The first steps taken in the Digital Light Research involved selecting the proper digital light.
LED's were selected to be the lighting source. Factors such as rapid rise and fall time, long lifetime, and the increasing availability made this decision easy.
Next LED's from several different manufacturer's and vendors were evaluated for the strength and purity of their so-called white LEDs. The following images show the spectrums of each LED evaluated. Several common light sources are provided for reference.
Joker 2000 Light
We selected the Nichia SPW300bs based upon its color, price, size, and availbility. After selecting the Nichia LED, we proceeded to build a prototype setup using a breadboard and several thousand of the LEDS. We used this setup to verify our hypothesis that LEDs did in fact correct the color variation problems.
Prototype LED Breadboard Setup
Once this was verified, we proceeded to find a vendor that could provide boards. We found a distributor called West Virginia Signal and Light, who distribute boards made by Precision Solar Controls. They were willing to modify their traffic light boards to include white LED's instead of the standard Red, Green and Yellow.
Precision Solar Controls LED Traffic Lights
We ordered 20 of these white traffic lights. The in-house engineering staff helped to design the control circuits, and to mount the LEDS in the ceiling.
Once the ceiling was constructed, several test runs were conducted to compare the results of the Tele-Immersion system using the normal room lighting, vs using the LED Digital Light to illuminate the room. The frames were then diff'ed(1-2, 2-3, 3-4, etc). The norm was calculated for each of these diff'ed images. The mean and standard deviation of these norm values were then calculated. This table documents those results for a subset of the frames captured. The results clearly show that the LED digital light provides much higher quality and consistent images from frame to frame.