30. Japanese Empathetic Dialogue Speech Corpus (STUDIES)

Data DOI

Producer, Project

Assist. Prof. Yuki Saito, The University of Tokyo

This corpus contains simulated Japanese dialogue empathetically uttered to the interlocutor by three voice actors. Dialogue lines were collected through crowdsourcing.

Chatting with a teacher and students
For the purpose of AI Tutor's speech synthesis, dialogues were recorded based on a situation where a female teacher of a cram is chatting with her students (male or female) in between studying.

The 8 hours of speech data contain 150 long dialogues (10-20 turns) and 720 short dialogues (4 turns).

This corpus also includes read speech of ITA corpus (phoneme balance sentences) by one female who acted teacher.
Call center dialogues －NEW
For the purpose of AI call-center operator's speech synthesis, dialogues were recorded based on a situation where a female operator in a call center talk to a customer.

The speaker was one female voice actor, the same as the teacher in 1. (Only the speech of the operator was recorded.)

The 6.5 hours of speech data contain 820 situation-oriented complaints handling dialogues (2-12 turns) and 600 positive attentive-listening dialogues (4 turns).

Speaker

One professional male speaker and two professional female speakers

Recording environment

Studio

Speech file format

WAV format (48 kHz, 16 bit, Mono)

Distribution media

1 DVD

Licensing

For research purpose only

Price

No fee

Further information

http://sython.org/Corpus/STUDIES/

http://sython.org/Corpus/STUDIES-2/

Speech sample for test listening