普通话言语测听材料的智能化研究

Computer assisted application of mandarin speech test materials

目的应用计算机软件编制出性能稳定可靠、操作简便、智能的普通话言语测听系统，并对其进行临床验证。方法利用Cool Edit Prov 2．1专业音频软件，对具有等价性的普通话言语测听材料的CD音频进行数字化无损分割，采用Delphi编程语言在计算机中实现言语测听智能系统。最后经过听力计放大输出，在隔声室（双室）中进行测试。测试55例耳科正常人的7张单音节词表识别率，应用SPSS统计软件对词表等价性进行分析，验证智能系统的可行性。结果计算机智能系统测试方法中，测试表间的差异不具有统计学意义（F=1．582，P〉0．05），由此得知7张单音节词表之间具有良好的等价性。结论该智能测听系统初步具有临床应用的可行性。

Objective To design an intelligent speech test system with reliability and convenience using the computer software and to evaluate this system. Methods First, the intelligent system was designed by the Delphi program language. Second, the seven monosyllabic word lists recorded on CD were separated by Cool Edit Pro v2. 1 software and put into the system as test materials. Finally, the intelligent system was used to evaluate the equivalence of difficulty between seven lists. Fifty-five college students with normal hearing participated in the study. Results The seven monosyllabic word lists had equivalent difficulty （ F = 1. 582,P 〉0. 05） to the subjects between each other and the system was proved as reliability and convenience. Conclusions The intelligent system has the feasibility in the clinical practice.