Changes in the fundamental frequency and intensity were measured for utterances of the five Japanese vowels and a sentence of greeting and a song by 42 deaf and hard of hearing children (aged 6 to 11), and their characteristics were studied in comparison with those of 9 normal children (aged 7 to 12). Characteristic differences between defective speech and normal speech were found in the relationships between the following : variation among fundamental frequency of each of the five vowels for each subject and his hearing loss, average fundamental frequency in greeting for each subject and his age, the total duration and the duration of voiced portions in greeting, and accuracies of rhythm and note in singing. Correlation between these physical characteristics. physiological characteristics and the evaluations of the speech of deaf and hard 00 hearing children was examined by means of principal component analysis.

View full abstract

First and second formant frequencies (FI and F2) of the five Japanese vowels uttered in isolation by 35 deaf and hard of hearing children (aged 6 to 11) were measured utilizing an audio spectrograph, and the results were compared with those of 8 normal children (aged 7 to 12). The distribution of the formant frequencies of each kind of vowel on the Fz - F2 plane shows that those of the afflicted children deviate significantly from the normal range especially for /i/ and /0/. The tendency toward perceptual confusion regarding the type of vowel corresponds well to the deviation of the formant frequencies. The relations between the formant frequencies of the five Japanese vowels of each of the defective children were classified into the following six types : F2 range reduced, F2 range reduced and rotated, F2 range reduced and /o/ and /a/ close together, F2 range reduced and /i/ and /e/ close together, F2 range neutralized, and F2 range reduced with FI raised. Based on a muscle contraction model of the speech organs, the results can be explained in terms of the nature of the imperfection of the articulatory movement.

View full abstract

The functional properties of vibratory sense of skin were studied for a deaf aid by means of psychophysical experiments with respects to the two-point threshold, intensity difference limen and masking. From the experiments, both two-point threshold and difference limen were minimum at the finger tip. The auditory masking characteristics by FM tone and the tactile masking by scanning vibratory stimulation were similar to each other. Based on The above results, a small tactile vocoder, in which the vibration pattern was produced by a 16x3 array stimulators of 200Hz that can be fitted on a single finger tip, was designed using highly efficient piezoelectric vibrators. After 20 to 30 minute training time, the rate of correct response was 91% in the articulation test for vowels and 66% for monosyllables /ku/, /su/, /nu/, /hu/, /ru/.

View full abstract

The 2-dimensional pitch-intensity indicator is an apparatus for use by the deaf who train to control vocal chord vibration efficiently by sustained utterances. By means of this indicator, the pitch (fundamental frequency) signal is extracted using an improved "Gruenz-type" pitch detector in which the time constant of integration is varied automatically to fit the speech signal. Pitch and intensity signals are displayed on a 12" CRT screen. The training was conducted for about 5 weeks with 8 deaf children (7-9 years old) by a teacher of the school for the deaf. The results of the analysis of the training effect are as follows: (1) The utterances of the deaf children trained with the indicator are superior in stability of pitch and intensity to those of the deaf children without it. (2) Three months after the end of training, it was shown that the utterances of the deaf children the indicator became more stable than at the end of training. These results can also be verified by listening to the utterances of the children.

View full abstract

This is a study of a speech training apparatus for the deaf. This speech trainer which makes, full use of eyesight instead of hearing, is expected to give excellent training, by comparing the pronunciations of normal persons with those of the deaf. In summary, the pitch extractor is constracted from an integrator and squarer and other logic blocks, and a change in the speaker's pitch frequency can be displayed in real time. A spectral display system effective for extraction of the voice characteristics, was developed for this purpose. This system is a device to project on a Brauntube, a spectral pattern of the voice. further, a vowel decision system is devised that can decide the clearness of the vowels.

View full abstract

This report deals with the development of a speech synthesizer for a physically very handicapped person. The synthetic method of this apparatus is on the basis of simulation of speech mechanism. By operating 27 keys of alphabetical letters, any Japanese sentence can be synthesized. This apparatus consists of an operating unit and a synthesizing unit ; the former is connected with the latter by means of wireless device and this is portable. A vocal tract simulater is composed on the basis of T-networks, in which an operational amplifier and analog multiplier are used as composing elements. Furthermore, the control parameters of the vocal tract simulater can represent any vocal tract configulation. A simple compensating method is found out to improve the quality factor Q of networks in the vocal tract simulater. As a computer is not used to control the synthesizer, this apparatus is comparatively economic in its construction.

View full abstract

A distinctive dominance effect of the steady-state vowel /a/ sound in the non-verbal hemisphere in Western Ss was found by Tsunoda's tapping test. In contrast to these results, several studies carried out with Western Ss using Kimura's dichotic listening test did not show any distinctive tendency for the dominance effect of vowel sounds. In this experiment, the author intended to set up a tapping test analogous to the Kimura's test using syllables. Test subjects were 12 Westerners and 6 second-generation Japanese. The cerebral dominance of vowel /a/ without loading of /doop/ was non-dominant side, whereas in the loading condition, the original pattern disappered completely and a new pattern lateralized to the dominant hemisphere was observed. The grade of the dominancy shift of vowel sounds in the dichotic test mainly depends upon the experimental condition. If a technique free from the influence of the contralateral shift had been applied, more clear-cut results could have been obtained.

View full abstract

The transmission of speech by the usual microphone is almost impossible when the logarithm of the signal-to-noise ratio is negative. In order to obtain intelligible speech in such a case, it is possible to apply a method of speech transmission picking up directly the vibration of the articulatory organs using a vibration transducer which is insensitive to air-coupled noise. In this paper, it is experimentally confirmed that the speech spectrum on the forehead is the most similar to the airborne speech spectrum among the spectra obtained for the vibration of each part of the body. It is found that an articulation score of sixty percent is obtained using a vibration pickup on the forehead even when the signal-to-noise ratio is -35 dB, and furthermore the naturalness of the speech is improved with a compensating filter.

View full abstract

Instead of acquiring speech data in diving experiments involving many difficulties, synthetic vowels under 15 combinations of sound velocity c and density ρof the breathing gas were prepared for a listening test of vowel identification. Error rates and confusion matrices of the listeners' responses are discussed with respect to c and p. The main points of the results are as follows: 1) Although, the intelligibility of speech in a helium environment is greatly improved by restoring the distortion by c (an upward shift of formants proportional to c), the additional restoration of the distortion by p (non-linearity of the upward shift) becomes indispensable at a depth of 500m more. 2) Uni-directional confusion of the vowels pronounced in a helium atmosphere is also explained by the non linear upward shift of the formants.

View full abstract