End-User DLP Projector Colour Calibration

We have analyzed different approaches to calibrating 4-segment Digital Light Processing (DLP) projectors from the perspective of an end-user. A modification is introduced to Wyble and Rosen’s 1,2 model that improves its prediction of tristimulus values (XYZ) as a function of input RGB. Tests also show that Tamura, Tsumura, and Miyake’s Masking model, which was originally introduced to compensate for channel interaction in LCD monitors, performs as well as the improved Wyble-Rosen model in the forward direction (RGB to XYZ). For predicting RGB values given XYZ input data (backward direction); however, we find that the Masking model is more accurate. All the models considered in this paper involve only basic look-up tables and matrix multiplication and as a result are computationally efficient.