Investigating the cepstral acoustic characteristics of voice in healthy children

Cepstral Peak Prominence Smoothed - CPPS and Acoustic Voice Quality Index - AVQI in healthy and altered children's voices: comparation, relationship with auditory-perceptual judgment and cut-off points

PURPOSE

To compare the acoustic measurements of Cepstral Peak Prominence Smoothed (CPPS) and Acoustic Voice Quality Index (AVQI) of children with normal and altered voices, to relationship with auditory-perceptual judgment (APJ) and to establish cut-off points.

METHODS

Vocal recordings of the sustained vowel and number counting tasks of 185 children were selected from a database and submitted to acoustic analysis with extraction of CPPS and AVQI measurements, and to APJ. The APJ was performed individually for each task, classified as normal or altered, and for the tasks together defining whether the child would pass or fail in a situation of vocal screening.

RESULTS

Children with altered APJ and who failed the screening had lower CPPS values and higher AVQI values, than those with normal APJ and who passed the screening. The APJ of the sustained vowel task was related to CPPS and AVQI, and APJ of the number counting task was related only to AVQI and CPPS numbers. The cut-off points that differentiate children with and without vocal deviation are 14.07 for the vowel CPPS, 7.62 for the CPPS numbers and 2.01 for the AVQI.

CONCLUSION

Children with altered voices, have higher AVQI values and lower CPPS values, when detected in children with voices within the normal range. The acoustic measurements were related to the auditory perceptual judgment of vocal quality in the sustained vowel task, however, the number counting task was related only to the AVQI and CPPS. The cut-off points that differentiate children with and without vocal deviation are 14.07 for the CPPS vowel, 7.62 for the CPPS numbers and 2.01 for the AVQI. The three measures were similar in identifying voices without deviation and dysphonic voices.

Longitudinal Evaluation of Cepstral Peak Prominence in Children

OBJECTIVES

To evaluate whether the acoustic measure of cepstral peak prominence changes during typical development in children 2-7.

METHODS

Data were retrospectively analyzed from the Arizona Child Acoustic Database Repository in this longitudinal cohort study. The Repository contains longitudinal data recordings from 63 total children between 2 and 7 years of age. Thirty-one children met the inclusion criteria for the current analysis (at least five time points of usable speech data, no history of speech or language difficulties, no significant dysphonia, and were monolingual speakers of American English). Cepstral peak prominence measures were calculated in Praat for each child, at each timepoint. Additional acoustic measures of vocal fundamental frequency, vocal intensity, and stimuli length were also calculated. These measures were chosen as previous work has shown they may impact cepstral peak prominence values.

RESULTS

Linear mixed-effects regression models examined the relationship between cepstral peak prominence and age, after controlling for vocal fundamental frequency, vocal intensity, and stimuli length. Within-participant effects of age were found, indicating a trajectory change in which cepstral peak prominence increases with age in this population. This positive relationship between a cepstral peak prominence and age was nonlinear, with a steeper slope between age and cepstral peak prominence after 5 years of age.

CONCLUSIONS

This is the first study to examine the typical developmental trajectory of cepstral peak prominence children between 2 and 7 years, a critical period of vocal development. Cepstral peak prominence increased with age, suggesting an increase in periodicity of vocal fold vibration that coincides with the significant vocal fold structural changes occurring during this time. Outcomes present important normative information on vocal development, essential for effectively understanding the difference between what vocal changes are part of normative development and what changes indicate a voice disorder.

Cepstral Analysis of Voice in School-Aged Children

INTRODUCTION

Dysphonia in school-aged children is attributed primarily to hyperfunctional use of voice. These can be identified through effective protocols using both acoustic and auditory-perceptual analyses.

OBJECTIVE

The current study aimed to investigate voice characteristics in school children aged 4-17 years using auditory-perceptual rating and cepstral measures of voice.

STUDY DESIGN

This is a descriptive cross-sectional observational study.

METHOD

Four hundred and fifty-seven recordings of sustained phonation of /a/ in children and adolescents obtained in a quiet room using Zoom h1 voice recorder were analyzed using auditory-perceptual evaluation by three speech-language pathologists using Grade of overall dysphonia, Roughness, Breathiness, Asthenia, and Strain(GRBAS) rating scale. The samples were classified based on age into five groups: 1) 4 to 6; 11 years 2) 7-8; 11 years, 3) 9-11; 11years 4) 12-13; 11years and 5) 14-16; 11 years. PRAAT software was used to extract Cepstral Peak Prominence (CPP) and Cepstral Peak Prominence Smoothed (CPPS). Inter-rater reliability was assessed for both auditory-perceptual and acoustic analysis.

