Gamelan instrument sound recognition using spectral and facial features of the first harmonic frequency

Abstract

Principal component and spectral-based feature sets were applied to the recognition of gamelan instrument sounds using support vector machines (SVMs). The principal components were calculated on the basis of a segmented scalogram from the first harmonic frequency of the gamelan recordings. The segmented scalogram is assumed as a ''facial image'' of the gamelan instrument sound in a frontal pose, neutral expression, and normal lighting. The scalogram was computed from the gamelan sound signal using a continuous wavelet transform (CWT). The performance and contribution of the principal component and spectral-based features were compared using an F-measure. For the training phase, the feature sets were extracted from isolated tones that were recorded over the entire frequency range of four gamelan instruments (demung, saron, peking, and bonang families). Using 90%/10% splits between the training and validating data sets, model classifiers were constructed from the radial basis function (RBF) kernel SVM. The classifiers are composed of 28 separate One-AgainstOne multiclass classifiers. The experiment showed that the spectral-based feature set shows an average F-measure of 74.05% and the appearance-based feature yields 71.87%. For saron-only note tracking, the spectral-based feature set had an F-measure of 83.79%, higher than the demung-only note tracking, which yielded 63.89%.

Deep learning for audio signal-based tempo classification scenarios

“…The results of the tests indicate that spectrum-based feature sets have an average F-size that is greater than appearance-based features. While the recognition of distinct tones for the musical instrument demung (63.89%) is lower than that for the saron (83.79%), the former is higher [8].…”

“…The suggested approach demonstrates that it can generate an F-measure greater than 0.80 for some methods by adjusting the window length and selecting the proper dynamic threshold parameter [4]. Additionally, Tjahyanto et al [8] employs the principal component approach and spectrum-based feature sets as feature extraction for the sound classification of gamelan instruments with the SVM method on RBF kernel [8]. These four categories of gamelan instruments are demung, saron, peking, and bonang.…”

Gamelan Demung Music Transcription Based on STFT Using Deep Learning

“…Compared to the western musical instrument, the same Demung instrument has a slightly different frequency, resonance, and amplitude. It is because the production of Gamelan is handmade by the artisans [6] and the tuning practice isn't based on mathematical purity of interval vibration ratio [7]. Therefore, music transcription for the gamelan instrument is interesting and challenging.…”