Frame-Level Multi-Label Playing Technique Detection Using Multi-Scale Network and Self-Attention Mechanism
This addresses the challenge of detecting overlapping or mixed IPTs in polyphonic music, which is incremental as it builds on prior work focused on monophonic signals.
The paper tackles the problem of detecting instrument playing techniques (IPTs) in polyphonic instrumental solo pieces, specifically for the Guzheng, by formulating it as a frame-level multi-label classification and proposing a method using a multi-scale network and self-attention mechanism. The approach outperforms existing works by a large margin, indicating its effectiveness.
Instrument playing technique (IPT) is a key element of musical presentation. However, most of the existing works for IPT detection only concern monophonic music signals, yet little has been done to detect IPTs in polyphonic instrumental solo pieces with overlapping IPTs or mixed IPTs. In this paper, we formulate it as a frame-level multi-label classification problem and apply it to Guzheng, a Chinese plucked string instrument. We create a new dataset, Guzheng\_Tech99, containing Guzheng recordings and onset, offset, pitch, IPT annotations of each note. Because different IPTs vary a lot in their lengths, we propose a new method to solve this problem using multi-scale network and self-attention. The multi-scale network extracts features from different scales, and the self-attention mechanism applied to the feature maps at the coarsest scale further enhances the long-range feature extraction. Our approach outperforms existing works by a large margin, indicating its effectiveness in IPT detection.