The development of speech encipherment

This paper surveys some past and present principles of speech coding. The advent of digital logic systems and integrated circuit components is seen as a major turning point in the implementation of effective coding principles that could not previously be used satisfactorily. Various coding methods are discussed, covering frequency domain, time domain and analogue and digital output systems. The problems of key generator integrity, synchronization, received voice quality and recognizability, and channel bandwidth restrictions, and the security offered by different systems are considered. The author concludes that in the immediate future high security speech encipherment systems will be of the digital output crypto-vocoder type, with audio bandwidth analogue output coders offering an increasingly secure alternative for the majority of applications. Digital output coders will continue to be used effectively in those applications where the communication equipment and transmission path have the necessary characteristics to carry low-error-rate digital signals. Advances in analogue-to-digital speech conversion providing acceptable quality at bit rates below 9.6 k bit/s should eventually enable a simple digital output coder to be as effective as a cryptovocoder for audio band high security applications.