A corpus of mandarin pronunciation by Uyghur college students

In research on the acquisition patterns of mandarin by ethnic minority groups and the development of related speech technologies, high-quality and highly targeted pronunciation corpora serve as an important foundation. However, systematic corpora for mandarin pronunciation specifically designed for Uyghur college students remain scarce, making it difficult to meet the needs of in-depth analysis and practical applications of their pronunciation characteristics. To address this gap, this study constructs a dedicated corpus of mandarin pronunciation by Uyghur college students. The corpus includes recordings from 48 Uyghur college students from southern Xinjiang as primary speakers, with 2 Han students selected as reference speakers for standard mandarin pronunciation. Based on the “Outline for Putonghua Proficiency Test”, we carefully selected experimental materials to systematically cover 39 finals, 21 initials, as well as monosyllabic and disyllabic tone patterns in mandarin. Speech data were recorded using professional equipment in a standardized acoustic environment. After segmentation and denoising, the corpus comprises 14,550 audio files (with a total duration of 146 minutes and a data volume of 595.6MB), along with corresponding transcription files. The corpus strictly controls character selection principles and pronunciation variables, providing high-quality data support for analyzing the mandarin pronunciation patterns of Uyghur college students, developing mandarin pronunciation teaching resources, and optimizing related speech technologies. Overall, the corpus holds great value for promoting language education in ethnic minority regions and cross-lingual speech and information processing research.