嗄声度の定量的指標に関する研究

概要

論文の詳細を見る
As the degree of judged hoarseness increases, the noise component of the spectrogram increases and replaces the harmonic structure. The relationship between these two components has been quantified as harmonics-to-noise (H/N) ratio. The computation is straightforward; 50 consecutive pitch periods of a sustained vowel /a/ are averaged; H is defined to be the energy of the average waveform, while N is the mean energy of the differences between the individual periods and the average waveform.The noise the present method measures is not only those components between the harmonic components of the spectrum but also those which overlap the harmonic components. In addition, cycle-to-cycle pitch perturbations (jitter) contribute to the noise component.The subjects of this study consisted of 42 normal and 89 pathological cases with varying degrees of hoarseness pre- and post-operatively.The H/N ratio of the normal group ranges from 7.0 to 17.0 dB with a mean of 11.9dB. We expect 95% of normal subjects to have a H/N ratio greater than 7.4dB. According to this criterion with a 5 % false alarm rate, 63.2% of the preoperative phonatory samples were differentiated from normals.The H/N ratio, the spectrographic classification and jitter showed a significant correlation with the psychophysical measurement of the degree of hoarseness (r=0.809, 0.805 and 0.712, respectively). The result suggested that jitter is less suitable for assessing the degree of hoarseness than the H/N ratio and the spectrographic classification. Since the spectrographic evaluation is subjective and arbitrary, the H/N ratio is the best quantitative index of the degree of hoarseness. The H/N ratio also proved useful for objective assessment of the result of treatment for hoarseness.
耳鼻咽喉科臨床学会の論文

著者

湯本英二
愛媛大学医学部耳鼻咽喉科学教室

嗄声度の定量的指標に関する研究

スポンサーリンク

概要

著者

関連論文

スポンサーリンク