is talk of much needed fixes to the DOD:S, those game play fixes simply can't address the core engine problems and make them disappear.

Pretty is Not Enough

There are a myriad of reasons why games succeed or fail, but ultimately game play is at the core of any successful series. Although the Source Engine can offer some unprecedented visuals on low end hardware, its overall technology hobbles the game play of some potentially excellent games. In the end graphics and now the inclusion of HDRI are not enough to save this troubled engine. One can only guess if Valve is already hammering out Half-Life 3 with an completely retooled Source Engine. Somehow I imagine there will be more gamers with a "Wait and See" as opposed to lining up to pre-order.