RESULTS

Auditory-perceptual analysis revealed dysphonia in 7.8% of samples with higher rate in males than females. Inter-rater reliability for auditory-perceptual rating was found to be good (Intraclass Corelation Coefficient-0.83). Independent t test revealed statistically significant difference (P < 0.001) in both cepstral measures and mean values were lower in dysphonic than normal group. Gender effect was present for CPP in group 5(14-16;11 years) and CPPS in group 4 (12-13; 11 years). One-way analysis of variance within groups in males (P < 0.005) revealed statistical difference in both cepstral measures but not in females. Statistically significant difference was not found between ratings of both speech language pathologists for both CPP (P = 0.929) and CPPS (P = 0.965) values indicating the ratings to be reliable.

CONCLUSION

Pediatric dysphonia has received less attention when compared to adults. Assessing school-aged children for dysphonia using both auditory-perceptual and acoustic measures would aid in identifying those at risk to make appropriate referrals and plan further intervention.

Attention guided learnable time-domain filterbanks for speech depression detection

Depression, as a global mental health problem, is lacking effective screening methods that can help with early detection and treatment. This paper aims to facilitate the large-scale screening of depression by focusing on the speech depression detection (SDD) task. Currently, direct modeling on the raw signal yields a large number of parameters, and the existing deep learning-based SDD models mainly use the fixed Mel-scale spectral features as input. However, these features are not designed for depression detection, and the manual settings limit the exploration of fine-grained feature representations. In this paper, we learn the effective representations of the raw signals from an interpretable perspective. Specifically, we present a joint learning framework with attention-guided learnable time-domain filterbanks for depression classification (DALF), which collaborates with the depression filterbanks features learning (DFBL) module and multi-scale spectral attention learning (MSSA) module. DFBL is capable of producing biologically meaningful acoustic features by employing learnable time-domain filters, and MSSA is used to guide the learnable filters to better retain the useful frequency sub-bands. We collect a new dataset, the Neutral Reading-based Audio Corpus (NRAC), to facilitate the research in depression analysis, and we evaluate the performance of DALF on the NRAC and the public DAIC-woz datasets. The experimental results demonstrate that our method outperforms the state-of-the-art SDD methods with an F1 of 78.4% on the DAIC-woz dataset. In particular, DALF achieves F1 scores of 87.3% and 81.7% on two parts of the NRAC dataset. By analyzing the filter coefficients, we find that the most important frequency range identified by our method is 600-700Hz, which corresponds to the Mandarin vowels /e/ and /eˆ/ and can be considered as an effective biomarker for the SDD task. Taken together, our DALF model provides a promising approach to depression detection.

Voice Onset Time in Children With and Without Vocal Fold Nodules

PURPOSE

Voice onset time (VOT) of voiceless consonants provides information on the coordination of the vocal and articulatory systems. This study examined whether vocal-articulatory coordination is affected by the presence of vocal fold nodules (VFNs) in children.

METHOD

The voices of children with VFNs (6-12 years) and age- and gender-matched vocally healthy controls were examined. VOT was calculated as the time between the voiceless stop consonant burst and the vocal onset of the vowel. Measures of the average VOT and VOT variability, defined as the coefficient of variation, were calculated. The acoustic measure of dysphonia, cepstral peak prominence (CPP), was also calculated. CPP provides information about the overall periodicity of the signal, with more dysphonic voices having lower CPP values.

RESULTS

There were no significant differences in either average VOT or VOT variability between the VFN and control groups. VOT variability and average VOT were both significantly predicted by the interaction between Group and CPP. There was a significant negative correlation between CPP and VOT variability in the VFN group, but no significant relationship was found in the control group.

CONCLUSIONS

Unlike previous studies with adults, there were no group differences in average VOT or VOT variability in this study. However, children with VFNs who were more dysphonic had increased VOT variability, suggestive of a relationship between dysphonia severity and control of vocal onset during speech production.

Voice Therapy Improves Acoustic and Auditory-Perceptual Outcomes in Children

PURPOSE

This study employed acoustic measures as well as auditory-perceptual assessments to examine the effects of voice therapy in children presenting with benign vocal fold lesions.

METHODS

A retrospective, observational cohort design was employed. Sustained vowels produced by 129 children diagnosed with benign vocal fold lesions were analyzed, as well as connected speech samples produced by 47 children. Treatment outcome measures included Consensus of Auditory-Perceptual Evaluation of Voice (CAPE-V), jitter, shimmer, Noise-to-Harmonic Ratio (NHR), cepstral peak prominence (CPP), and Low-to-High Ratio (LHR) on sustained vowels, and CPP and LHR on connected speech.

