Digital audio scrambling system with error conditioning

Abstract

In the scrambling system, an analog audio signal is converted into a digital signal to provide a sequence of digital signal samples corresponding to the analog audio signal. Each digital signal sample is compressed to provide compressed signal samples having a sign bit, three exponent bits and seven mantissa bits. Each bit of each compressed signal sample is exclusive-OR'd with a unique keystream to thereby scramble the audio signal. A Hamming code generator generates code bits for correcting singular errors in a combination of the sign bit, the exponent bits and the code bits; and a parity bit generator generates a parity bit for detecting double errors in a combination of the sign bit, the exponent bits and the code bits and for further detecting an error in the most significant mantissa bit and/or the parity bit. The bits from a plurality of successive compressed, error-encoded signal samples are interleaved and serialized in order to separate the bits from any single sample by at least a predetermined duration associated with an FM discriminator click. The serialized, interleaved, error-encoded, compressed signal samples are combined to provide two-bit digital words. The digital words are converted to digital PAM data signals which when converted to an analog signal by digital-to-analog conversion, provide a pulse-amplitude-modulated signal having a level related to the binary value of the digital words. The digital PAM data signals are converted to an analog signal to provide the pulse-amplitude-modulated signal. The descrambler system descrambles the scrambled audio signal by a process that is the converse of the scrambling process. Singular errors in a scrambled signal sample are detected and corrected by a Hamming error corrector. Double errors in a scrambled signal sample are detected by a parity bit check and compensated for by repeating the last received error free signal sample.