Multiple maskers for speech masking in open-plan offices

Abstract

Sound masking is an effective method to reduce speech intelligibility for reaching acceptable speech privacy in open-plan offices. Speech-like masker is considered as a good choice to reduce speech intelligibility. Using one masker in masking system is the common way so far. In this paper, a new method was proposed, using randomly integrated speech frames (SF) of the speaker as masker. Because SF masker was based on speech of the same speaker, it is a good speech-like masker. We further studied the method of using multiple SF maskers. The performance of this new masking method was evaluated by objective measurement (STOI). The speech masked by one SF masker and multiple SF maskers were compared. Results showed that speech with SF masker could reduce speech intelligibility. Speech with multiple SF maskers led to poorer speech intelligibility than speech with only one SF masker. Masking speech by multiple SF maskers is an effective way to reduce speech intelligibility in open-plan offices. (Acknowledgement: This research was supported by the joint training PhD program of Chinese Academy of Sciences-Fraunhofer, and Fraunhofer Institute of Building Physics in Stuttgart, Germany.)