Coding of an audio signal (x) is provided where the coded bitstream (AS) comprises a parametric representation of the audio signal. One component of the parametric representation comprises tracks of linked sinusoidal components (CS) where subsequently linked components are coded differentially from parameters of linked signal components already determined. The coder scrambles (18) the differentially encoded frequency and/or amplitude values by mapping the differential values onto other differential values (CSe). By modifying these values, the bit-stream is still decode-able but results in tracks with random frequency and/or amplitude variation. Consequently, the signal will be degraded.