|Year : 2015 | Volume
| Issue : 2 | Page : 34-38
Cepstral analysis of voice in healthy aged individuals
Dhanya Monnappa, Radish Kumar Balasubramanium
Department of Audiology and Speech-Language Pathology, Kasturba Medical College (Manipal University), Mangalore, Karnataka, India
|Date of Web Publication||13-Jun-2016|
Dr. Radish Kumar Balasubramanium
Department of Audiology and Speech-Language Pathology, Kasturba Medical College (Manipal University), Mangalore, Karnataka
Source of Support: None, Conflict of Interest: None
| Abstract|| |
Introduction: The anatomical and physiological changes in the phonatory system due to aging have a major impact on voice. Widely researched parameters of voice such as fundamental frequency and the perturbation measures depend on the location of the exact pitch pulses, and it may yield unreliable results in case of a severely aperiodic voice. Aim of the Study: The present study focused on determining the cepstral characteristics of voice in healthy aging individuals. Methods: All the participants among the young adults, middle-aged adults, and old adults were required to sustain the vowel/a/. Voice samples were analyzed using Hillenbrand's Z tool software to obtain the cepstral measures. Results: Results showed that there was an increase in the cepstral measures with aging, thus suggesting that the harmonic structure of voice is not affected in healthy aged individuals. Gender also influences the cepstral measures in healthy aging population. Conclusion: The results of the present study could be used by the voice clinicians while assessing adults with voice disorders. It will also help the clinicians in delineating the age and gender differences in the vocal parameters.
Keywords: Cepstral measures, healthy aging, voice quality
|How to cite this article:|
Monnappa D, Balasubramanium RK. Cepstral analysis of voice in healthy aged individuals. J Laryngol Voice 2015;5:34-8
| Introduction|| |
The normal process of aging has a major impact on the voice. It changes the anatomical and physiological mechanism of the phonatory system. These changes include atrophy of the laryngeal muscles,  ossification and calcification of the cartilages of larynx,  thinning of vocal fold mucosa,  and reduction in the movement of the cricothyroid joint.  Most of the researchers have agreed that this age-related anatomical and physiological changes in voice will affect the acoustic properties of speech such as formant frequencies, fundamental frequency, and perturbation measures. , Fundamental frequency is one of those acoustic properties that is most widely studied in healthy aged individuals. In women, F0 remains considerably the same until menopause, after which the pitch decreases by 10-15 Hz. This decrease could be attributed to the hormonal changes leading to edema of the mucosa. In men, F0 rises considerably after middle age (approximately 35 Hz). ,
Another widely researched parameter is frequency and amplitude perturbation measures. These measures signify the cycle-to-cycle variations in the pitch period and amplitudes, respectively. These measures of periodicity are known to determine the vocal control and are the aspects that are most often being studied in aging voice.  It was observed that jitter and shimmer measures are increased in both males and female elderly individuals. Moreover, shimmer is found to be strongly associated with age.  However, these measures are reported to be less reliable because of their dependency on the exact location of the pitch pulses, which could lead to an uncertain result in case of a severely aperiodic voice. ,
Spectral-dependent measures such as the normalized noise energy, cepstral analysis, and the spectral amplitude measures are found to capture the harmonic organization of the spectrum and estimate the periodicity of voice based on the harmonics. Cepstrum is obtained by the Fourier transformation of the spectrum. ,,,,,, By subjecting the voice signal to fast Fourier transformation, the waveform which is a time domain signal is converted into a frequency domain signal, i.e., spectrum. Further, performing Fourier transformation to a spectrum, a que frequency signal cepstrum is obtained. The time ("quefrency") at the cepstral peak corresponds to the fundamental period of the signal. A signal whose spectrum shows a well-defined harmonic structure will show a very prominent cepstral peak. ,,,,,, One of the measures of cepstral analysis is the cepstral peak prominence (CPP), and it is the amplitude of the highest peak in the cepstrum. It is estimated by finding the difference in amplitude between the cepstral peak and the regression line below the peak. If each of the cepstra is first averaged for a given number of times and then cepstral peak is extracted, the smoothened CPP (CPPS) is obtained. These measures are less susceptible to the technical shortcomings that are faced by the perturbation measures. Moreover, these measures better depict the harmonic organization. ,,,,,, Hence, the present study is focused on characterizing the harmonic structure of the voice in the adults through cepstral analysis.
