Table of Contents
*This post may contain affiliate links. As an Amazon Associate we earn from qualifying purchases.
Developer: id Software
In this highly modern day and age, it can be difficult to come up with fresh ideas for computer games. After all, with the current technologies today, there is probably nothing you can think of that hasn’t been implemented yet. In fact a lot of recent ideas that are deemed innovating and unique have actually been done in previous projects that simply did not get due attention and publicity.
Technology development, meanwhile, produces mere rehashes of past hits that have been modernized and produced on a bigger scale. For instance, Morrowind is more of an iteration of Daggerfall compared to the sequel. Likewise, Return to Wolfenstein is an old idea adapted to new settings. These great games of the past have a lure that makes people keep coming back to them.
Thus, it is not surprising to see Doom 3 emerge. Except for the number 3 in the title, nothing is unexpected. The game was developed and released by the original maker, id Software, and not by a satellite studio such as Raven, which is currently busy on Quake 4. With a release date that is more than 10 years after its predecessor, the 3 in Dome 3 seems like an attempt to bind the two games somehow.
With new technologies, methods, and standards, Doom 3 easily detaches itself from the original version of Doom. In fact, the game Painkiller can be thought as a more likely heir to the old Doom than the supposed new version.
In the previous Doom, the 3 dimensionality of the environment didn’t seem artificial to ordinary folks. It was obviously fake, but id Software’s true 3D engine was displayed first in Quake. It was high quality trickery. Still fake, but it would be difficult to tell it apart from reality. One thing unique about the first Doom engine was the ability to create wide open spaces and fill them with crowds of monsters without negatively affecting performance, and despite the games preference for close, dark passageways.
The problem with the Doom 3 engine is that it can barely build significant open spaces and it is more inclined to render only 3 monsters at a time, max. Thus, in this aspect, the Doom 3 engine is closer to the fabled Quake, instead of its older namesake. However, the game was named Doom 3 nevertheless, to justify the Mars setting.
Because most of Doom 3’s basic designs are decided by its strengths and limitations, it is a good idea to discuss first the new game from Carmac. Like all the company’s past releases, Doom 3 also raises the bar in terms of graphics. However, the main research area this time wasn’t in geometry, textures or even ultra-fashionable shaders. Carmac has, for the first time, has instead tried to come up with a lighting system that is unified.
In a unified lighting system, to the uninitiated, all sources of light are processed in a similar way, which is in real time at all times, and well-calculated. This is unlike past id Software engines. No preliminary light map generation is done, which is strange for all Quake engines. You can set a source of light in the Doom 3 editor and view the results immediately, including surface illumination, and the shadows that various objects cast.
The new lighting system has the primary advantage of being able to render shadows for each frame, and in real time. You can see monster casting dynamic shadows, moving sources of light to shade the environment correctly, and view movements of passing by lamps. This is actually the next step in computer graphic development.
The system definitely has its own limitations that are not only linked to the present video chip capacities, but also to the need for support compatibility with past generations of hardware. Carmac opted for shadow volumes as an algorithm for shadow rendering. The good thing is that the video chip’s 8-bit stencil is enough for the given operation. A stencil buffer is, in fact, supported by every video card that runs on 32-bit colors, beginning with old-timers like Riva TNT. After all, shadow volumes are easier to use by comparison. Objects are made to cast their own shadows and put peculiarities at a minimum.
Glitches in the Matrix
Shadow volumes have their own set of cons. First, rendering the shadows is dependent on the fillrate. The more surfaces, the more rendering time is required. This is an area where GeForce video cards excel because of their built-in ability to make the fillrate double when writing to stencil and Z buffers. This makes it possible for the GeForce FX 59×0 video card fare well in Doom 3, although it doesn’t do as well in many tests. The top fillrate for stencil and Z buffers is around 13 gigapixels/second. This explains the total domination over the Radeon competition in Doom 3.
Likewise, it is difficult to render the soft, muzzy shadows that happen in reality with the use of shadow volumes. In fact, each light source must be processed several times and later averaged. This is obviously not an effective algorithm. Even Carmac acknowledges that shadow mapping, an alternative shadow rendering algorithm, has the potential to oust shadow volumes in future id Software engines since the issue of muzzy shadows can easily be solved through shadow mapping. The algorithm, however, comes with its own limitations that need to be addressed by newer video chips.
To sum up the previous 2 sections, it can be concluded that shadow volumes are not well-suited for vast open space shadow rendering. The shadows behind windows are most of the time blurred with light reflected by a lot of objects. In addition, there is Earth atmosphere’s effect when scattering and refracting light rays to contend with. Open space shadows can virtually be endless, and the bigger the shadow, the more pixels is required by the video chip for rendering; the more the fillrate rises to have the shadow rendered.
It is therefore obvious that unless video cards improve soon, the best shadow volume niche will be filled with narrow games characterized by minimal open spaces, and sharp shadows cast by bright spotlights.