Objective measurement of display formats: multidimensional and multimodel user perception models

Comparing the effectiveness of display formats, especially displays set in different sensory modalities and containing complex combinations of dimensions, can be likened to the proverbial task of comparing apples and oranges. Dissimilar displays can be compared if a “unit-less” dimension can be found which describes how well critical information is expressed, compared to other information contained in the display. The signal-to-noise ratio (SNR) is just such a measure. Fourier power spectra can be computed for energy imparted by the display of critical information (signals) and the remainder of the display (noise). By computing the SNRs for each feature channel (modality or dimension), complex SNRs can be computed to describe the salience of the signal. Also considered is the similarity of signal and noise as expressed in the Pearson product moment correlation coefficient. Computational examples of such display SNRs are presented and discussed