32. Kobe University Japanese-Chinese Comparative MRI Movies corpus (KUJC-MRI)

Data DOI


Producer, Project

Graduate School of Intercultural Studies, Kobe University

(Principal Investigator: Prof. Toshiyuki Sadanobu, Co-Investigator: Prof. Shunyue SHU and Prof. Ryoko Hayashi)

Grant-in-Aid for Scientific Research (A) "Developing teaching materials for education of spoken Japanese based on contrastive studies among Japanese English and Chinese" supported by JSPS, Japan during FY 2004–2006.

Grant-in-Aid for Scientific Research (A) "Speech grammar based on speaker's characters" supported by JSPS, Japan during FY 2007–2010.

Grant-in-Aid for Scientific Research (A) "Study on spoken Japanese based on situation and its pedagogical application" supported by JSPS, Japan during FY 2011–2015.

"Comprehensive research project on spoken words" supported by Graduate School of Intercultural Studies, Kobe University during FY 2012-2014.


This database contains images and movies of articulatory movements by Japanese and Chinese speakers. The data were recorded by a medical MRI system using an external trigger synchronous imaging method.

  Vol.1:Single Japanese sounds (uttered for each "SEI-ON," "DAKU-ON," "YO-ON" and each set of five syllables)
  Vol.2:Japanese vowel sounds / Chinese basic vowel sounds (single vowel sounds uttered individually)
  Vol.3:Japanese geminate consonants and moraic nasals(iQpai, saNsai, etc.)
  Vol.4:Similar sounds within the Japanese language (猶予yuuyo/言うよiuyo, 肩kata/勝ったkatta, etc.)
  Vol.5:Similar sounds in the Japanese and Chinese languages (J_曖昧aimai/C_暧昧ai4mei4, etc.)
  Vol.6:Japanese Tongue Twister ("Bugu-bagu")*Japanese speakers only

The speakers were native speakers of Japanese (Tokyo and Kinki dialects) and native speakers of Chinese (Mandarin and Northeast dialect). The Chinese speakers were also learners of Japanese, so they uttered Japanese utterances in addition to Chinese utterances.


1-6 speakers depending on the task.

Japanese native speakers: 5 Tokyo dialect speakers and 3 Kinki dialect speakers (4 males and 4 females)

Chinese native speakers (Japanese learners): 3 Chinese Mandarin speakers and 3 Chinese Northeastern speakers (2 males and 4 females)

Recording environment

MRI system

(before 2007) Shimazu Marconi, MAGNEX ECLIPSE PowerDrive250 (1.5T)

(after 2008) Siemen.s., MAGNETOM Verio (3T)

Video file format

Sagittal plane, 2–8 slices (depending on task)

- Movie: MP4 format(256x256/512x512 pixels, sound is uncompressed 44.1kHz, about 60/30/15fps depending on the task)

- Still image: TIFF format (8bit, 256x256 pixels)

- Sound: WAV format (44.1kHz, 16bit, Stereo) *One or more utterances in a repetitive speech

Distribution media



For research purpose only


No fee


All documents are written in Japanese.

Sample data

Japanese male speaker, Japanese vowels "aiueo"

Go to corpora list