Logistic modeling of regional variability in Russian vowels

Abstract

This paper investigates the acoustic and prosodic-structural features of Russian vowels in South Russian and Moscow phonetic variants. The aim of the study is to identify key features that distinguish South Russian and Moscow phonetic variants using machine learning methods. The research, based on the reading of neutral phrases by native speakers from Moscow, Moscow Oblast, and Southern Russia, involved a comprehensive analysis of speech data: annotation, acoustic measurements, construction of a logistic regression model, and assessment of feature permutation importance. Scientific novelty lies in the quantitative assessment of the contribution of various speech parameters – temporal, spectral, and structural – to the task of binary classification of regional pronunciation variants. The study revealed that the main differentiating feature is the temporal characteristics of vowels: absolute duration is significantly associated with the Moscow variant, whereas the relative duration of the vowel within the segment structure is associated with the South Russian variant. Spectral parameters demonstrated minimal differentiating potential.

Research materials

1. Boersma P., Weenink D. Praat: doing phonetics by computer (Version 6.1.03) (Computer software). 2025. http://www.praat.org/

References

1. Бондарко Л. В. Фонетика современного русского языка: учеб. пособие для студентов вузов. СПб.: Изд-во Санкт-Петербургского университета, 1998.
2. Златоустова Л. В. Фонетическая структура слова в потоке речи. Казань: Изд-во Казанского университета, 1962.
3. Норма произношения в узусе и кодификации: коллективная монография / отв. ред. М. Л. Каленчук, Д. М. Савинов. М., 2021.
4. Трубецкой Н. С. Основы фонологии. М.: Аспент пресс, 2000.
5. Фант Г. Акустическая теория речеобразования / пер. с англ. Л. А. Варшавского и В. И. Медведева; под ред. В. С. Григорьева. М.: Наука, 1964.
6. Arkin G., Abdukelim T., Yilahun H., Hamdulla A. Machine learning-driven acoustic feature classification and pronunciation assessment for Mandarin learners // Applied Sciences. 2025. Vol. 15 (11). https://doi.org/10.3390/app15116335
7. Peterson G. H., Barney H. L. Control methods used in a study of the vowels // Journal of the Acoustical Society of America. 1952. Vol. 24 (2). https://doi.org/10.1121/1.1906875
8. Tanner J., Sonderegger M., Stuart-Smith J. Multidimensional acoustic variation in vowels across English dialects // Proceedings of the 19th SIGMORPHON Workshop. 2022. https://doi.org/10.18653/v1/2022.sigmorphon-1.8
9. Themistocleous C., Protopsaltis P. Dialect classification from a single sonorant sound using deep neural networks // Frontiers in Communication. 2019. Vol. 4. https://doi.org/10.3389/fcomm.2019.00064

Author information

About this article

Publication history

• Received: November 9, 2025.
• Published: November 28, 2025.

Keywords

• региональная вариативность
• русские гласные
• акустическая фонетика
• логистическая регрессия
• темпоральные характеристики гласных
• regional variability
• Russian vowels
• acoustic phonetics
• logistic regression
• temporal characteristics of vowels

Copyright

© 2025 The Author(s)
© 2025 Gramota Publishing, LLC

User license

Creative Commons Attribution 4.0 International (CC BY 4.0)