Abstract: Disclosed is a method and apparatus for inserting a data signal into an audio signal to provide a tagged signal. The method includes removing a band of frequencies centred at a predetermined notch frequency from the audio signal, spectrally shaping the data signal so that it takes on the precise shape and magnitude of the envelope of the audio signal at the removed band, and inserting the shaped data signal into the removed band of the audio signal. The method may be used to identify an audio segment or it may be used to encode the audio signal with other desired data. The method of the invention provides for the data signal to be virtually inaudible to the listener of the audio segment yet robust enough to survive severe audio signal processing.