STAR TREK ONLINE
GRAPHICAL DEVELOPMENT SECRETS:

Achieving the look of your game.

• Gain confidence that the team can achieve the look. Star Trek has a distinct visual style, which is essential for games, movies, art, any media based upon the collective series. This style is not necessarily easy to replicate, for all of it being as iconic and well-known as it is.
• Opportunity to experiment and innovate. VLD allows the artists and lighting designers to test familiar techniques and invent new methods by which to achieve the distinctive look of Star Trek .
• Understand the technology. Many of the methods of generating lights, sheen and shadows are quite complex; by experimenting, the team can learn the limits of the engine, the styles of generating lights, and where new capabilities may be needed.

• Concept Art. Absolutely essential for any form of electronic visual media, the STO team utilized a number of stills taken from the various Star Trek series in their efforts to match the visual style for STO . Special attention was paid to the techniques by which the distinctive visual effects of the starship interiors, lighting and surfaces.
• Research Techniques. Whether by researching the Nvidia and ATI graphic sites, by playing other games, by attending conferences, or ideally, by doing all three, it's vitally important to be familiar with the various methods and techniques by which cutting-edge graphics are achieved. PE paid special attention to the Doom 3 and Source engines for STO, particularly the use of dynamic lighting and shadows.
• Test Techniques. Another essential part of the process, the team must be certain that the chosen methods actually create the desired results, by using the creation tools, such as Maya, whenever possible. This stage allows a design team to select which techniques will actually be used, and which will be discarded. PE, for example, rejected the use of direct lighting when the results were rather less than the effects PE was looking for, settling instead on baked radiosity. Pieragostini also showed screenshots of some of the test images as examples.

• Objects must be rigid - not a particularly large limitation when lighting starships, but worth mentioning.
• Lights must be at infinity - also not a particularly problematic limitation, as lights in space are infinite anyway.
• Per-vertex lighting only - a much larger problem, as the STO team found vertex counts skyrocketing on account of having to add large numbers of unnecessary vertices in order to have the starships lit correctly, thus increasing poly counts to levels the engine would have difficulty handling regularly.
• Lighting environments were approximated - the results of using Pre-computed Radience Transfer were not intuitive, in terms of effect, and would create strange lighting artifacts. As a result, the lighting artists were unable to achieve the results they desired.
• Cast shadows were too diffuse - even when an object rested upon the surface it cast its shadow upon, the shadow was too diffuse to be realistic; there seemed to be no effective way to remedy this problem.