Methods of correlation of any number of input functions or signals to be compared are discussed. It is shown that correlation values can be obtained from any one of several arithmetic processes which measure the degree of similarity of the inputs and that these processes can be mechanized. Included are a simple addition, a sign-coincidence scheme, and two processes based on analysis of variance. The discussion of analysis of variance procedures as applied to signals shows that the answers can be interpreted in terms of statistical significance, that measurements of coherent signal power can be made, and that the accuracy of such measurements is greatly improved as the number of input signals is increased. Signal-to-noise ratio is shown to be uniquely related to statistical significance, and signal detectability is plotted as a function of number of inputs and signal duration. Several means of presentation of data are suggested and illustrated.

This content is PDF only. Please click on the PDF icon to access.

First Page Preview

First page PDF preview
You do not currently have access to this article.