Aplikasi Pembelajaran Hiragana Bahasa Jepang Berbasis Android Menggunakan Speech Recognition

Abstract

The alternative to learning Japanese is to use mobile learning. Based on comparisons made on the 7 popular Japanese language learning applications in the PlayStore, there is no Japanese language learning application that has a speech recognition feature for Japanese sentence pronunciation. In this study Japanese hiragana learning application was designed for sentences using speech recognition. This application uses google cloud speech-to-text for speech recognition features. The algorithm that is applied in this application is the levenshtein distance to compare the user's voice input with the text in the database. Android based application for users and the web for admin. Based on the results of testing the accuracy and success of detection, the ideal distance of the application receiving the user's voice input is 50 cm in a noise-free environment with a successful detection of 100%. In testing before and after using the application there was an increase in vocabulary knowledge of 44.48% after users used the application. Based on usability testing using SUS, the result is 73.01 which indicates that the application falls into the acceptable category.

Keywords

• japanese
• hiragana
• levenshtein distance
• speech recognition