On the use of 'processing-gain' as a performance-measure for spread-spectrum systems

Theoretical analyses are presented to show that the use of 'processing gain' (the ratio of spread bandwidth to information bandwidth) as the figure of merit in evaluating the performance of such basic spread-spectrum techniques as direct-sequence modulation, frequency-hopping and time-hopping may sometimes yield misleading results. Analytical expressions for the performance of binary noncoherent FSK systems which use frequency-hopping spectrum-spreading are employed to demonstrate that, for cases involving noise jamming, tone jamming or swept-FM jamming signals, actual performance gain may vary as much as 6 dB from a calculated 'processing gain' of 27 dB.