The University of British Columbia

The authors develop a practical method of evaluating the power spectral density of guided scrambling line-coded sequences. Analysis of block-coded signals is reviewed, and it is shown that it is impractical to analyse continuous guided scrambling coded sequences with expressions presented elsewhere in the literature. Alternative expressions for sequence autocorrelation are derived which depend on the relationship of several code parameters. Two relationships are considered in detail, and it is shown how expressions derived in these instances can be applied to other guided scrambling codes. Using these expressions, the spectral characteristics of several continuous guided scrambling codes are evaluated. These spectra demonstrate the important result that variation in the average statistics of the coded sequences with varying source-stream statistics decreases as the degree of the scrambling polynomial increases