Validation of an Android-Based Korean Voice Evaluation App Through Comparison with a Laboratory Acoustic Analysis System


Objective: Mobile-based voice assessment tools have become increasingly relevant as interest in telehealth and remote clinical screening grows. Despite this trend, most prior work has concentrated on English-language applications operating on iOS platforms, leaving limited options for Korean speakers and Android users. This study investigated whether the app can provide trustworthy acoustic measurements comparable to those obtained through a laboratory-standard system, thereby examining its feasibility as a remote voice-screening tool.

Methods: Sixty adults without reported voice problems participated. Each speaker produced a sustained /a/ vowel for 3 seconds and read the second sentence of a Korean reading passage (“Fall”). Speech was captured simultaneously through a smartphone and a Computerized Speech Lab (CSL; KayPENTAX) system. Smartphone recordings were processed using the new app and also analyzed with the Analysis of Dysphonia in Speech and Voice (ADSV) software. CSL recordings were evaluated using ADSV to serve as a benchmark. Acoustic features examined included cepstral and spectral measures: fundamental frequency (F0), Cepstral Peak Prominence (CPP), CPP standard deviation (CPPsd), Low-to-High spectral energy ratio (L/H ratio), and L/H ratio standard deviation. To evaluate measurement agreement across platforms and analysis methods, intraclass correlation coefficients (ICC) were calculated.

Results: A strong level of agreement was observed. The ICC among app-based smartphone analysis, smartphone-ADSV analysis, and CSL-ADSV analysis, with ICCs exceeding .70 for all acoustic parameters, indicating that the mobile tool yielded results comparable to the established laboratory system.

Conclusions: Results suggest that the Android-based Korean voice assessment app can reliably analyze voice recordings collected remotely. Individuals who face barriers to traditional clinical evaluation due to cost, travel restrictions, or limited service access may benefit from convenient self-monitoring and early vocal-health screening through this platform. Continued refinement and broader clinical testing will be important next steps in establishing its clinical utility and integration into telepractice environments.

Jaeock
Jeon Gue
Seung Jin
Yeon-Woo
Kim
Park
Lee
Lee