Age-related changes in voice have been extensively studied on fundamental frequency and perturbation measures of voice. However, because of poor reliability on these measures, it is difficult to estimate the exact changes that take place with aging. The evidence related to the harmonic structure of voice of elderly individuals is scanty. Thus, the present study is carried out to determine the cepstral characteristics of voice associated with aging. Furthermore, the results of the study will help the voice clinicians in distinguishing the aging voice from that of the disordered voice and helps in the better assessment and management of voice in elderly population. Thus, the aim of the study was to estimate the cepstral measurements of voice in healthy aged population.
| Methods|| |
One hundred and twenty participants were recruited for the study. They were divided into three groups based on the WHO age classification. Each group consisted of 40 adults (20 males and 20 females) in which Group 1 had young aged adults between the age range of 18 and 40 years, Group 2 had middle-aged adults between the age range of 40 and 60 years, and Group 3 had geriatrics between the age range of 61 and above. All the participants included in the study were adults having normal voice, and none of them had any voice problems, history of vocal abuse or misuse, gastroesophageal reflux disorder, smoking habits, exposure to toxic or chemicals fumes, and any respiratory problems.
Hillenbrand's CPPS Software, Speech Tool (Z tool) 1.65 based on the Hillenbrand's algorithm (Hillenbrand, Western Michigan University) was used for cepstral analysis.
For the recording of the voice sample, all the participants were seated on a chair in a quiet room. The voice samples were recorded using a dynamic, sensitive microphone, which was kept at a distance of 10 cm from the participant's mouth. All the samples were recorded with 22.1 kHz and 16-bit quantization. All the participants were instructed to phonate and sustain the vowel/a/at their habitual loudness and pitch for a minimum duration of 5 s. They were also instructed to avoid singing it. A demonstration was given on how to carry out the task. A short period of familiarization was given for the participants to understand the instructions and they were repeated if the participants did not follow.
The voice samples were also fed into the Z tool software for the cepstral analysis and the two cepstral measures such as CPP and the CPPS were obtained. The above-mentioned procedures were kept constant and were carried out for all the participants in the three groups [Figure 1].
All the statistical analyses were done using SPSS version 16.0 (SPSS Inc (Chicago, IL)). Mean and standard deviation were calculated for the cepstral measures to obtain normative data for all the groups. Further, cepstral measures were subjected to multivariate analysis of variance, followed by Bonferroni t-test to see a significant difference, if any, across the group and gender.
| Results|| |
The present study aimed at estimating the changes in cepstral measures (CPP and CPPS) of voice with healthy aging. The results are presented under the following sections. The descriptive statistics of the cepstral measures, CPP, and CPPS across the three groups are shown in [Table 1].
|Table 1: Descriptive statistics for cepstral measures across the groups and gender |
Click here to view
As shown in [Table 1], the mean CPP and CPPS values were the largest for Group 3, followed by Group 1 and Group 2, respectively. Among the gender, males had higher cepstral values when compared to females.
The results of multivariate analysis of variance have revealed that there was a significant difference (P < 0.05) seen across the groups. The CPP (t [df = 2] =19.93, P = 0.000) and CPPS (t [df = 2] =31.34, P = 0.000) scores showed a significant statistical difference among the three groups. A significant difference was also noted for gender in the CPPS scores (t [df = 1] =4.62, P = 0.034) across the three groups. The group and gender interaction showed a statistical significant difference for CPP (t [df = 2] =7.50, P = 0.001) and CPPS (t [df = 2] =13.02, P = 0.000) scores.
The results of the post hoc analysis for the cepstral measures, CPP, and CPPS for the vowel are shown in [Table 2].
|Table 2: Results of post hoc analysis of cepstral measures across the three groups |
Click here to view
[Table 2] indicates a significant difference for cepstral measures between Group 1 and Group 2 for both CPP and CPPS scores. The analysis also showed a significant difference for Group 3 when compared to Group 2 for both the cepstral measures and also when compared to Group 1 for CPP scores.
| Discussion|| |
The present study examined the cepstral characteristics of voice in healthy aged individuals in comparison to the young adults and middle-aged adults. The comparison of the cepstral measures across the three age groups has revealed a significant main effect, suggesting that the cepstral scores increase with age in both genders. The larger cepstral scores in older age groups compared to middle-aged and young-aged would imply that the voice quality gets better with advancing age. However, no standard perceptual voice analysis was performed to support the same. This finding could also be attributed to the fact that only normal individuals were recruited in the present study, and it was ensured that none of them had any voice problems, history of vocal abuse or misuse, gastroesophageal reflux symptoms, smoking habits, exposure to toxic or chemical fumes, and any respiratory problems.
