| dc.description.abstract | Acoustic guitar can produce sound waves with different types of tones. High-low tone is determined by the fundamental frequency of the sound wave. Human sense of hearing can distinguish high-and low tones, but can not know for sure what kind of tone is heard by him. This study developed a guitar chord recognition system. This study uses MFCC as feature extraction methods and the codebook as pattern recognition method for identification of the guitar chord. Cluster technique used in this study is the K-means clustering. Data obtained from MFCC are clustered using the K-Means method, and the model for classification is constructed using the codebook method. The parameters used in MFCC is the number of cepstral coefficients, overlap, and the time frame, while an important parameter in the K-Means is the number of clusters. This study uses 8640 guitar chord sounds from 24 classes. Each of the data is made up of two chords which will be tested. Simulation results show that the maximum accuracy obtained is 98.89% obtained on 26 cepstral coefficients, overlap 0.4, time frame is 30 ms, on 100 clusters. | en |
