抄録
It is sixteen years since a new method of articulation hearing test was published, and during these years many studies have been made about it, that is, studies on the test words from the quantitative and qualitative points of view, on the relationship between the articulation curve and the kind of hearing impairment, on the relationship between pure tone audiogram and speech audiogram, and other studies.
In 1957, a standardized word list was enacted in Japan which consists of 36 of one figures for measurement of speech reception threshold and 50 monosyllables for speech discrimination test.
Five reports which dealt with the articulation test by means of one hundred monosyllables which included all representative Japanese monosyllables except for phonetically similar ones were investigated. Comparing the discrimination score with each other settled in file (consonants) and rank (vowels) in these five reports,“y” file (ya, yu, yo) showed a remarkably low discrimination score in Tonotsuka's report, all the files except “n” file (na, ni, nu, ne, no) were in fairly good agreement in Fukushima's and Ebihara's results, and “s” and “t” files showed a low discrimination score in Goto's report. Much difference was recognized in “n” and “h” files in all the reports. Remarkable discordance in them about contracted sounds (kya, sha, etc.) was observed, but it was difficult to find its general tendency of mistaking them along a certain line. Concerning the voiceless group,“i” rank showed a remarkably good discrimination score in Fukushima's report. About contracted sounds,“u” rank showed a bad score in Tonotsuka's report while a good score in Goto's.
Different monosyllables have their own tendency of being well or ill heard. When test words are recorded and played back, the personal quality of a speaker's voice, the intensity of each monosyllable and the surrounding noise blur the distinctive peculiarities of the monosyllable and exercise on the discrimination score a powerful influence.
To clarify the personal tendency of hearing a monosyllable, the author performed the following examinations.
The experimental methods:
The subject: ten healthy students, 13 to 22 years old, whose hearing was normal at the air conduction test. Five of them were male and others female.
The pure tone audiometry: to check the changes of pure tone threshold during the discrimination test, the pure tone air conduction audiometry was performed before discrimination test. The difference between the arithmetic mean of hearing loss of the first time for 500, 1000 and 2000 cps and that of the others never exceeded 10 dB.
The speech discrimination test: 100 monosyllables which included all representative Japanese monosyllables except for phonetically similar ones were recorded on a magnetic tape by a male and a female announcer of NHK (Nippon Broadcasting Corporation). Changing the order of monosyllables, 5 tables for each voice were prepared. The interval of pronunciation was decided to be 2 sec. because the subjects were students with normal hearing and the program of the test was very long.
The discrimination test was performed by 10 dB step from 10 dB to 50 dB, and repeated 10 times by male voice and the same times by female voice on each ear.
The examination was performed in the speech audiometric laboratory in the Otorhinolaryngological Clinic of Kyushu University.
The results of the test at 20, 30 and 50 dB for each voice were settled in a confusion matrix. The results at 10 dB and 40 dB of 5 female subjects were settled in 2 confusion matrixes concerning the female voice only.
The results
1. The articulation scores of a certain monosyllable showed a wide variation even when the normal ear was tested by the same voice, using the same equipment. Wider variation was recognized when female voice was used.
