人耳辨识非语言声目标能力的实验研究

Investigation on human ear's capability for identifing non-speech objects

以主观听觉测试实验为手段,研究人耳听觉系统辨识能力与声音特性、评价主体训练程度及噪声干扰的关系.实验结果发现:人耳辨识声音能力主要由声信号的谱时结构决定,人耳对于谐音信号的辨识能力明显强于非谐音,对于稳态信号的辨识能力略强于瞬态信号;评价主体的训练可以提高辨识效果及准确性,但与声音特性密切相关;不同的噪声干扰对人耳目标辨识能力的影响主要取决于声音频谱结构;比对测听方式能提高非受训人员的辨识准确性,但不影响受训人员的辨识效果.

Investigation on human ear＇s capability receives much attention in auditory science and target automatic recognition. In this paper, the relationship between auditory identification, nature of sound, training level and ambient noise are investigated by the test of subjective auditory capability approach. The test results are as follows： Human ear＇s capability of identification is dependent on sound spectral structure and it will be better when the signal is of harmonic nature and in steady-state. The identification accuracy, being closely related with sound characteristic, can be improved by training of juries. The auditory identification capability under noise environment lies on sound spectral structure.