Assessment of musical noise using localization of isolated peaks in time-frequency domain

Abstract

Musical noise is a recurrent issue that appears in spectral techniques for denoising or blind source separation. Due to localised errors of estimation, isolated peaks may appear in the processed spectrograms, resulting in annoying tonal sounds after synthesis known as "musical noise". In this paper, we propose a method to assess the amount of musical noise in an audio signal, by characterising the impact of these artificial isolated peaks on the processed sound. It turns out that because of the constraints between STFT coefficients, the isolated peaks are described as time-frequency "spots" in the spectrogram of the processed audio signal. The quantification of these "spots", achieved through the adaptation of a method for localisation of significant STFT regions, allows for an evaluation of the amount of musical noise. We believe that this will pave the way to an objective measure and a better understanding of this phenomenon.

An Improved Measure of Musical Noise Based on Spectral Kurtosis

“…Musical noise is an annoying distortion for the listener, who perceives it as unnatural and warbling fluctuations [1]. Countermeasures and improved methods were developed (e.g., [2,3]) as well as objective measures trying to predict the perceived audio quality degradation, e.g., [4].…”

Multivariate iterative hard thresholding for sparse decomposition with flexible sparsity patterns

“…One shortcoming of a batch processing is that every frame is processed with the same number of iteration, and the stopping criterion y t −Dx n t 2 2 > cannot be readily applied. This is of utmost importance in the denoising case, since the condition y t −Dx n t 2 2 > reflects our knowledge about the noise level, and can lead to artifacts when it is not respected [26]. In order to deal with the stopping criterion for each frame t, we add a masking matrix M ∈ {0, 1} M ×T in the gradient descent step.…”