Multichannel audio coding using sinusoidal modelling and compressed sensing

This paper explores the potential of applying compressed sensing (CS) to multichannel audio coding. In this context, we consider how sinusoidally-modelled multichannel audio signals might be encoded using compressed sensing, as opposed to directly encoding the sinusoidal parameters (amplitude, frequency, phase) as current state-of-the-art methods do. The results, obtained from listening tests using 80 sinusoids per frame with no residual noise signal, show that such a model can achieve equal or better performance to that of the state-of-the-art methods. Given that CS can lead to novel coding systems where the sampling and compression operations are combined into one low-complexity step, this can be considered as an important step towards applying the CS framework to audio coding applications.