DOI 10.1016/j.jvoice.2026.04.009
Algorithm Verification and Concurrent Validity of a Web-Based Platform for Multiparametric Acoustic Voice Quality Indices.
Аннотация
OBJECTIVES
To evaluate whether established multiparametric acoustic voice quality indices, namely, the acoustic voice quality index (AVQI), acoustic breathiness index (ABI), cepstral spectral index of dysphonia (CSID), and smoothed cepstral peak prominence (CPPS), retain their concurrent validity with expert perceptual ratings when deployed through a free, web-based analysis platform.
STUDY
DESIGN
Cross-sectional validation study using a publicly available voice database.
METHODS
Voice samples from 290 speakers in the Perceptual Voice Qualities Database were analyzed using PhonaLab, a free web-based platform that implements these indices via Parselmouth (a Python interface to Praat). Sustained vowels and connected speech (consensus auditory-perceptual evaluation of voice [CAPE-V] sentences) were processed to extract AVQI, ABI, CSID, CPPS, and related parameters. Algorithm agreement with desktop Praat was assessed using Pearson correlations and intraclass correlation coefficients (ICCs). Spearman correlations with CAPE-V and GRBAS perceptual ratings were calculated. Receiver operating characteristic curve analysis was performed as a secondary, exploratory analysis.
RESULTS
Algorithm agreement between PhonaLab and desktop Praat was strong to excellent (r ≥ 0.96, ICC ≥ 0.94 for all primary indices). Correlations between acoustic indices and perceptual ratings were consistent with published validation studies: AVQI with CAPE-V Overall Severity (rs = 0.73) and GRBAS (Grade, Roughness, Breathiness, Asthenia, Strain) Grade (rs = 0.75); ABI with CAPE-V Breathiness (rs = 0.75); CSID with GRBAS Grade (rs = 0.68); CPPS from sustained vowels with CAPE-V Breathiness (rs = - 0.69). DeLong tests showed that ABI significantly outperformed single parameters for breathiness detection (P < 0.001).
CONCLUSIONS
Validated multiparametric acoustic indices retain their established concurrent validity with perceptual ratings when deployed through PhonaLab's web-based environment, supporting the feasibility of delivering research-validated acoustic analysis through browser-based platforms.
Подготовка и перевод публикации
Обработка статьи · этапы, ошибки и управление
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Тип публикации указан по данным источника. Уровень доказательности автоматически не присваивается.