English

Does k Matter? k-NN Hubness Analysis for Kernel Additive Modelling Vocal Separation

Sound 2018-04-09 v1 Audio and Speech Processing

Abstract

Kernel Additive Modelling (KAM) is a framework for source separation aiming to explicitly model inherent properties of sound sources to help with their identification and separation. KAM separates a given source by applying robust statistics on the selection of time-frequency bins obtained through a source-specific kernel, typically the k-NN function. Even though the parameter k appears to be key for a successful separation, little discussion on its influence or optimisation can be found in the literature. Here we propose a novel method, based on graph theory statistics, to automatically optimise kk in a vocal separation task. We introduce the k-NN hubness as an indicator to find a tailored k at a low computational cost. Subsequently, we evaluate our method in comparison to the common approach to choose k. We further discuss the influence and importance of this parameter with illuminating results.

Keywords

Cite

@article{arxiv.1804.02325,
  title  = {Does k Matter? k-NN Hubness Analysis for Kernel Additive Modelling Vocal Separation},
  author = {Delia Fano Yela and Dan Stowell and Mark Sandler},
  journal= {arXiv preprint arXiv:1804.02325},
  year   = {2018}
}

Comments

LVA-ICA 2018 - Feedback always welcome

R2 v1 2026-06-23T01:16:15.365Z