Ken-Ichi Sakakibara

A Method for Automatic Detection of Vocal Fry

Vocal fry (also called creak, creaky voice, and pulse register phonation) is a voice quality that carries important linguistic or paralinguistic information, depending on the language. We propose a set of acoustic measures and a method for automatically detecting vocal fry segments in speech utterances. A glottal pulse-synchronized method is proposed to deal with the very low fundamental frequency properties of vocal fry segments, which cause problems in the classic short-term analysis methods. The proposed acoustic measures characterize power, aperiodicity, and similarity properties of vocal fry signals. The basic idea of the proposed method is to scan for local power peaks in a ldquovery short-termrdquo power contour for obtaining glottal pulse candidates, check for periodicity properties, and evaluate a similarity measure between neighboring glottal pulse candidates for deciding the possibility of being vocal fry pulses. In the periodicity analysis, autocorrelation peak properties are taken into account for avoiding misdetection of periodicity in vocal fry segments. Evaluation of the proposed acoustic measures in the automatic detection resulted in 74% correct detection, with an insertion error rate of 13%.

Clinical analysis of presbylarynx—Vocal fold atrophy in elderly individuals

OBJECTIVE The proportion of elderly Japanese is increasing significantly and aging society greatly influences the medical services. The number of elderly patients who visited out Voice Outpatient Clinic has been increasing in recent years. The aim of the present study was to analyze the trend in the number of patients with aging-related vocal fold atrophy, and to assess the characteristics of senile change of vocal function. METHODS A retrospective review of the 1999-2005 medical records of patients aged > or =65 years was conducted, and patients with presbylarynx were analyzed. We evaluated age, sex, occupation, maximal phonation time (MPT), and mean airflow rate (MFR) during comfortable phonation to assess the physiological characteristics of senile change of vocal function. We compared MFR of presbylarynx with those of normal elderly person. RESULTS The number of patients aged > or =65 who visited our clinic in the 7-year study period was 361, and 72 (20%) were diagnosed with vocal fold atrophy. They included 47 (65%) men and 25 (35%) women, with a mean age of 71 years (range, 65-89). Among these patients, 27 (38%) were still employed at the time of consultation, while 17 (24%) were retired, at least 26 jobs required voice skills; and 15 patients (21%) were involved in singing and recitation of poems. MPT correlated negatively with age, while MFR correlated positively with age. CONCLUSION The number of elderly patient with vocal fold atrophy increased gradually almost every year. The male patients account for 65%. Thirty-three percent of patient of the present study with presbylarynx were still in the workforce. MFR correlated positively and significantly with age. MFR was higher in presbylarynx than normal elderly person.

Evaluation of Vocal Fold Vibration With an Assessment Form for High-Speed Digital Imaging: Comparative Study between Healthy Young and Elderly Subjects

OBJECTIVES/STUDY DESIGN We conducted a prospective study with a subjective assessment form for high-speed digital imaging (HSDI) to elucidate the features of vocal fold vibrations in vocally healthy subjects and to clarify gender- and age-related differences. METHODS Healthy adult volunteers participated in this study. They were divided into young (aged 35 and younger) and elderly (aged 65 and older) groups, and the scores of an assessment form for HSDI characteristics elaborated at our institution were statistically analyzed. RESULTS Twenty-six young subjects (males: 9, females: 17; mean age: 27 years) and 20 elderly subjects (males: 8, females: 12; mean age: 72 years) were assigned to our study. Posterior gap and posterior-to-anterior longitudinal phase difference were characteristic to young females, whereas in young males, mucosal wave, anterior-to-posterior longitudinal phase difference, and supraglottic hyperactivity were frequent. In elderly males, axis shift, asymmetry, supraglottic hyperactivity, increased mucosal wave, lateral phase difference, and anterior-to-posterior longitudinal phase difference were frequent; and in elderly females, high incidence of lateral phase difference, atrophic change, anterior gap, and asymmetry were observed. CONCLUSIONS The results show that the behaviors of vocal fold vibrations were diverse even in healthy subjects with no vocal complaints or history of laryngeal diseases, and hence, the diversity of vocal fold vibrations in normal subjects must be taken into account in evaluating vocal fold vibrations.