RESULTS

Following voice therapy, significant improvements in CAPE-V ratings (p < 0.001) were observed. Additionally, jitter (p = 0.041), NHR (p = 0.019), and CPP (p < 0.01) on sustained vowels, and CPP (p = 0.002), and LHR (p = 0.008) on connected speech significantly improved following voice therapy. CPP increased with age in males but did not change in females. CAPE-V ratings and perturbation measures indicated that dysphonia was more severe in younger children pre and post-therapy.

CONCLUSIONS

Auditory-perceptual and acoustic measures demonstrated improved voice quality following voice therapy in children with dysphonia. CPP effectively quantified voice therapy gains and allowed for analysis of connected speech, in addition to sustained vowels. These findings demonstrate the value of CPP as a tool in assessing therapy outcomes and support the efficacy of voice therapy for children presenting with vocal fold lesions.

LEVEL OF EVIDENCE

4 Laryngoscope, 133:977-983, 2023.

Update on the pediatric adverse vocal behavior voice disorders: a clinical practice review

The main causes of voice disorders in children with adverse vocal behavior include benign lesions of the vocal folds caused by voice abuse or misuses, such as vocal fold nodules, vocal fold polyps, and laryngitis. Long-term voice disorders can affect the physical and mental health of children. Reviewing the literature of the last two decades on "Vocal Fold Nodules," "Vocal Fold Polyp," "Voice disorder," "Voice Abuse," "Voice Misuse," "Pediatrics," and "Children" with the appropriate Boolean operators.  Conclusion: A total of 315 results were returned on an initial PubMed search. All articles from 2000 to 2022 written in English or Chinese were screened. Duplicate articles, those relating to adults only or concerned with the malignant lesion of the vocal cord, were excluded, resulting in 196 articles of interest. Relevant references and books have also been consulted, and we provide a review of the pathogenesis, diagnosis, and treatment of these maladaptive vocal behavioral voice disorders. What is Known: • Hoarseness is the most common voice symptom in children, and there are various causes of hoarseness in children. However, there is a lack of reviews on voice disorders caused by adverse vocal habits in children. • Voice training is a conservative treatment method for children with voice disorders , and it is important to clarify the factors that influence the effectiveness of voice training for children. What is New: • This review of the personality and family characteristics of children with adverse vocal behavioural voice disorders provides a valuable guide to the clinical planning of subsequent treatment. • This article discusses and summarises some of the factors that may influence the effectiveness of voice training in children and collates some of the scales and questionnaires currently used in children that are important in predicting the effectiveness of voice training.

A Practical Guide to Calculating Cepstral Peak Prominence in Praat

The acoustic measure of cepstral peak prominence (CPP) is recommended for the analysis of dysphonia. Yet, clinical use of this measure is not universal, as clinicians and researchers are still learning the strengths and limitations of this measure. Furthermore, affordable access to specialized acoustic software is a significant barrier to universal CPP use. This article will provide a guide on how to calculate CPP in Praat, a free software program, using a new CPP plugin. Important external factors that could influence CPP measures are discussed, and suggestions for clinical use are provided. As CPP becomes more widely used by clinicians and researchers, it is important to consider external factors that may inadvertently influence CPP values. Controlling for these external factors will aid in reducing variability across CPP values, which will make CPP a valuable tool for both clinical and research purposes.

[Detection of speech pathology based on parameters of analysis of dysphonia in speech and voice]

Objective:To analysis speech pathology based on dysphonia in speech and voice（ADSV）. Methods:The acoustic signals of continuous vowels and continuous speech of one-hundred and thirteen individuals were collected, including 93 vocal cord polyps cases, 20 glottis laryngeal carcinoma cases and 47 volunteers without speech sound disorders. Cepstral peak prominence（CPP）, CPP standard deviation（CPP SD）, L/H spectral ratio（L/H ratio）, L/H ratio standard deviation（L/H ratio SD） and cepstral/spectral index of dysphonia（CSID） were analyzed by ADSV to explore the role of these parameters in the recognition of speech pathology. Results:In the acoustic signal of continuous vowels, CPP and L/H ratio in normal group were higher than those in pathological voice group（P<0.001）, while CPP SD and CSID were lower than those in pathological voice group（P<0.001）, CPP and CSID areas under ROC curve were 0.95 and 0.99, respectively, which were important acoustic parameters for diagnosing pathological voice. In continuous speech acoustic signals, CPP, CPP SD and L/H ratio in the normal group were all higher than those in the speech disorders group（P<0.001）, and the area under the curve of CPP SD was 0.90, which showed high accuracy in diagnosing pathological voice. The ADSV voice analysis parameters CPP, CPP SD, CSID, and L/H ratio also showed significant differences between the vocal cord polyp group and the glottic laryngeal cancer group. The results of the discriminant analysis model show that the use of ADSV voice parameters can distinguish vocal cord polyps and laryngeal cancers. Conclusion:The ADSV voice analysis parameters can not only distinguish the voice signals of the normal group and the pathological group, but also distinguish different types of pathological voices. It has high sensitivity and specificity in diagnosing pathological voices.

