Time Compression Multiplexing for Loop Transmission of Speech Signals

This paper considers the use of time compression multiplexing (TCM) for telephone loop communications. The analysis treats TCM transmission degradations in general, and applies the results to telephone speech transmission. To gain an understanding of the nature of TCM degradations, transmission through a linear, time invariant network and demultiplexing is considered. The channel signals are then assumed to be wide-sense stationary random processes, and expressions for the spectral properties of the TCM distortion and interchannel crosstalk are determined. To accurately evaluate the performance of a TCM system, a practical and economical equalizer is proposed. It is designed to meet speech transmission requirements for the telephone loop plant. A worst case application is chosen to evaluate the system performance. The evaluation of the TCM transmission performance utilizes both the random signal model and a computer simulation using digitized speech. The random signal analysis shows that the spectra of the distortion and crosstalk are significant for human speech. Results of the digitized speech simulation demonstrate that the subjective effect of the signal distortion is not severe, but the crosstalk is somewhat intelligible and more severe than expected.