Page 43 - Spring2019

course of maturation observed in behavioral speech-recog- nition data and improvements in executive control (reviewed by Crone, 2009).
Role of Linguistic Experience
and Knowledge
It has been suggested that the ability to use the information provided by the peripheral auditory system optimally re- quires years of experience with sound, particularly exposure to spoken language (e.g., Tomblin and Moeller, 2015). In a recent study, Lang et al. (2017) tested a group of 5- to 6-year- old children and found a strong relationship between recep- tive vocabulary and speech recognition when the masker was two-talker speech masker. As shown in Figure 2, children with larger vocabularies were more adept at recognizing sen- tences presented in a background of two competing talkers than children with more limited vocabularies. Results from previous studies investigating the association between vo- cabulary and speech recognition in a steady noise masker have been somewhat mixed (e.g., Nittrouer et al., 2013; Mc- Creery et al., 2017). The strong correlation observed by Lang et al. (2016) between vocabulary and speech recognition in a two-talker masker may reflect the greater perceptual and linguistic demands required to segregate and attend to target speech in a speech masker or to the spectrotemporally sparse cues available in dynamic speech maskers.
A second line of evidence that immature language abili- ties contribute to children’s increased difficulty recognizing speech when a few people are talking at the same time comes from studies that have compared children’s and adults’ ability to recognize speech based on impoverished spectral and/or temporal information (e.g., Eisenberg et al., 2000; Buss et al., 2017). For example, adults are able to recognize band-pass- filtered speech based on a narrower bandwidth than children (e.g., Eisenberg et al., 2000; Mlot et al. 2010). One interpreta- tion for this age effect is that children require more informa- tion than adults in order to recognize speech because they have less linguistic experience.
This hypothesis was recently tested by assessing speech recog- nition in a two-talker masker across a wide age range of chil- dren (5-16 years) and adults using speech that was digitally processed using a technique designed to isolate the auditory stream associated with the target speech (Buss et al., 2017). Children and adults showed better performance after the sig- nal processing was applied, indicating that sound source seg-
8 6
4 2
0
90
100 110
Receptive Vocabulary Score (PPVT)
Figure 2. Receptive vocabulary scores and thresholds for sentence recognition in a two-talker masker are shown for 30 young children (5-6 years) tested by Lang et al. (2017). There was a strong associa- tion between performance on these two measures (r = −0.75; P < 0.001), indicating that children with larger vocabularies showed bet- ter speech recognition performance in the presence of two competing talkers than children with smaller vocabularies. SNR, signal-to-noise ratio; PPVT, Peabody Picture Vocabulary Test.
regation negatively impacts children’s speech recognition in a speech masker. The child/adult difference in performance per- sisted, however, providing evidence of developmental effects in the ability to reconstruct speech based on sparse speech cues.
Implications
The negative effects of environmental noise on children’s speech understanding in the classroom are well documented, leading to the development of a classroom acoustics standard by the Acoustical Society of America (ASA) that was first ap- proved by the American National Standards Institute (ANSI) in 2002 (ANSI S12.60). Although this and subsequent stan- dards recognize the negative effects of environmental noise in the classroom on children’s speech understanding, they focus exclusively on noise sources measured in unoccupied classrooms (e.g., heating and ventilation systems, street traf- fic). The additional sounds typically present in an occupied classroom, such as speech, are not accounted for. As argued by Brill et al. (2018) in an article in Acoustics Today, meeting the acoustics standards specified for unoccupied classrooms might not be adequate for ensuring children’s speech under- standing in occupied classrooms, in which multiple people are often talking at the same time. This is problematic be- cause, as anyone who has spent time in a classroom can at- test, children spend most of their days listening and learning with competing speech in the background (e.g., Ambrose et al., 2014; Brill et al., 2018).
120 130
Spring 2019 | Acoustics Today | 41
Speech Reception Threshold (dB SNR)