Arytenoid Adduction for Correcting Vocal Fold Asymmetry: High-Speed Imaging

Objectives We hypothesized that high-speed digital imaging provides a quantitative method to evaluate the effect of arytenoid adduction for the correction of asymmetric and irregular vocal fold vibration in unilateral vocal fold paralysis. Methods Six subjects with unilateral vocal fold paralysis participated in the study (4 male, 2 female; mean [±SD] age, 52.5 ± 21.3 years). Videokymographic and laryngotopographic methods for image analysis were performed for highspeed recordings of vocal fold vibration for visualizing the glottal vibratory patterns, and for quantifying the frequency of vibration of each vocal fold, respectively. Comparisons of the paralyzed and the normal vocal folds were made before and after arytenoid adduction. Results Analysis of the laryngotopographs revealed 2 distinct frequencies of vibration for the paralyzed and the contralateral vocal folds for all subjects before surgery. After arytenoid adduction, the vibration frequencies became identical or nearly identical in all subjects. Conclusions Asymmetric vibration in vocal fold paralysis was exemplified by differences in vibration frequency between the vocal folds. The present data showed that after arytenoid adduction the vibration frequencies and the vibratory patterns of the contralateral vocal folds approached symmetry. This surgical procedure could improve the functional symmetry of the larynx for phonation.

Phase Difference of Vocally Healthy Subjects in High-Speed Digital Imaging Analyzed With Laryngotopography

BACKGROUND To elucidate the vibratory patterns of the vocal folds in normophonic subjects, we conducted a prospective study with laryngotopography, an analyzing technique for high-speed digital imaging to intuitively grasp spatial characteristics of vocal fold vibrations by pixel-wise discrete Fourier transform for brightness curve. MATERIALS AND METHODS Vocally healthy volunteer subjects with neither vocal complaint nor history of laryngeal diseases were divided into young (aged 35 years and younger) and elderly groups (aged 65 years and older), and high-speed digital images of phonations at a conversational frequency were analyzed with laryngotopography: phase differences (PDs) in the anterior-posterior (longitudinal) direction and in the left-right (lateral) direction were assessed. RESULTS Twenty-six young subjects (nine men and 17 women with a mean age of 27 years) and 20 elderly subjects (eight men and 12 women with a mean age of 73 years) were assigned to our study. In longitudinal PD, posterior-to-anterior opening type was significantly frequent in young women (all subjects: 43% and young women: 94%; P<0.001), whereas anterior-to-posterior opening type was significantly frequent in elderly men (all subjects: 39% and elderly men: 88%; P<0.001). There were no age- or gender-related differences in lateral PD (all subjects: 65%; P=0.880). DISCUSSIONS Both longitudinal and lateral PDs were widely observed even in vocally healthy subjects. Furthermore, the age- and gender-specific characteristics were also classified. Laryngotopography serves as a powerful tool to quantitatively characterize the vibratory patterns of vocal folds in vocally healthy subjects.

Characteristics of Vocal Fold Vibrations in Vocally Healthy Subjects: Analysis With Multi-Line Kymography

PURPOSE In this study, the authors aimed to analyze longitudinal data from high-speed digital images in normative subjects using multi-line kymography. METHOD Vocally healthy subjects were divided into young (9 men and 17 women; Mage = 27 years) and older groups (8 men and 12 women; Mage = 73 years). From high-speed digital images of phonation at a conversational frequency kymograms were created at 5 different levels of the vocal fold and were analyzed to determine the opening/closing longitudinal phase difference, open quotient, and speed index. Then age- and gender-related differences of these parameters were analyzed statistically. RESULTS Young women frequently showed a pattern of posterior-to-anterior glottal opening and anterior-to-posterior glottal closure, and older women demonstrated various opening and closing patterns. Both young men and older men were similar to older women. The open quotient was maximal at the most posterior glottal level in young women, but it tended to be maximal at the anterior glottis in the other subgroups. The mean value of the 5 open quotients was largest in young women. The mean speed index had a large negative value in older subjects. CONCLUSION This study provides the first information about age-related differences of longitudinal oscillatory characteristics of the vocal folds obtained with high-speed digital imaging.

Collagen injection for correcting vocal fold asymmetry: high-speed imaging.

