Speech scrambling employing adaptive rate sampling

A new speech scrambling technique is described. The speech is spectrally scrambled at the transmitter to produce a defined multiple bandpass structure. The reconstruction ´key´ comprises a knowledge of the spectral frequency bands of the scrambled elements together with a unique sampling rate which will allow correct recombination of these elements. To prevent unauthorised reconstruction, the bandpass structure and sampling rate can be changed irregularly according to an appropriate key sequence.