Spatial release from masking for Chinese-native listeners in English speech recognition

Spatial release from masking for Chinese-native listeners in English environment is investigated in psychoacoustic experiments for two kinds of interferences from different directions and with different signal to noise ratio(SNR) signals.Both targets and interferences were produced by loudspeakers in an anechoic room and the correct rates of targets were obtained by listeners' recognition.The results of experiments showed that correct rate was above 99%when only speech targets were in front of listeners,the correct rate was about 57%when speech targets and speech interferences were both in front of listeners,and the correct rate was about 96%when speech targets and speech interferences were produced randomly from a direction of±60°.When speech targets and interferences were both in front of listeners,it is found that the correct rate dropped from 96%to 34%with decreasing SNR from 0 dB to—12 dB under the condition that noise interferences were in front of listeners.While speech interferences were in front of listeners,the correct rate first dropped,then rose by 27%and after that,resumed its declining.When the noise and speech interferences were produced from a direction of 60°,the correct rate dropped respectively from 99%to 80%and from 98%to 91%with SNR decreasing from—4 dB dB to—16 dB.The conclusions are that the spatial separation has an obvious gain effect on English speech intelligibility of Chinese-native listeners,and the correct rate dropped with SNR decreasing under most situations.This agrees well with the conclusion of the related researches on other native languages.