Evaluation Of Chinese Speech Intelligibility In Elementary School Classrooms

Chinese speech intelligibility was investigated in28classrooms from9differentelementary schools in Guangzhou, China. Subjective and objective Chinese speechintelligibility were evaluated by children in grades2,4and6children (nominally7~8,9~10,and11~12years old) in total28elementary classrooms from these schools. SubjectiveChinese speech intelligibility scores and corresponding objective speech intelligibilityparameters such as speech transmission index (STI), A weighted signal to noise ratio (S/N(A)),and useful-to-detrimental sound ratio with different early sound limits (Ut) were calculated.The relationships between Chinese speech intelligibility scores and STI, S/N(A), Utfordifferent age groups were established. The results indicate that there are high correlationsbetween Chinese speech intelligibility scores and STI S/N(A), U50and U80for grades2,4and6children respectively. STI, S/N(A), U50and U80can evaluate Chinese speech intelligibility inelementary classrooms for children. The Chinese speech intelligibility scores are increasingwith the increase of STI, S/N(A), U50and U80, and increase with the age increasing. Under thesame STI condition, the Chinese speech intelligibility scores of children are lower that theseof adults. However the differences in Chinese speech intelligibility scores between childrenand adults are decreasing with the increase of STI. To obtain95%subjective Chinese speechintelligibility scores, the required value of STI, S/N(A), U50and U80for grade2is0.75,16.4dBA,7.4dB and10.1dB, for grade4is0.69,14.0dBA,4.5dB and6.9dB, and for grade6is0.63,11.2dBA,2.8dB and5.0dB.To investigate the effect of different types of noise on Chinese speech intelligibility ofchildren, two classrooms with a big difference in reverberation times (T30) were chosen.Subjective Chinese speech intelligibility was evaluated by children in grades2,4and6children in these two classrooms. Five different types of noises such as speech shaped noise,fan noise, impact noise, traffic noise, and babble noise were used at two signal-to-noise ratios(0dBA,10dBA) to change the listening conditions. The results showed that children’ssubjective Chinese speech intelligibility scores of the same age group in10dBA SNR werehigher than those in0dBA at the same T30, and the difference was more obvious when T30ishigher. Subjective Chinese speech intelligibility scores of the same age group with0.83s T30are higher than those with1.30s under the same SNR and noise type conditions. The lowerSNR is, the more obvious the difference is. Subjective Chinese speech intelligibility scoresare increasing with the increase of age under the same listening condition. The differences ofChinese speech intelligibility scores between different age groups were more obvious when the SNR is lower under the same T30. Under the impact noise condition, the subjectiveChinese speech intelligibility scores of different age groups were not less than those of theother4types of noise. Because that the high frequency energy of speech shaped noise andbabble noise are higher than the other three types of noise at the same A weighted sound level,children’s subjective Chinese speech intelligibility scores were lower than these of the other3types of noise under the longer reverberation and higher noise level condition.