Frame-level Instrument Recognition by Timbre and Pitch

TL;DR

This paper presents a convolutional neural network approach for frame-level instrument recognition in multi-instrument music, utilizing pitch information to improve accuracy, and demonstrates significant performance gains on the MusicNet dataset.

Contribution

It introduces a novel method that incorporates pitch data into CNN-based instrument recognition, enhancing multi-instrument detection at the frame level.

Findings

Significant performance improvement over baseline methods.

Incorporating pitch information aids in instrument prediction.

Analysis shows pitch helps distinguish instruments effectively.

Abstract

Instrument recognition is a fundamental task in music information retrieval, yet little has been done to predict the presence of instruments in multi-instrument music for each time frame. This task is important for not only automatic transcription but also many retrieval problems. In this paper, we use the newly released MusicNet dataset to study this front, by building and evaluating a convolutional neural network for making frame-level instrument prediction. We consider it as a multi-label classification problem for each frame and use frame-level annotations as the supervisory signal in training the network. Moreover, we experiment with different ways to incorporate pitch information to our model, with the premise that doing so informs the model the notes that are active per frame, and also encourages the model to learn relative rates of energy buildup in the harmonic partials of different instruments. Experiments show salient performance improvement over baseline methods. We also report an analysis probing how pitch information helps the instrument prediction task. Code and experiment details can be found at \url{https://biboamy.github.io/instrument-recognition/}.