THE USE OF TAPE RECORDINGS BY LAW ENFORCEMENT AGENCIES HAS EXPANDED TO CONSIDERABLE PROPORTIONS. THE AVAILABLE TECHNIQUES--AS WELL AS CONDITIONS IN THE FIELD--HAVE LED TO SUBSTANTIAL PROBLEMS RELATIVE TO THE DECODING OF SPEECH ONCE IT IS RECORDED ON TAPE. THE SPEECH OFTEN CANNOT BE UNDERSTOOD DUE TO THE PRESENCE OF NOISE, COMPETING SIGNALS, LOW RECORDING LEVELS, AND OTHER FACTORS. THERE ARE TWO GENERAL APPROACHES TO THE ENHANCEMENT OF SPEECH ON TAPE RECORDINGS. ONE APPROACH--DIGITAL SPEECH PROCESSING--CAN BE UTILIZED WHERE THE SPEECH IS SEVERELY DEGRADED. HOWEVER, TECHNIQUES OF THIS TYPE DEMAND SUBSTANTIAL ENGINEERING SOPHISTICATION AND COMPUTING POWER FOR PROPER APPLICATION. A SIMPLER METHOD--ONE THAT CAN BE CARRIED OUT WITH ONLY MODEST EQUIPMENT AND THE REASONABLE TRAINING OF PERSONNEL--IS BASED ON THE DIRECT PROCESSING OF THE SPEECH SIGNAL ON THE TAPE. THIS APPROACH UTILIZES VARIABLE SPEED TAPE RECORDERS, FREQUENCY BIASING EQUIPMENT, AND FILTERS TO IMPROVE SPEECH INTELLIGIBILITY. SUCH APPROACHES VARY IN THEIR EFFECTIVENESS, OFTEN WORKING WELL, SOMETIMES NOT SO WELL, DEPENDING ON THE CONDITIONS OF THE TAPE RECORDING. HOWEVER, A FACILITY CAPABLE OF PERMITTING GOOD TAPE PROCESSING OF THIS TYPE CAN BE ESTABLISHED AT EVEN RELATIVELY SMALL CRIME LABORATORIES. GRAPHIC DATA AND REFERENCES ARE PROVIDED. (AUTHOR ABSTRACT MODIFIED--KBL)