Indian Musical Instrument Audio Dataset: Flute, Harmonium, Sarod, and Tabla
Year: 2026
Type: other
Abstract: This dataset comprises 1,214 audio recordings of four Indian classical instruments: Flute, Harmonium, Sarod, and Tabla. It was collected for instrument classification and music information retrieval research, as no existing public corpus covered all four instruments with sufficient volume. Recordings were gathered through two channels. The first is performance videos sourced from YouTube, Facebook, and Instagram using instrument specific search terms, filtered for isolated instrument sound with minimal background noise or vocals, and converted from video to audio (MP3/M4A). The second is recordings contributed directly and willingly by individual artists in semi controlled acoustic environments. Class distribution: Flute (300 files, 24.7%), Harmonium (313 files, 25.8%), Sarod (301 files, 24.8%), Tabla (300 files, 24.7%). Each clip is up to 5 seconds in duration and contains a single, clearly audible instrument. This is a simple, straightforward dataset intended for academic and research use. See the included README for full methodology and known limitations.
Authors: Sharika Ali, Nath, Bidhan, Barua, Adity
Institution: Premier University
Language: English
Topic: Music and Audio Processing
Subfield: Signal Processing
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7204698153