Development and evaluatıon of the Turkısh mobıle real-noıse speech dıscrımınatıon test

Abstract

BackgroundWhile there is a list of monosyllabic words available for use in speaking tests in our country, the list of sentences is limited. The aim of this study is to evaluate the effectiveness of the mobile application we developed by creating a list of 4-word sentences that are frequently encountered in daily life, well-known, and spectrally balanced, to be used in speech tests in children aged 8–12 with normal hearing. A total of 50 participants, 25 girls (mean age 9.56 ± 1.56) and 25 boys (mean age 10.60 ± 1.55) with normal hearing, aged 8–12, were included in the study. Familiarity was evaluated by listening to a list of 200 sentences consisting of 4-word sentences to 70 children in 7 geographical regions of Turkey, and the sentences that 80% of the children did not understand were excluded. A list of 60 balanced sentences was determined by performing spectral and coherence analysis of the sentences. The recording of each sentence at 80 dB, 60 dB, and 40 dB intensity levels, respectively, increased in 5 dB steps between the − 10/ + 10 signal-to-noise ratio value and was combined separately with 3 different environmental noises (classroom, shopping mall, amusement park) and uploaded to the developed mobile application.ResultsFor all participants, speech comprehension scores at 80 dB, 60 dB, and 40 dB sound intensity were lowest in the amusement park environment. Speech comprehension scores are lowest in noise levels of 40 dB in amusement park and shopping mall environments. Speech comprehension scores are lowest in noise levels of 60 dB in the classroom environment. There is a statistically significant difference between ambient noises (p < 0.05). When male and female participants are compared, it is seen that male participants’ comprehension scores in noise in the shopping mall environment are higher than female participants. There is a statistically significant difference between genders (p < 0.05).ConclusionWhen the speech in noise tests developed in the literature are examined, even though the word lists are equalized with the phonemic equalization method, they do not reflect the reality of daily life. The Turkish Mobile Speech Discrimination in Real Noise 4-Word Sentence Test that we have developed will contribute to the literature. It is recommended that future research compare the results by adding different ambient noises.