Low Resolution TV: Subjective Effects of Noise Added to a Signal

For the design of a television communication channel it is desirable to have a figure of merit for comparison of channels. As our sophistication in the design of communication channels increases, so we must also increase our sophistication in defining and measuring a suitable figure of merit. As a measure of merit we may use the power spectrum N(u) of the error, or noise, added in the chamiel which can be measured for all frequencies, a>, in a given transmission system. However, the ultimate receiver is a person viewing the picture, and his sensitivity to noise superimposed on the picture depends upon the distribution with frequency of that noise. This dependence of the viewer's sensitivity to noise can be considered equivalent to a linear filter and a linear detector. We will call this sensitivity function a subjective noise-weighting function, W(«), defined on the video bandwidth, 0 to 12.1,2,3'4 This subjective noise-weighting function gives the value at each frequency of the relative contribution of noise to an overall figure of merit. We define this figure of merit as 1 /P, where 233