Suggestions of Multi-Talker Babble Noise Using Korean Speech Sound

Lee, Kyoung Won; Lee, Jae Hee; Lee, Junhak

doi:2008.4.1.24

Audiol > Volume 4(1); 2008 > Article

Research Paper

Audiology 2008;4(1): 24-27.
Published online: June 30, 2008.
doi: http://doi.org/10.21848/audiol.2008.4.1.24

한국어음을 이용한 다화자잡음의 개발시안

이경원^1,2,3, 이재희^1,2,4, 이정학^1,2

¹한림국제대학원대학교 청각학과
²한림국제대학원대학교 부설 청각언어연구소
³한림대학교 일반대학원 청각학전공
⁴Speech and Hearing Sciences, Indiana University

Suggestions of Multi-Talker Babble Noise Using Korean Speech Sound

Kyoung Won Lee^1,2,3, Jae Hee Lee^1,2,4, Junhak Lee^1,2

¹Department of Audiology, Hallym Institute of Advanced International Studies, Seoul, Korea,
²Hallym Institute of Audiology & Speech Pathology, Hallym Institute of Advanced International Studies, Seoul, Korea,
³Graduate Program in Audiology, Hallym University, Chuncheon, Korea,
⁴Department of Speech and Hearing Sciences, Indiana University, Indiana, U.S.A

Correspondence

Kyoung Won Lee ,Tel: (016) 661-2087, Fax: (02) 3453-6618, Email: leekw@hallym.ac.kr

Received: May 14, 2008; Accepted: June 2, 2008. Published online: June 30, 2008.

ABSTRACT

The purpose of the present study was to construct a multitalker babble stimuli based on Korean speech as well as to compare the spectrum of the Korean babble stimuli with that of English babble noise. The development of this multitalker babble noise would also provide better insights into the evaluation of hearing assistive devices, the evaluation of central auditory processing disorder (CAPD), and auditory training for Korean hearing-impaired listeners. Sixteen young normal-hearing speakers (8 males and 8 females) whose first language is Korean participated. All the participants were told to speak the sentences used in standardized Korean speech audiometry, newspaper, and magazine with conversational level for about two minutes. Each sentence was normalized to have the same RMS energy before forming Korean babble noise. The long-term average speech spectra (LTASS) of all utterances were analyzed with computerized speech lab (CSL) that is one of the leading speech analysis systems for acoustic phonetics. Results have showed that there were no differences among Korean LTASS of sentences, newspaper, and magazine (p<0.05). As expected, we found significant differences in spectral features between male-spoken and female-spoken LTASS. The level difference of the Korean LTASS was approximately 37 dB between low and high frequencies, indicating larger intensity in Korean that in English. Differences of 16 dB were also found in the LTASS of between Korean and English multitalker babble noise, suggesting the possible differences of acoustic properties between Korean and English speech.

Key Words: Babble noise·Multi-talker·LTASS (long term average speech sound).

중심단어: 다화자잡음· Multi-talker babble· 장기평균어음스펙트럼· LTASS.