Audio declipping performance enhancement via crossfading

Some audio declipping methods produce waveforms that do not fully respect the physical process of clipping, which is why we refer to them as inconsistent. This letter reports what effect on perception it has if solution by inconsistent forced consistent postprocessing. We first propose a simple sample replacement method, then identify its main weaknesses and an improved variant. The experiments show vast majority significantly benefit from proposed approach in terms objective perceptual metrics. In particular, SS PEW method based social sparsity combined with performs comparable top class, but at computational cost one order magnitude lower.