Click to download the sample audio files

The Humanities Research Institute (HRI) has developed and validated the Chung-Ang Auditory Database of Korean Emotional Speech (CADKES), a Korean emotional speech database that covers a range of vocal emotions collected at two levels of emotional intensity. Within CADKES, 18 actors were recorded using 25 sentences with strong and moderate intensities for six emotions (“neutral,” “happiness,” “sadness,” “anger,” “fear,” and “boredom”), resulting in 5400 sentences. The validation experiment involved emotion category identification and naturalness rating tasks. The data were subjected to a set of statistical analyses to assess identification accuracy, inter- and intra-rater reliabilities, the relationship between identification accuracy and naturalness degree, and the possible effects of intensity, listener gender, speaker gender, and emotion categories on emotion perception. 

An article on the CADKES was published in IEEE Access (Y. Nam and C. Lee, “Chung-Ang Auditory Database of Korean Emotional Speech: A Validated Set of Vocal Expressions with Different Intensities,” in IEEE Access, vol. 10, pp. 122745–122761, 2022). 

Sample audio files are available for download or playback on the HRI website. If you have any questions about the CADKES, please contact Youngja Nam at btoj19@cau.ac.kr or Chankyu Lee at leeck@cau.ac.kr.