Method of and arrangement for deriving characteristic values from a sound signal

Claim 1

An arrangement for deriving characteristic values from a sound signal having a transducer circuit, which circuit receives the electrical signal derived from the sound signal and on whose output short-time-spectrum signal values are available, which values represent the energy of the electrical signal in consecutive equal time intervals for each of a plurality of adjacent frequency ranges, comprising: a memory (9) having an address input; a clock supplying a clock signal; multiple parallel switch means (13) to connect said address input of said memory to the output (8) of said transducer circuit (3, 5, 7) under the control of said clock signal; said multiple parallel switching means being a series of switches connected in parallel, one of which connects said memory address input to said transducer circuit output when a sound signal is present on said output to receive said short-time-spectrum signal values and the number of the associated frequency ranges as addresses; an address circuit (15, 17) which consecutively addresses the memory location (19, 21) of said memory under the control of said clock signal via said switching means; a data output (10) for said memory; a data output of said memory (10) connected to said address circuit; said address circuit forming some of the contents of said addressed memory locations of each frequency range; an adder (11) connected to said data output of said memory such that said adder increments the content of each addressed memory location by one unit and stores the result at said location, thus storing the number of times that the corresponding signal has appeared on said transducer output; a result memory (31);