A constrained matching pursuit approach to audio declipping

Abstract

We present a novel sparse representation based approach for the restoration of clipped audio signals. In the proposed approach, the clipped signal is decomposed into overlapping frames and the declipping problem is formulated as an inverse problem, per audio frame. This problem is further solved by a constrained matching pursuit algorithm, that exploits the sign pattern of the clipped samples and their maximal absolute value. Performance evaluation with a collection of music and speech signals demonstrate superior results compared to existing algorithms, over a wide range of clipping levels.

VLSI Design of Approximate Message Passing for Signal Restoration and Compressive Sensing

“…In addition to CS, sparse signal recovery has been employed in the restoration of signals corrupted by impulse noise and/or saturation [2]- [4]. We next show how signal restoration can be formulated as a sparse signal recovery problem and demonstrate the suitability of AMP for this particular application.…”

“…2 A direct comparison is difficult, because the ASICs in [15] perform sparse signal recovery in 0.5 ms of signals with roughly 12 to 18 significant entries and problems of dimension 200×256; moreover, both applications have different precision requirements.…”

“…Here, the matrix A ∈ R M ×Na sparsifies the signal s = Aa to be restored and the matrix B ∈ R M ×N b sparsifies the corruptions on the signal s. Signal restoration now amounts to the recovery of the sparse vector x from (4) using, BPDN or BPDN * , for example, followed by computing s = Aa. In certain cases, one can identify the locations of the corrupted entries in z prior to recovery, which typically results in improved restoration performance [2]- [4].…”

“…Prominent examples are the restoration of audio signals or images from saturation, impulse noise, or narrowband interference [2]- [4], signal separation [5], de-noising [6], de-blurring [7], super-resolution [6], and in-painting [8], as well as compressive sensing (CS) [9], [10]. CS has recently gained significant attention in the research community by enabling the sampling of sparse signals using fewer measurements than the Nyquist rate suggests.…”

Design and Implementation of Different Audio Restoration Techniques for Audio Denoising Applications

“…It usually occurs in phone sound transmissions. It may occur in two forms, [1]. Hard clipping-In cases where the signal is strictly limited at the threshold, producing a flat cutoff, Fig-1 shows Hard clipping and normal signal waveforms.…”