Abstract
A series of experiments dealing with sound quality judgments has been carried out over a long period in the Laboratory of Audiology at Nagoya University, in order to show that vowels possess two fundamental qualities, one forming the basis of phoneme identification and the other of voice identification. We present some results of quality experiments in which Japanese vowels and Japanese speakers are used and in which distortion of the band-elimination type is employed. The most general conclusion is that voice identification is far more suitable than phoneme articulation as a criterion in the technological problem of estimation or rating of speech-communication systems and speech-processing devices.
Get full access to this article
View all access options for this article.
