Speech Accent Archive

The Speech Accent Archive is a dataset that contains recordings of speakers from various linguistic backgrounds, used to evaluate the performance of speech separation models on different accents and languages not seen during training.

Papers using Speech Accent Archive (3)

Modeling L1 Influence on L2 Pronunciation: An MFCC-Based Framework for Explainable Machine Learning and Pedagogical Feedback2025 · 2 citesSvarah: Evaluating English ASR Systems on Indian Accents2023 · 3 citesImproving the transferability of speech separation by meta-learning2022