Durham School of Architectural Engineering and Construction

 

Date of this Version

12-2014

Document Type

Article

Comments

A DISSERTATION Presented to the Faculty of The Graduate College at the University of Nebraska In Partial Fulfillment of Requirements For the Degree of Doctor of Philosophy, Major: Architectural Engineering, Under the Supervision of Professor Lily M. Wang. Lincoln, Nebraska: December, 2014

Copyright (c) 2014 Zhao Ellen Peng

Abstract

The current classroom acoustics standard (ANSI S12.60-2010) recommends core learning spaces not to exceed background noise level (BNL) of 35 dBA and reverberation time (RT) of 0.6 second, based on speech intelligibility performance mainly by the native English-speaking population. Existing literature has not correlated these recommended values well with student learning outcomes. With a growing population of non-native English speakers in American classrooms, the special needs for perceiving degraded speech among non-native listeners, either due to realistic room acoustics or talker foreign accent, have not been addressed in the current standard. This research seeks to investigate the effects of BNL and RT on the comprehension of English speech from native English and native Mandarin Chinese talkers as perceived by native and non-native English listeners, and to provide acoustic design guidelines to supplement the existing standard.

This dissertation presents two studies on the effects of RT and BNL on more realistic classroom learning experiences. How do native and non-native English-speaking listeners perform on speech comprehension tasks under adverse acoustic conditions, if the English speech is produced by talkers of native English (Study 1) versus native Mandarin Chinese (Study 2)? Speech comprehension materials were played back in a listening chamber to individual listeners: native and non-native English-speaking in Study 1; native English, native Mandarin Chinese, and other non-native English-speaking in Study 2. Each listener was screened for baseline English proficiency level, and completed dual tasks simultaneously involving speech comprehension and adaptive dot-tracing under 15 acoustic conditions, comprised of three BNL conditions (RC-30, 40, and 50) and five RT scenarios (0.4 to 1.2 seconds).

The results show that BNL and RT negatively affect both objective performance and subjective perception of speech comprehension, more severely for non-native listeners than for native listeners. While the presence of foreign accent is generally detrimental, an interlanguage benefit was identified on both speech comprehension and the self-report frustration and perceived performance ratings, specifically for non-native listeners with matched foreign accent as the talker. Suggested design guidelines for BNL and RT are identified for attaining optimal speech comprehension performance to improve classroom acoustics for the non-native English-speaking population.

Advisor: Lily M. Wang

Share

COinS