AI RESEARCH
Learning-free L2-Accented Speech Generation using Phonological Rules
arXiv CS.CL
•
ArXi:2603.07550v1 Announce Type: new Accent plays a crucial role in speaker identity and inclusivity in speech technologies. Existing accented text-to-speech (TTS) systems either require large-scale accented datasets or lack fine-grained phoneme-level controllability. We propose a accented TTS framework that combines phonological rules with a multilingual TTS model. The rules are applied to phoneme sequences to transform accent at the phoneme level while preserving intelligibility. The method requires no accented.