Sound masking performance of time-reversed masker processed from the target speech

Noise is typically used as a masker to reduce speech intelligibility of ambient speech in offices. However, the efficiency of noise in masking speech is not satisfactory. In this paper, the masking produced by time reversal of the target speech was investigated. The masking efficiency, the subjective annoyance and the cognitive load produced by time-reversed maskers were measured to evaluate their performance. The results of the intelligibility test show that a time-reversed masker with the same target-to-masker ratio (TMR) as a noise masker is more efficient in reducing speech intelligibility. However, the results of a short-term memory test show that the cognitive load is not decreased when time-reversed maskers are used to mask the target speech, while the perceived annoyance of time-reversed maskers is no better, and sometimes worse, than that of the noise masker.