Development of a hybrid design method for protecting speech privacy in open-plan spaces

本研究针对目前尚未建立的混响声和掩蔽噪声效果评估方法，研究了一种用于保护“语音隐私”的混合设计方法，这种方法通常用于表示房间内机密语音的泄漏问题。具体而言，两种评价方法的充分性；(1)通过提高背景噪声水平来代替混响声的效果；(2)通过对不同类型的听力测试来评估掩盖不愉快和可理解噪声的效率。此外，根据听力测试的结果，对设计过程中考虑的物理值预测单词可理解性分数的等可理解性轮廓进行了修订。

In the present study, a hybrid design method for protecting "speech privacy", that is often used for representing the problem of the leakage of confidential speech in rooms, is investigated focussing on evaluation methods of effects of reverberation sound and masking noise which has not been established. Specifically, adequacy of the two evaluating method; (1) replacing the effect of reverberation sound with an increase in background noise level, (2) evaluating the efficiency of masking noise from both unpleasantness and intelligibility, were confirmed on the basis of several kinds of listening tests. Furthermore, equal-intelligibility-contours that enable us to predict word intelligibility scores from physical values considered in designing process were revised according to the results of the listening tests.