Perception of Japanese consonant-vowel syllables in reverberation: Comparing non-native listeners with native listeners

Abstract

Long reverberation degrades the intelligibility of speech sounds. Previous studies have reported that non-native listeners have difficulty in understanding speech in reverberation more than native listeners. In the results of the previous studies, there was the possibility that lower identification scores for non-native listeners were attributed to non-native phonemes which did not exist in their native languages. The current study investigated the identification of Japanese consonant-vowel (CV) syllables in reverberation for native English listeners whose native language has counterparts to most or all Japanese consonants. The current study used 62 CV syllables as stimuli. The reverberation time of the reverberant condition was 2.7 s. The results showed that the correct answer rate for non-native listeners declined in the reverberant condition more than that for native listeners. There was significant difference between native and non-native listeners in the correct answer rate of /m, r, k, d, s/ in reverberation. The results suggested that non-native listeners had disadvantage in listening to nonnative consonants even if their native languages had counterparts to the consonants. In addition, the results suggested that native English listeners might had advantage in finding acoustic cues of place of articulation in adverse environments because of the inventory of English.Ã The current paper includes the results for native listeners in Arai [1,2]. The parts of the results for non-native listeners in the current paper were preliminarily published [3]. y

Differences in speech intelligibility in noise between native and non-native listeners under ambisonics-based sound reproduction system

“…It is of no surprise that non-native listeners struggle more to understand speech that is not in their native language when compared to native listeners. Previous studies have found non-native listeners to perform poorer than native listeners in a range of speech perception measures, such as the ability to discriminate and identify phonetic contrasts, spoken words recognition and listening to unfamiliar and synthetic speech [1,2,3,4]. Even when a fluent non-native listener can understand speech similarly to a native listener in quiet and optimal conditions, studies have found that the non-native listener performs poorer in the acoustically adverse environments such as listening in noise [5,6,7,8,9] due to speech masking.…”