ter Keurs M, Festen J M, Plomp R
Department of Oto-rhino-laryngology, Free University Hospital, Amsterdam, The Netherlands.
J Acoust Soc Am. 1993 Sep;94(3 Pt 1):1307-14. doi: 10.1121/1.408158.
This paper examines the relations among the spectral contrast needed for speech intelligibility, hearing loss for speech in noise, and auditory filter bandwidth. Fifteen hearing-impaired listeners with relatively flat, mild-to-moderate sensorineural losses and eight normal-hearing listeners participated in the study. The spectral contrast needed for speech intelligibility was determined by reducing spectral contrast in the speech signal and measuring the reduction in contrast beyond which the speech-reception threshold (SRT) for sentences in noise increases. Reduction of spectral contrast was accomplished by smearing the envelope of the squared short-time fast Fourier transform by a convolution with a Gaussian-shaped filter, and overlapping additions to reconstruct a continuous signal. Auditory filter bandwidth was determined by estimating auditory filter shapes at center frequencies of 0.8, 1.6, and 3.2 kHz, using a notched-noise masking paradigm. The results show that limited resolution of spectral contrast is only loosely associated with hearing loss for speech in noise. Moreover, the correlations between the SRT for unsmeared speech and the auditory filter bandwidth at various frequencies were weak.
本文研究了言语清晰度所需的频谱对比度、噪声环境下言语听力损失与听觉滤波器带宽之间的关系。15名患有相对平坦的轻至中度感音神经性听力损失的听力受损听众和8名听力正常的听众参与了这项研究。通过降低语音信号中的频谱对比度，并测量对比度降低到何种程度时噪声中句子的言语接受阈值（SRT）会增加，来确定言语清晰度所需的频谱对比度。频谱对比度的降低是通过用高斯形滤波器进行卷积来涂抹平方短时快速傅里叶变换的包络，并通过重叠相加来重建连续信号来实现的。使用陷波噪声掩蔽范式，通过估计0.8、1.6和3.2 kHz中心频率处的听觉滤波器形状来确定听觉滤波器带宽。结果表明，频谱对比度的有限分辨率与噪声环境下的言语听力损失只是松散相关。此外，未涂抹语音的SRT与不同频率下的听觉滤波器带宽之间的相关性较弱。
