Deteksi Nada Dasar Alat Musik Panting Menggunakan Compressive Sensing dengan MFCC dan SVM

Abstract

During this time, Panting’s calibration process has been done with mere reliance on human hearing.  Thus, this study is intended to assist the calibration process of Panting. The expirement has been done by using Compressive Sensing, MFCC and SVM. Moreover, the six scenarios are applied to the classification process are classification using five audios with different compression ratios and the original audio. This study results the best state of the Panting’s note recognition system is given by using the compressed audio with ratio of 2,5%. It is proven by its 97,96% of accuracy that is computed in 0,06274 second of duration.