The CPP is a measure of amplitude difference (in dB) between the cepstral peak and the corresponding regression line below the peak. This measure will provide information regarding the harmonic structure of a voice sample. ,,,,,, In the present study, CPP values increased with age, with older adults obtaining a higher mean scores (20.37 dB) followed by the middle-aged adults (17.54 dB) and the young-aged adults (18.4 dB), respectively. The greater value of CPP would imply that the voice samples of the older adults in the present study had a well-defined harmonic structure.
The CPPS is the peak prominence obtained from the cepstrum, which is averaged over a given number of frames. This measure is also a good predictor of dysphonia and gives information regarding the harmonic organization of a voice sample. ,,,,,, The results of the present study have shown that the CPPS measure also tends to increase with age. The older adults have obtained highest mean values (9.545 dB) followed by young-aged adults (9.09 dB) and middle-aged adults (7.06 dB), respectively. Both CPP and CPPS were better in predicting the age-related changes in voice in adults in the present study. The findings suggest that the older adults in the study had good quality of voice which could be attributed to the kind of projective voice they have developed over the years.
The comparison of the cepstral measures across the gender indicated a significant main effect for CPP measure. The females had smaller amplitudes for the cepstral measures across all the three groups, which would indicate that their harmonic structure is poor compared to that of the males. In most of the females, the glottal configuration is such that there is an opening at the posterior border throughout the glottal cycle.  This configuration is said to modify the glottal waveform such that there is a presence of a turbulent noise source at the region of the glottis, which has the amplitude (at high frequency) similar to that of the glottal source.  This, in turn, would decrease the harmonic organization of voice in females. Along with this, the hormone-related factors such as premenstrual dysphonia and menopause might have decreased the estrogen level in females.  This deficit will cause the laryngeal tissues and the blood vessels in the larynx to enlarge in turn affecting the quality of voice, thus lowering the cepstral values. However, these factors were not considered in the present study.
After the age of 65 years, the voice is said to decline in a much similar way like the other body functions. , However, the voice does not always reflect the changes that might occur in the physical functioning of the body. There is no particular characteristic that would define every voice that is aging. The anatomical or the physiological changes due to aging could vary from person to person. Few individuals will have a well-preserved voice even at their 80's whereas few others' voices begin to decline from their 50's. There are studies which have reported that the aged individuals having good physical condition have the quality of voice which is similar to that of the younger individuals.  There are also evidences that the singers could maintain their voices till their seventh decade. The voice quality degradation is said to depend on various factors such as the food habits, lifestyle, alcohol consumption, and smoking habits. , Genetics is also one of the important factors which could influence the age-related changes in voice. All these factors would have played a role in the voice of the older adults in the present study, thus yielding a normal voice quality.
| Summary and Conclusion|| |
The aim of the present study was to estimate the cepstral measures for healthy aging population. Results revealed that the older adults obtained higher cepstral values followed by middle-aged adults and young-aged adults, respectively. This finding was attributed to the fact that the participants included in the study were individuals having normal voices and thus there was an improvement in the harmonic structure of the voice that was noted with age. Gender differences were noted in cepstral measures. It showed that females had smaller cepstral measures indicating a poorer quality of voice when compared to males. The present study focused on obtaining the data for sustained phonation; however, the quality of the voice could be different during a speaking situation. Thus, this measure must be validated in the continuous speech samples too. During the voice recording, the loudness of the voice was not controlled; this would have led to the better values in the measured parameters. The age range of Group 3 participants was between 61 and 75 years. Thus, the age-related decline in the voice during the eighth and ninth decade of life is not accounted in the study.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Thomas LB, Harrison AL, Stemple JC. Aging thyroarytenoid and limb skeletal muscle: Lessons in contrast. J Voice 2007;8:56-9.
Hatley W, Samuel E, Evison G. The pattern of ossification in the laryngeal cartilages: A radiological study. Br J Radiol 1965;38:585-91.
Noell G. On the problem of age-related changes of the laryngeal mucosa. Arch Ohren Nasen Kehlkopfheilkd 1962;179:361-5.
Casiano RR, Ruiz PJ, Goldstein W. Histopathologic changes in the aging human cricoarytenoid joint. Laryngoscope 1994;104 (5 Pt 1):533-8.
Rastatter MP, Jacques RD. Formant frequency structure of the aging male and female vocal tract. Folia Phoniatr (Basel) 1990;42:312-9.
Sebastian S, Babu S, Oommen NE, Ballaraj A. Acoustic measurement of geriatric voice. J Laryngol Voice 2012;2:81-4.
Endres W, Bambach W, Flösser G. Voice spectrograms as a function of age, voice disguise, and voice imitation. J Acoust Soc Am 1971;49:1842-8.
