A Study of Audio-to-Text Conversion Software Using Whispers Model
Amma Liesvarastranta Haz, Evianita Dewi Fajrianti, Nobuo Funabiki, Sritrusta Sukaridhoto
Abstract
This paper explores the potential of utilizing the Whispers model to create unified interfaces for audio-to-text in the context of Natural Language Processing (NLP). It offers possibilities for accurately converting spoken language into written texts. Whispers model by OpenAI is a state-of-the-art model in the field of NLP and is employe. In this study, various metrics and criteria are considered to evaluate the performance of the developed audio-to-text conversion software, including loading time, stress test, and transcription accuracy and speed. The proposal is capable of handling up to 180 concurrent users with an average response time of 309 ms and 471.5 requests per second. The results and findings of this study provide valuable insights into the effectiveness and limitations of the Whispers model in the context of audio-to-text conversion.