Objectives We hypothesized that high-speed digital imaging with videokymographic and laryngotopographic analysis would provide a quantitative method to evaluate the effect of collagen injection for the correction of asymmetric and irregular vocal fold vibration in unilateral vocal fold paralysis. Methods Videokymographic and laryngotopographic analysis was performed for high-speed digital recordings of vocal fold vibration for visualizing the glottal vibratory patterns, and for quantifying the frequency of vibration of each vocal fold, respectively, including comparisons between the paralyzed and normal vocal folds before and after surgery. This included prospective observations of 11 subjects with unilateral vocal fold paralysis (4 male, 7 female; mean ± SD age, 67.1 ± 12.0 years) using high-speed digital image analysis before and after collagen injection. Results Analysis of the laryngotopographs revealed 2 distinct frequencies of vibration for the paralyzed and contralateral vocal folds for 8 of the 11 subjects before surgery. After collagen injection, the vibration frequencies became identical, despite asymmetric vibration amplitudes. Asymmetric vibration amplitudes were also observed in the other 3 subjects before surgery, but the amplitudes became symmetric after collagen injection, despite a persistent phase shift. Conclusions Asymmetric vibration in vocal fold paralysis was exemplified by differences in vibration frequency and amplitude between the vocal folds. The present study showed that after collagen injection, these aspects of vibratory patterns improved toward symmetry. This surgical procedure could improve the functional symmetry of the larynx for phonation.

Estimation of inferior-superior vocal fold kinematics from high-speed stereo endoscopic data in vivo

Despite being an indispensable tool for both researchers and clinicians, traditional endoscopic imaging of the human vocal folds is limited in that it cannot capture their inferior-superior motion. A three-dimensional reconstruction technique using high-speed video imaging of the vocal folds in stereo is explored in an effort to estimate the inferior-superior motion of the medial-most edge of the vocal folds under normal muscle activation in vivo. Traditional stereo-matching algorithms from the field of computer vision are considered and modified to suit the specific challenges of the in vivo application. Inferior-superior motion of the medial vocal fold surface of three healthy speakers is reconstructed over one glottal cycle. The inferior-superior amplitude of the mucosal wave is found to be approximately 13 mm for normal modal voice, reducing to approximately 3 mm for strained falsetto voice, with uncertainty estimated at σ ≈ 2 mm and σ ≈ 1 mm, respectively. Sources of error, and their relative effects on the estimation of the inferior-superior motion, are considered and recommendations are made to improve the technique.

Vocal Fold Vibration in Vocal Fold Atrophy: Quantitative Analysis With High-Speed Digital Imaging.

INTRODUCTION Vocal fold vibrations of vocal fold atrophy (VFA), a rapidly increasing voice disorder owing to worldwide societal aging, have not been clarified by high-speed digital imaging (HSDI). METHODS The HSDI method was performed on 46 patients (33 males and 13 females) with VFA and 20 vocally healthy subjects (8 males and 12 females), and the obtained data were quantitatively evaluated by frame-by-frame analysis, laryngotopography, single- and multi-line kymography, and glottal area waveform. RESULTS Overall, patients with VFA revealed larger open quotients, larger lateral phase difference, larger integral glottal width (the average glottal width over a glottal cycle), and smaller speed index than vocally healthy subjects. Some gender difference was noted: in males, lateral phase difference was not significant; and in females, integral glottal width and speed index were not significant. Correlation study revealed moderate correlations between HSDI-derived parameters and conventional acoustic or aerodynamic parameters. CONCLUSIONS The combination of multiple HSDI analysis methods was effective in documenting the characteristics of vocal fold vibrations in VFA. The knowledge of general vibratory characteristics and gender difference is beneficial for the appropriate clinical care of VFA.

Quantification of Vocal Fold Vibration in Various Laryngeal Disorders Using High-Speed Digital Imaging.

OBJECTIVE To quantify vibratory characteristics of various laryngeal disorders seen by high-speed digital imaging (HSDI). METHODS HSDI was performed on 78 patients with various laryngeal disorders (20 with polyp, 16 with carcinoma, 13 with leukoplakia, 6 with vocal fold nodule, and 33 with others) and 29 vocally healthy subjects. Obtained data were quantitatively evaluated by frame-by-frame analysis, laryngotopography, digital kymography, and glottal area waveform. RESULTS Overall, patients with laryngeal pathologies showed greater asymmetry in amplitude, mucosal wave and phase, smaller mucosal wave, and poorer glottal closure than vocally healthy subjects. Furthermore, disease-specific vibratory disturbances that generally agreed with the findings in the literature were quantified: comparing polyp with nodule, differences were noted in longitudinal phase difference, amplitude, and mucosal wave. In comparison with leukoplakia and cancer, nonvibrating area was more frequently noted in cancer. CONCLUSIONS The HSDI analysis of various voice disorders using multiple methods can help phonosurgeons to properly diagnose various laryngeal pathologies and to estimate the degree of their vocal disturbances.