Implementasi Speech Recognition Pada Aplikasi Iqra’ Berbasis Android

Masnur, T. Waldi Firmansyah (2020) Implementasi Speech Recognition Pada Aplikasi Iqra’ Berbasis Android. Other thesis, Universitas Islam Riau.

Text T. Waldi Firmansyah.pdf - Submitted Version Download (1MB)

Abstract

Reading or reading the Qur'an is an obligation for every Muslim. However, over the times, the interest of individuals in religious knowledge, especially in reading iqra 'decreases, so that new innovations are needed on how to create an interactive iqra learning media using smartphones. The purpose of this study is to implement speech recognition with the Pocketsphinx framework as a learning media for iqra on Android Studio. The method used for sound processing in pocketsphinx is Mel Frequency Cepstral Coeficient and Hidden Markov Model. The method of collecting speech data in this study was to record 6 speakers directly by using an audio recorder using a smartphone in a quiet room. For vocab data by taking a sample of 73 words hijaiyah in iqra '2 and 3. The results of this study are how the system can determine whether the words spoken by users are true or false. In application testing found several factors that can affect accuracy, namely the level of complexity of the words, the distance of the sound source to the microphone and the sound source of the examiner. So that the accuracy of the system is better it is recommended that the number of trained speakers be added, the use of recording equipment with good quality and recording in quiet room conditions.

Actions (login required)

View Item