Signal Processing for the Detection of Acoustic Transients
James F. Bartram · The Journal of the Acoustical Society of America · 1970
There is a variety of possible ways of classifying sounds. One approach, which is explored in this paper, is to use the first-order probability distribution. Specifically, a mathematical model is constructed, consisting of a family of first-order probability distributions with a parameter called the “distribution index,” with which to select a particular member of the family. It is then assumed that the set of real acoustic sounds can be described by this model and characterized by the distribution index. A signal processor is derived, based on the distribution-index concept. Finally, this signal processor is tested with real input sounds and is found to respond characteristically to ambient noise, sinusoidal signals, and sharp transients, as is desired.