Work

Synthesis of everyday conversational speech based on fine-tuning with a corpus for speech synthesis

Year: 2024

Type: article

Abstract: In this letter, we propose a separate modeling of prosodic and segmental features for everyday conversational speech synthesis, addressing challenges posed by low-quality recordings in the Corpus of Everyday Japanese Conversation (CEJC). Initially, the FastSpeech 2 model is trained on the conversation corpus and subsequently fine-tuned on a corpus for speech synthesis. Experimental results show that this fine-tuning approach enhances synthesis quality while preserving the nuances of everyday conversations.

Source: Nippon Onkyo Gakkaishi/Acoustical science and technology/Nihon Onkyo Gakkaishi

DOI: 10.1250/ast.e24.35

Authors: Hiroki Mori, Kota Furukawa

Institution: Utsunomiya University

Language: English

Topic: Speech Recognition and Synthesis

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: diamond

Funders: National Institute for Japanese Language and Linguistics, Japan Society for the Promotion of Science

Awards: Everyday conversation speech synthesis, Speech-laughの発生機序の解明