PulseAugur
EN
LIVE 07:44:41

New PASQA model targets pitch-accent errors in speech quality assessment

By PulseAugur Editorial · [2 sources] ·

Researchers have developed PASQA, a new model designed to specifically assess the quality of speech by focusing on pitch-accent correctness. Unlike existing models that predict general naturalness, PASQA is trained on a synthetic Japanese dataset with introduced accent errors. This approach allows PASQA to more accurately evaluate accent quality, even with unseen speakers, and shows better agreement with human judgments on accent correctness. AI

IMPACT This model could improve the evaluation of synthetic speech and speech synthesis systems by focusing on a critical aspect of naturalness: pitch-accent accuracy.

RANK_REASON The cluster contains an academic paper detailing a new model for speech quality assessment.

Read on arXiv cs.CL →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

New PASQA model targets pitch-accent errors in speech quality assessment

COVERAGE [2]

  1. arXiv cs.CL TIER_1 English(EN) · Masaya Kawamura, Yuma Shirahata, Kentaro Mitsui, Reo Shimizu ·

    PASQA: Pitch-Accent-Focused Speech Quality Assessment Model Trained on Synthetic Speech with Accent Errors

    arXiv:2606.20137v1 Announce Type: cross Abstract: Existing mean opinion score (MOS) prediction models typically predict utterance-level naturalness MOS and can be insensitive to localized pitch-accent errors. We propose Pitch-Accent-focused Speech Quality Assessment (PASQA), whic…

  2. arXiv cs.CL TIER_1 English(EN) · Reo Shimizu ·

    PASQA: Pitch-Accent-Focused Speech Quality Assessment Model Trained on Synthetic Speech with Accent Errors

    Existing mean opinion score (MOS) prediction models typically predict utterance-level naturalness MOS and can be insensitive to localized pitch-accent errors. We propose Pitch-Accent-focused Speech Quality Assessment (PASQA), which explicitly targets pitch-accent correctness. To …

RELATED ENTITIES

RELATED TOPICS