Dillier N, Spillmann T
Audiophonologische Abteilung, ORL-Klinik, Universitätsspital Zürich.
HNO. 1988 Aug;36(8):335-41.
A newly developed battery of 12 multiple-choice tests for the German language (graded according to difficulty) comprising identification of sentences, words and syllables was applied to 8 patients with single-channel cochlear implants and to 35 deaf hearing-aid users. Phoneme confusion matrices were analysed using sequential information transmission analysis (SINFA) and multi-dimensional scaling (SINDSCAL). Acoustic parameters of the digitally stored test items were compared with the outcome of SINFA and SINDSCAL. It was found that cochlear implant and hearing-aid users relied on the same perceptual dimensions which were correlated to acoustic signal properties of the speech sounds. Best perceived acoustic features were time and amplitude structure of the speech sound (envelope). Fundamental and first formant frequency are evaluated both by implant and hearing-aid users. The use of higher frequency spectral information depends on the amount of residual hearing.
一套新开发的针对德语的由12项多项选择题测试组成的测试（根据难度分级），包括句子、单词和音节识别，应用于8名单通道人工耳蜗植入患者和35名使用助听器的聋人。使用序列信息传输分析（SINFA）和多维标度法（SINDSCAL）分析音素混淆矩阵。将数字存储的测试项目的声学参数与SINFA和SINDSCAL的结果进行比较。结果发现，人工耳蜗植入患者和使用助听器的人依赖相同的感知维度，这些维度与语音的声学信号特性相关。最佳感知声学特征是语音的时间和幅度结构（包络）。人工耳蜗植入患者和使用助听器的人都评估基频和第一共振峰频率。高频频谱信息的使用取决于残余听力的程度。
