Implementation of a COVAREP-Integrated Application for Voice Analysis Using LLMs

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

This paper presents VocalCheck, a mobile application designed to monitor voice conditions using recordings made with a smartphone. The initial phase involved a literature review and an evaluation of existing voice processing applications, which informed the project’s design requirements. Based on this analysis, an application was developed to record and preliminarily analyze voice samples using parameters provided by the Collaborative Voice Analysis Repository (COVAREP) for the Speech Technologies Toolkit. Additionally, an experimental study was conducted to assess the effectiveness of large language models (LLMs), specifically Microsoft Copilot and ChatGPT, in generating MATLAB scripts for computing voice quality parameters. Although these tools proved helpful for general programming tasks, their performance in speech signal processing was unsatisfactory.