Sovani P, Mukundan G. Comparison of postmenopausal voice changes across professional and non-professional users of the voice. S Afr J Commun Disord 2010;57:76-81.
Linville SE. Vocal Aging. Australia: Singular Thomson Learning; 2001.
Xue SA, Deliyski D. Effects of aging on selected acoustic voice parameters: Preliminary normative data and educational implications. Educ Gerontol 2001;27:159-68.
Hillenbrand J. A methodological study of perturbation and additive noise in synthetically generated voice signals. J Speech Hear Res 1987;30:448-61.
Rabinov CR, Kreiman J, Gerratt BR, Bielamowicz S. Comparing reliability of perceptual ratings of roughness and acoustic measure of jitter. J Speech Hear Res 1995;38:26-32.
Hillenbrand J, Houde RA. Acoustic correlates of breathy vocal quality: Dysphonic voices and continuous speech. J Speech Hear Res 1996;39:311-21.
Hillenbrand J, Cleveland RA, Erickson RL. Acoustic correlates of breathy vocal quality. J Speech Hear Res 1994;37:769-78.
Radish Kumar B, Bhat JS, Prasad N. Cepstral analysis of voice in persons with vocal nodules. J Voice 2010;24:651-3.
Balasubramanium RK, Bhat JS, Fahim S 3 rd
, Raju R 3 rd
. Cepstral analysis of voice in unilateral adductor vocal fold palsy. J Voice 2011;25:326-9.
Balasubramanium RK, Bhat JS, Srivastava M, Eldose A. Cepstral analysis of sexually appealing voice. J Voice 2012;26:412-5.
Balasubramanium R, Bhat JS. Cepstral characteristics of voice before, during and after menopause. Int J Phonosurg Larngol 2014;4:50-4.
Chandran S, Hanna J, Lurie D, Sataloff RT. Differences between flexible and rigid endoscopy in assessing the posterior glottic chink. J Voice 2011;25:591-5.
de Krom G. A cepstrum-based technique for determining a harmonics-to-noise ratio in speech signals. J Speech Hear Res 1993;36:254-66.
Kadakia S, Carlson D, Sataloff RT. The effect of hormones on the voice. J Sing 2013;69:571-4.
Kahane JC. Anatomic and physiologic changes in the aging peripheral speech mechanism. In: Beasley DS, Davis GA, editors. Aging: Communication Processes and Disorders. New York: Grune & Stratton; 1981.
Ramig LA, Ringel RL. Effects of physiological aging on selected acoustic characteristics of voice. J Speech Hear Res 1983;26:22-30.
Gorham-Rowan MM, Laures-Gore J. Acoustic-perceptual correlates of voice quality in elderly men and women. J Commun Disord 2006;39:171-84.
[Table 1], [Table 2]
|This article has been cited by|
||Acoustic Voice Quality Index (AVQI) in the Measurement of Voice Quality: A Systematic Review and Meta-Analysis
| ||Thirunavukkarasu Jayakumar, Jesnu Jose Benoy |
| ||Journal of Voice. 2022; |
|[Pubmed] | [DOI]|
||Effect of Gender on Acoustic Voice Quality Index 02.03 and Dysphonia Severity Index in Indian Normophonic Adults
| ||Srushti Shabnam,M. Pushpavathi |
| ||Indian Journal of Otolaryngology and Head & Neck Surgery. 2021; |
|[Pubmed] | [DOI]|
||Vocal Changes of Men and Women from Different Age Decades: An Analysis from 30 Years of Age
| ||Aline Oliveira Santos,Juliana Godoy,Kelly Silverio,Alcione Brasolotto |
| ||Journal of Voice. 2021; |
|[Pubmed] | [DOI]|
||A Cross-sectional Study of Perceptual and Acoustic Voice Characteristics in Healthy Aging
| ||Benjamin G. Schultz,Sandra Rojas,Miya St John,Elaina Kefalianos,Adam P. Vogel |
| ||Journal of Voice. 2021; |
|[Pubmed] | [DOI]|
||Effect of Age and Gender on Acoustic Voice Quality Index Across Lifespan: A Cross-sectional Study in Indian Population
| ||T. Jayakumar,Jesnu Jose Benoy,H. Mohamed Yasin |
| ||Journal of Voice. 2020; |
|[Pubmed] | [DOI]|
||Correlation Between Subjective and Objective Parameters of Voice in Elderly Male Speakers
| ||Mahesh Soumya,Srirangam Vijayakumar Narasimhan |
| ||Journal of Voice. 2020; |
|[Pubmed] | [DOI]|