Conversational Speech Naturalness Predictor

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Xu, Anfeng, Gaur, Yashesh, Kanda, Naoyuki, Ouyang, Zhicheng, Zmolikova, Katerina, Raj, Desh, Merello, Simone, Sun, Anna, Kalinli, Ozlem
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866915827360464896
author Xu, Anfeng
Gaur, Yashesh
Kanda, Naoyuki
Ouyang, Zhicheng
Zmolikova, Katerina
Raj, Desh
Merello, Simone
Sun, Anna
Kalinli, Ozlem
author_facet Xu, Anfeng
Gaur, Yashesh
Kanda, Naoyuki
Ouyang, Zhicheng
Zmolikova, Katerina
Raj, Desh
Merello, Simone
Sun, Anna
Kalinli, Ozlem
contents Evaluation of conversational naturalness is essential for developing human-like speech agents. However, existing speech naturalness predictors are often designed to assess utterances from a single speaker, failing to capture conversation-level naturalness qualities. In this paper, we present a framework for an automatic naturalness predictor for two-speaker, multi-turn conversations. We first show that existing naturalness estimators have low, or sometimes even negative, correlations with conversational naturalness, based on conversational recordings annotated with human ratings. We then propose a dual-channel naturalness estimator, in which we investigate multiple pre-trained encoders with data augmentation. Our proposed model achieves substantially higher correlation with human judgments compared to existing naturalness predictors for both in-domain and out-of-domain conditions.
format Preprint
id arxiv_https___arxiv_org_abs_2603_01467
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Conversational Speech Naturalness Predictor
Xu, Anfeng
Gaur, Yashesh
Kanda, Naoyuki
Ouyang, Zhicheng
Zmolikova, Katerina
Raj, Desh
Merello, Simone
Sun, Anna
Kalinli, Ozlem
Audio and Speech Processing
Evaluation of conversational naturalness is essential for developing human-like speech agents. However, existing speech naturalness predictors are often designed to assess utterances from a single speaker, failing to capture conversation-level naturalness qualities. In this paper, we present a framework for an automatic naturalness predictor for two-speaker, multi-turn conversations. We first show that existing naturalness estimators have low, or sometimes even negative, correlations with conversational naturalness, based on conversational recordings annotated with human ratings. We then propose a dual-channel naturalness estimator, in which we investigate multiple pre-trained encoders with data augmentation. Our proposed model achieves substantially higher correlation with human judgments compared to existing naturalness predictors for both in-domain and out-of-domain conditions.
title Conversational Speech Naturalness Predictor
topic Audio and Speech Processing
url https://arxiv.org/abs/2603.01467