A study of speech intelligibility and indoor environmental assessment in Hong Kong classrooms

Abstract

The indoor acoustical environment is not only related to productivity, health, and comfort, but also is related to acoustical quality in a space. The education of every citizen is essential to modern societies. Most formal education takes place in the classrooms, where a high level of acoustical quality is required. This thesis provides a systematic investigation of classroom speech intelligibility, sound field prediction, acoustical environment assessment, indoor environmental assessment with objective experiments, subjective questionnaires, and acoustic simulation methods in Hong Kong classrooms. To achieve the research objectives, several sub-works were conducted: (a) an investigation of the effects of speech transmission index (STI) on speech intelligibility; (b) the effects of acoustic descriptors on speech intelligibility; (c) a new combined sound field prediction method; (d) assessment model of acoustical environment; (e) assessment model of indoor environmental quality and its relationships with environmental factors. In order to investigate the effects of STI, speech intelligibility tests were conducted in 9 middle school classrooms and 11 university classrooms in Hong Kong. Meanwhile, objective acoustical measurements were performed in each listening position and testing conditions in each classroom. The relationship between subjective speech intelligibility scores (SI) an STI was discussed based on regression models. The effects of different age groups on the speech intelligibility were compared. The results show that SI increases with the increase of STI value for all age groups. The SI increase as the age increases under the same STI condition. The differences between age groups are decreased with the increase of STI values. English speech intelligibility scores in Hong Kong are always lower compared with native language studies under the same values of STI. Better STI values and better acoustical environment are needed because English is not the native language for students in Hong Kong but the official educational language. In order to investigate the effects of acoustical descriptors, Speech intelligibility tests were conducted in 9 secondary school classrooms and 18 university classrooms, and the acoustical measurements were performed in these classrooms. Subjective speech intelligibility tests were obtained from phonetically balanced (PB) word lists on a total of 672 students and acoustic descriptors such as signal-to-noise ratio (SNR), early decay time (EDT), and sound clarity (C80) were conducted in different listening positions in each classroom. The relationships between SI and acoustical descriptors were fitted based on non-linear curve fitting regression models. The "S" form regression model was selected with modification as the basic regression equation to describe the effects of SNR on speech intelligibility. The combination effects of SNR with reverberation condition and sound clarity condition on speech intelligibility were investigated. The impact of different age groups and linguistic environment on speech intelligibility were discussed. The results reveal that SI increases with the increase of SNR value for all age groups. The results indicate that nearly 0.06s increasing in EDT values will be correlated to a 1% decrease in SI. Furthermore, the results also suggest that a 1 dB increasing in C80 values will be correlated to a 1.23% increase in speech intelligibility scores. The SI increases as the age increases under the same SNR condition. The speech intelligibility scores are always lower than the comparison research results with a constant reverberation value as well as sound clarity value for an equal SNR value. Classroom acoustical parameters have a significant impact on speech intelligibility. In practice, applications of sound field predictions can provide the predicting level and spectral content of the sound in buildings, which are essential to acoustical design and acoustic environmental assessment. Therefore, a new combination method for sound field prediction is proposed for simulating sound fields during the whole audio frequency domain in small classrooms. An optimization approach based on the genetic algorithm is employed for optimizing the transition frequency of the combined sound field prediction method in classrooms. The selected optimization approach can identify the optimal transition frequency so that the combined sound field prediction can obtain more efficient and accurate prediction results. The proposed combined sound field prediction method consists of a wave-based method and geometric acoustic methods separated by the transition frequency. In low frequency domain (below the transition frequency), the sound field is calculated by the finite element method (FEM), while a hybrid geometric acoustic method is employed in the high frequency domain (above the transition frequency). The proposed combined prediction models are validated by comparing them with previous results and experimental measurements. The optimization approach is illustrated by several examples and compared with traditional combination results. Compared to existed sound field prediction simulations in classrooms, the proposed combination methods take the sound field in low frequencies into account. The results demonstrate the effectiveness of the proposed model. Apart from the speech intelligibility investigations and sound prediction methods mentioned above, the overall acoustical environment satisfaction evaluation was developed. An assessment model based on a multi-layer fuzzy comprehensive evaluation method (FCE) of the classroom acoustical environment is proposed. The model classifies five major factors affecting the overall assessment model into several subsets alternatives. The weightings of these main criteria and alternatives were collected through questionnaires among students based on the analytic hierarchy process methodology (AHP). An evaluation score was calculated from the proposed model with the weightings generated from the AHP method. It indicates that classrooms in PolyU need to be improved. The weightings generated from the AHP method can be considered for the importance of each alternative. The assessment model can provide proper recommendations to universities for acoustic treatment so as to increase the acoustic quality of the educational environment. As the acoustical environment is a key part of the indoor environment assessment. Indoor environmental quality (IEQ) is co-determined by several environmental factors (thermal, indoor air, lighting, and acoustics). In the last part, a four-layer IEQ assessment model for university classrooms was proposed based on fuzzy comprehensive evaluation (FCE) methods. The assessment model was evaluated based on a survey with a sample of 224 respondents in selected eight university classrooms in Hong Kong. Besides, objective measurements were performed in each classroom. Several parameters were included, such as operative temperature, CO2 concentration, illuminance level, and A-weighted background noise level in the measurements. Then a set of prediction formulas were proposed to illustrate the relationships between IEQ and the environmental factors. The analysis results showed that the quality of the thermal environment was the most essential factor in the indoor environment. The results also discussed the significance rankings of sub-factors based on the weightings calculated from the analytic hierarchy process (AHP). The methods can give proper suggestions to authorities to manage the appropriate treatment and improve the indoor environmental quality. It is also useful for indoor environment design based on the proposed prediction formulas.