PENGEMBANGAN SISTEM PENGENALAN SUARA SECARA REAL-TIME PADA KOMPUTER MENGGUNAKAN BAHASA PEMROGRAMAN C

IMAM MUCHLIZAR, Dr. Ir. Risanuri Hidayat, M.Sc.; Agus Bejo, S.T., M.Eng., D.Eng.

2019 | Skripsi | S1 TEKNIK ELEKTRO

Sebagian besar komputer masih dikontrol dengan menggunakan peralatan tertentu seperti mouse dan keyboard, padahal teknologi pengenalan suara sudah cukup canggih dan dapat diimplementasikan pada komputer sebagai control system. Paper ini akan membahas tentang pembuatan sistem pengenalan suara menggunakan bahasa pemrograman C yang mampu merekam, menampilkan, dan mengenali data hasil rekaman suara secara real-time. Library yang digunakan pada program adalah PortAudio yang merupakan Application Programming Interface (API) untuk merekam suara. Metode Mel Frequency Cepstral Coefficient (MFCC) digunakan untuk mengekstrak ciri dari data audio dan Artificial Neural Network (ANN) digunakan untuk membuat sistem data matching pada sistem pengenalan. Sistem pengenalan suara yang dirancang memiliki persentase akurasi kata "atas" 93%, kata "bawah" 98%, kata "kanan" 86%, kata "kiri" 93%, kata "maju" 78%, kata "mundur" 86%, kata "hidup" 98%, kata "mati" 78%, kata "cepat" 93% dan kata "lambat" 90%.

Most computers are still controlled using certain equipment such as a mouse and keyboard, even though speech recognition technology is quite sophisticated and can be implemented on a computer as a control system. This paper will discuss the making of a speech recognition system using C programming language that is capable of recording, displaying, and recognizing sound recording data in real-time. The library used in the program is PortAudio which is an Application Programming Interface (API) for sound recording. The Mel Frequency Cepstral Coefficient (MFCC) is used as a method to extract features from audio data and Artificial Neural Network (ANN) is used to create a data matching system in the recognition system. The voice recognition system that is designed has an accuracy percentage of the word "atas" 93%, the word "bawah" 98%, the word "kanan" 86%, the word "kiri" 93%, the word "maju" 78%, the word "mundur" 86%, the word "hidup" is 98%, the word "mati" is 78%, the word "cepat" is 93% and the word "lambat" is 90%.

Kata Kunci : Bahasa pemrograman C, pengenalan suara, MFCC, ANN