Moore B C, Glasberg B R
Department of Experimental Psychology, University of Cambridge.
Br J Audiol. 1988 May;22(2):93-104. doi: 10.3109/03005368809077803.
Hearing impairment of cochlear origin is usually associated with loudness recruitment. As a consequence, the dynamic range between threshold and the highest comfortable level is smaller than normal. To ensure that low-level sounds can be heard, while avoiding discomfort at high levels, a hearing aid with automatic gain control (AGC) is required. This paper compares four different systems for implementing AGC, and compares each of them with unaided listening and with linear amplification. The systems were evaluated by measuring thresholds for understanding speech in quiet and in five types of background sound: speech-shaped noise, 12-talker babble, cafeteria noise, traffic noise and a single competing speaker. The first system used a new dual-action AGC (called dual front-end AGC) operating on the whole speech signal. A slow-acting control voltage (recovery time 5 s) held the average level of speech at the output constant, regardless of the input level. In response to sudden intense transients, a fast-acting control voltage (recovery time 150 ms) reduced the gain rapidly and then returned the gain to the value set by the slow-acting component. In the second system, referred to as the mark II aid, the output of the dual front-end AGC was split into two frequency bands, and fast-acting (syllabic) compression was applied in the high-frequency band only. The bands were then recombined. The third system resembled the mark II aid except that fast-acting compression was applied in both bands. The fourth system resembled the 2-channel aid evaluated in previous trials (Moore, 1987). It was similar to the third system, but had only single-action front-end AGC with a recovery time of 400 ms. Six subjects with moderate sensorineural hearing loss accompanied by recruitment were used. Best results overall were obtained using the mark II aid. Speech reception thresholds (SRTs) in noise were, on average, 4 dB better than for linear amplification and 2.4 dB lower than for the previous 2-channel aid. There was a significant advantage of having fast-acting AGC in the high-frequency band, but no advantage of having AGC in the low-frequency band.
耳蜗源性听力障碍通常与响度重振有关。因此,阈值与最高舒适水平之间的动态范围比正常情况小。为确保能听到低强度声音,同时避免高强度声音时的不适,需要一款带有自动增益控制(AGC)的助听器。本文比较了四种不同的AGC实现系统,并将它们分别与无助听和线性放大进行比较。通过测量安静环境以及五种背景声音(言语噪声、12人交谈声、自助餐厅噪声、交通噪声和单个竞争说话者)下理解言语的阈值来评估这些系统。第一个系统使用了一种新的双作用AGC(称为双前端AGC),作用于整个语音信号。一个慢作用控制电压(恢复时间5秒)使输出端语音的平均水平保持恒定,而不管输入水平如何。响应突然的强烈瞬态,一个快作用控制电压(恢复时间150毫秒)迅速降低增益,然后将增益恢复到由慢作用组件设置的值。在第二个系统中,称为Mark II助听器,双前端AGC的输出被分成两个频段,仅在高频段应用快作用(音节)压缩。然后将这些频段重新组合。第三个系统与Mark II助听器类似,只是在两个频段都应用了快作用压缩。第四个系统类似于之前试验中评估的双通道助听器(Moore,1987年)。它与第三个系统相似,但只有单作用前端AGC,恢复时间为400毫秒。使用了六名伴有重振的中度感音神经性听力损失患者。总体而言,使用Mark II助听器获得了最佳结果。噪声中的言语接受阈值(SRT)平均比线性放大时好4分贝,比之前的双通道助听器低2.4分贝。在高频段有快作用AGC有显著优势,但在低频段有AGC没有优势。