Problem
Research questions and friction points this paper is trying to address.
Vocal Source Separation
Multi-Singer Mixtures
Singer-Informed Separation
Target Singer Extraction
Singer-Informed Vocal Source Separation for Multi-Singer Music Mixtures
🤖 AI Summary
This study addresses the challenge of target vocal separation in multi-singer mixtures by proposing a singer-informed separation framework. The method leverages short reference recordings of the target singer to generate embeddings that guide the model via feature concatenation or FiLM modulation, effectively focusing on the target while suppressing interference. Additionally, a high-quality duet dataset is constructed to facilitate training. Experimental results demonstrate that this approach overcomes the limitations of traditional single-vocal separation, improving the target SI-SDR from 0.33 dB to 5.58 dB with significant gains in FAD scores. These findings validate the framework’s effectiveness and perceptual quality advantages in accurately extracting target vocals within complex multi-singer scenarios.
📝 Abstract
Music source separation systems typically extract a single vocal track and do not distinguish between multiple singers. We study singer-informed vocal source separation for multi-singer mixtures. Our framework introduces a short enrollment recording of a target singer to guide separation through a learned embedding. The singer embedding is incorporated using feature concatenation or feature-wise linear modulation (FiLM), enabling the model to focus on the target singer while suppressing interference. We construct a duet dataset based on DAMP-VSEP with quality filtering and non-overlapping enrollment segments. Experiments on solo and duet settings show that while baseline models perform well for single-singer mixtures, the proposed method improves target-singer extraction in multi-singer cases, increasing target-singer SI-SDR from 0.33 dB to 5.58 dB. Fréchet Audio Distance (FAD) further shows improved perceptual quality and better alignment with target audio distributions. Code and checkpoints are available at https://github.com/jocelynxu01/singer-separation-paper.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Singer-Informed Separation
Singer Embedding
Feature-wise Linear Modulation (FiLM)
Multi-Singer Mixture
Vocal Source Separation
🔎 Similar Papers
No similar papers found.
