This study presents the results of an acoustic analysis of Oromo vowels of the northern dialect. The data for the study were collected from 19 speakers (nine female, 10 male) who produced the vowels in the same phonetic environment. Such acoustic measures as duration, fundamental frequency and the first three formant frequencies were extracted for the analysis. In a linear mixed-effects model, each acoustic parameter was modelled as a function of the fixed effects such as gender and vowel quality, and of participants’ random effects. The model shows a main effect of gender on all the acoustic measures, with the female speakers producing the vowels with significantly greater duration and formant frequencies. The model also indicates the main effect of vowel quality on all the acoustic measures with the exception of duration. The proportion of variances explained by gender and vowel quality is found to be large for fundamental frequency and the first formant frequency respectively. As regards the classification of the vowels, Support Vector Machine reveals that the time-varying frequency does not have an advantage over a steady state in separating vowels of both genders. However, it is generally fairly effective in classifying vowels of the dialect of the language.

本研究介绍了北部方言奥罗莫元音的声学分析结果。该研究的数据收集自 19 名说话者（9 名女性，10 名男性），他们在相同的语音环境中发出元音。提取持续时间、基频和前三个共振峰频率等声学测量值进行分析。在线性混合效应模型中，每个声学参数都被建模为固定效应（例如性别和元音质量）以及参与者随机效应的函数。该模型显示了性别对所有声学测量的主要影响，女性说话者发出的元音具有明显更长的持续时间和共振峰频率。该模型还表明元音质量对除持续时间之外的所有声学测量的主要影响。发现由性别和元音质量解释的方差比例对于基频和第一共振峰频率分别很大。对于元音的分类，支持向量机表明，时变频率在分离两性元音方面并不比稳态具有优势。然而，它在对语言方言的元音进行分类方面通常相当有效。