Spectral Denoising for Microphone Classification

TL;DR

This paper introduces a denoising approach to improve microphone classification accuracy in noisy environments, demonstrating a 25% average accuracy boost across various noise levels.

Contribution

It presents a novel integration of denoising techniques into microphone classification, identifying the most effective method for noisy conditions.

Findings

25% average accuracy improvement with denoising

Effective denoising methods identified for spectral and time domains

Robust performance across different SNR levels

Abstract

In this paper, we propose the use of denoising for microphone classification, to enable its usage for several key application domains that involve noisy conditions. We describe the proposed analysis pipeline and the baseline algorithm for microphone classification, and discuss various denoising approaches which can be applied to it within the time or spectral domain; finally, we determine the best-performing denoising procedure, and evaluate the performance of the overall, integrated approach with several SNR levels of additive input noise. As a result, the proposed method achieves an average accuracy increase of about 25% on denoised content over the reference baseline.