A method of processing a spectrally-encoded digital audio signal
comprising band data components representing audio contributions in
respective frequency bands comprises the steps of altering a subset
comprising one or more of the band data components; and generating
recovery data to allow the original values of the altered band data
components to be reconstructed.