METHOD OF EXTRACTING SOUND SOURCE INFORMATION

摘要

A method of extracting sound source information capable of detecting as a quantitatively clearly interpreted quantity a property of a fixed point from a filter center frequency to an output's instantaneous frequency from instantaneous data. A method of extracting sound source information using the fixed point of mapping of a frequency onto an instantaneous frequency; wherein a value, obtained by partial-differentiating in a time direction a partial differentiation in a frequency direction of an instantaneous frequency obtained by instantaneous frequency frequency differentiating circuits (3, 10) for each filter (2, 9) and a partial differentiation in a frequency direction of each filter output obtained by instantaneous frequency time frequency differentiating circuits (4, 11), is properly weighted and is subjected to a short-time weighted integration in a time direction, whereby an estimated value of a carrier wave-to-noise ratio for each filter is calculated by carrier wave-to-noise ratio calculating circuits (5, 12) to determine a carrier wave-to-noise ratio and obtain an estimated value of an evaluation quantity.