Validity and Reliability of the Dutch Children's Voice Handicap Index-10

OBJECTIVES

Voice-related quality of life (Qol) questionnaires provide the clinician with information regarding the impact of voice disorders on the patient's well-being. The available voice-related QoL tools for Dutch-speaking children are parent-proxy in nature. However, the use of proxy measurements has been debated in the literature. The Children's Voice Handicap Index-10 (CVHI-10) is a self-reported QoL tool for dysphonic children. Therefore, the aim of this study is to develop and validate a Dutch version of the CVHI-10.

STUDY DESIGN

Observational, prospective, cross-sectional study.

METHODS

The original version of the CVHI-10 was translated and adapted to Dutch according to the recommendations of the Quality of Life Special Interest Group - Translation and Cultural Adaptation group. Subsequently, the questionnaire was individually completed by 77 children (dysphonic group: n = 30, control group: n = 47) between eight and 14 years. In order to investigate test-retest reliability, 50% of the participants were asked to complete the questionnaire twice with an interval of 2 weeks. Internal consistency, test-retest reliability and construct validity were calculated. A receiver operating characteristic (ROC) analysis was conducted to check the sensitivity and specificity levels of the instrument.

RESULTS

Internal consistency measured with Cronbach's alpha coefficient was 0.745. Test-retest reliability measured with intraclass correlation coefficients was 0.718. Mean total CVHI-10 score was 6.17 ± 2.7 in the dysphonic group and 2.68 ± 2.6 in the control group. The difference in total score between the groups was statistically significant (P < 0.001), suggesting that the tool has good construct validity. ROC analysis demonstrated moderate diagnostic accuracy (area under the curve = 0.869) and suggested a cut-off score of 3.5.

CONCLUSIONS

The Dutch CVHI-10 is the first self-reported voice-related QoL tool for dysphonic Dutch-speaking children. It is a valid, reliable and sensitive tool to assess the impact of a voice disorder on the child's well-being.

Smoothed Cepstral Peak Analysis of Brazilian Children and Adolescents Speakers

INTRODUCTION

Childhood and adolescence are essential stages in the development of voice and speech quality; therefore, it is essential to understand the vocal changes that occur during this period. Frequency-based measurement methods like cepstral measurements stand out among the methods described, which are able to identify fo and estimate the periodicity and noise in the acoustic wave without establishing individual cycles of the sound wave.

METHODS

Two hundred seventy-one recordings (128 female and 131 male) from children and adolescents aged 5 to 18 years with no vocal complaints were analyzed. Three speech-language pathologists assessed the vocal quality and determined as appropriate for the age. The recordings were divided into six age groups (G1:5-7; G2:8-9; G3:10-11; G4:12; G5:13-15 and G6:16-18 years old). Sustained production of the vowel /a/ were inspected and edited using the PRAAT software. Then, it was extract de Cepstrum Peak Prominence Smoothed (CPPS) using a script in the same software. A Two-way ANOVA was applied to investigate the effect of sex, age and sex*age interaction, followed by Bonferroni's correction for each gender separately. Finally, the Student's t test for independent samples was performed to compare genders within each age group.

RESULTS

Male children and adolescents from G5 and G6 had higher CPPS measures than G1, G2 and G3 (P ≤ 0.001). In addition, G6 also had higher values than G4 (P ≤ 0.001). There was no difference between age groups in the female group. In turn, sex differences were reported from 12 years of age onwards, with higher CPPS values found for male participants compared to female participants (P ≤ 0.01).

CONCLUSION

Vocal changes that usually occur from childhood to adolescence are reflected in the acoustic CPPS measure in males, resulting in higher values in the 13 to 18 years old. On the other hand, no changes in CPPS values were observed in the age groups of female participants. Males have higher CPPS values than females and that sex differences are reported after 12 years of age.