SPEECH SCRAMBLING METHODS CAN BE CATEGORIZED AS DIGITAL OR ANALOG. IN DIGITAL SCRAMBLERS, THE VOICE SIGNAL IS FIRST CONVERTED TO DIGITAL FORM AND THE RESULTING BIT STREAM IS THEN ENCRYPTED. ANALOG SCRAMBLERS MODIFY THE VOICE SIGNAL BY OPERATING ON THE SIGNAL IN AMPLITUDE, TIME, OR FREQUENCY DOMAINS. THE SCRAMBLED SIGNAL CAN BE TRANSMITTED IN THE SAME BANDWIDTH AS THE ORIGINAL CLEAR SIGNAL. SCRAMBLING METHODS IN THE ANALOG CATEGORY INCLUDE SIMPLE FREQUENCY INVERSION, FREQUENCY-HOPPING INVERSION, BANDSPLITTING, AND TSP. THE MAJOR ADVANTAGE OF TSP, WHERE THE VOICE SIGNAL IS DIVIDED INTO SHORT SEGMENTS THAT ARE TRANSMITTED IN SCRAMBLED ORDER, IS RELATED TO THE UNCHANGED NATURE OF THE SIGNAL'S SPECTRAL CHARACTERISTICS IN THE SCRAMBLING PROCESS. THE SLIDING-WINDOW VERSION OF TSP IS SUPERIOR TO THE HOPPING-WINDOW VERSION IN THAT IT SPREADS SIGNAL ELEMENTS OVER A GREATER DISTANCE, GIVEN SEGMENT LENGTH AND TIME DELAY. SECURITY ASPECTS OF THE SLIDING-WINDOW TSP SCRAMBLER AND THE INTRODUCTION OF SPEECH SCRAMBLERS INTO EXISTING POLICE TELECOMMUNICATION NETWORKS ARE DISCUSSED. SUPPORTING DATA AND ILLUSTRATIONS ARE PROVIDED. REFERENCES ARE CITED. (DEP)