Method for forming a digital signal representative of an original audio signal and a method for searching for such digital signals within a sound database in a manner analogous to searching for text using a web browser

摘要

Method has the following steps: digitization (102) of an original audio signal for a given duration, sectioning of the signal into time slices, such that an originating matrix (101) is formed that contains respectively samples in the time slices, determination of the frequency distribution of the signal amplitude in each of the time slices to form a repartition matrix and then to further process the signals in terms of frequency and amplitude until a matrix melody is obtained representative of the original selected audio signal. An Independent claim is made for a method for recognizing a part of an audio signal forming at least a part of a melody from a sound database.