Voice Conversion for Likability Control via Automated Rating of Speech Synthesis Corpora

Hitoshi, Shinnosuke, Satoru · 2025

Read Voice Conversion for Likability Control via Automated Rating of Speech Synthesis Corpora by Hitoshi, Shinnosuke, Satoru (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-01356.txt.

Description

Suda, Takamichi, and Fukayama convert voices toward likability while keeping the speaker who they are. Their 2025 paper notes that a likable voice smooths social dealings from dating to advertising. A tool that serves sample likable voices for each audience could coach speakers in real time.

Labels are the bottleneck, so a likability predictor is trained on an existing rated dataset. That predictor then auto-rates a large synthesis corpus, scaling training data cheaply. Predictor scores track human ratings closely. Listener and metric tests show converted speech gains likability while identity and words stay intact. The authors present likability as a dial that voice conversion can now turn on purpose, not a fixed trait of a voice.

Catalog record

CREATORHitoshi, Shinnosuke, Satoru
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE20 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-01356

Browse the full library · About the Tower of Babel