A Native-Reference Phone-Class Geometry for Second-Language Pronunciation Analysis

2026-09-24

AI Summary

This study addresses the lack of interpretability and the reliance on paired recordings in automatic pronunciation quality assessment. We propose a novel label-free, unpaired native-reference geometric approach. Specifically, features are extracted via self-supervised representation learning, and a native reference geometric space is constructed by combining singular value decomposition with context-dependent phoneme-class averaging. Second-language speech is then projected onto this space to directly quantify pronunciation deviation. Experiments on two datasets demonstrate that the proposed deviation distance exhibits a significant negative correlation with overall fluency and pronunciation quality. This work thereby achieves interpretable pronunciation analysis without requiring paired recordings.

Abstract

Automatic speaking assessment systems can provide holistic proficiency scores, but often lack interpretable measures that characterize pronunciation quality. We propose a native-reference phone-class geometry for measuring second language (L2) pronunciation deviation without requiring pronunciation labels, read-aloud prompts, or matched recordings of the same text from native and L2 speakers. Given a native speech corpus, we average frame-level self-supervised representations for each context-dependent phone-class and use singular value decomposition (SVD) to derive a compact native-reference coordinate system. For each L2 utterance, we compute the corresponding averages and project them into the native-reference space. We then demonstrate that the distances between L2 and native-reference coordinates for matched phone-classes show consistent negative correlations with holistic speaking proficiency on the Dev subset of the Speak and Improve Corpus 2025 (Spearman's $ρ\!=\!-0.53$) and with pronunciation quality on the learner subset of the English Read by Japanese Students dataset ($ρ\!=\!-0.34$). These findings suggest that the proposed geometry captures acoustic-phonetic information relevant for proficiency rating while remaining applicable to spontaneous L2 speech without matched native recordings.

Problem

• Second-language pronunciation analysis
• Automatic speaking assessment
• Pronunciation deviation
• Interpretable measures
• Speaking proficiency

Innovation

• Self-supervised representations
• Phone-class geometry
• Singular value decomposition
• Second-language pronunciation assessment
• Native-reference space