Predominant Musical Instrument Classification based on Spectral Features

This work aims to examine one of the cornerstone problems of MusicalInstrument Retrieval (MIR), in particular, instrument classification. IRMAS(Instrument recognition in Musical Audio Signals) data set is chosen for thispurpose. The data includes musical clips recorded from various sources in thelast century, thus having a wide variety of audio quality. We have presented avery concise summary of past work in this domain. Having implemented varioussupervised learning algorithms for this classification task, SVM classifier hasoutperformed the other state-of-the-art models with an accuracy of 79%. We alsoimplemented Unsupervised techniques out of which Hierarchical Clustering hasperformed well.