Audio Inpainting

Abstract

The objective of audio inpainting is to conceal missing or heavily distorted segments in an audio signal. This is ideally done by reconstructing the original signal or, at least, by inferring a meaningful surrogate signal. We present here a novel approach applying sparse modeling in the time-frequency (TF) domain which was recently published. In particular, we devise a dictionary learning technique which learns the dictionary from reliable parts around the gap with the goal to obtain a signal representation with improved TF sparsity. The presented method is based on a basis optimization technique that deforms a given Gabor frame such that the sparsity of the analysis coefficients of the resulting frame is maximized (with respect to a certain sparsity measure). Furthermore, we modify the SParse Audio INpainter (SPAIN) such that it is able to exploit the improved TF sparsity and—in turn—benefits from dictionary learning. As an alternative, we combine dictionary learning with weighted ℓ1-minimization adapted for audio inpainting to compensate for the loss of energy within the gap after restoration. Our experiments demonstrate that the developed methods achieve significant gains in terms of signal-to-distortion ratio (SDR) and objective difference grade (ODG) compared with several state-of-the-art audio inpainting techniques.