Audio Signal Enhancement Research And Realization Based On K-SVD Algorithm

With the surging development of information technology and the rapid popularization of Internet,there's an increasing demand for information in our daily life.According to the traditional Nyquist sampling theorem,to avoid losing the necessary information when acquiring the signal,a signal must be sampled at least two times the bandwidth of itself.However,in many applications,such as imaging systems and high-speed analog-to-digital converters,the high sampling rate may result in unnecessary waste of resources and channels.Compressed sensing came into being with the higher demand for information storage and transmission,which combined sampling with compression.With the employment of non-adaptive linear projections and the proper optimization algorithm,it can preserve the structure of the signal and then lead to a quite satisfactory reconstructed result.Sparse representation of signals,as one of the core issues of compressed sensing,has become a hot topic among researchers since it can lead to a simple and effective reservation of signal structure.Based on lots of relevant literature,this paper gives a thorough study and discussion on sparse representation of signals and the recovery algorithm.Then,on the basis of summary and research on existing sparse dictionaries,the K-SVD algorithm,along with its improved algorithms,has been introduced in this paper.In real life,the continuous development of acoustic technology constantly improves the requirement of people's hearing experience.Thus,Dolby 5.1 surround sound system came into being.What's more,a low bitrate multi-channel audio transmission scheme based on compressed sensing technology is proposed.Experimental results show that our scheme has good effect on practical applications.It can effectively realize signal transmission and reconstruction at low sampling rates while reducing storage.At the same time,for the white noise interference which may be encountered in the signal transmission process,this scheme has a strong enhancement effect.The system has strong robustness under the noise interference,and can guarantee the signal transmission quality under the condition of low sampling rate.In order to verify the feasibility of the audio enhancement part in the audio transmission scheme,the advantages and disadvantages of improved K-SVD algorithm based on Bayesian method are briefly analyzed.As an unsteady signal,audio signal will inevitably be disturbed by noise in the transmission process.How to minimize the interference while retaining the detailed information in the observation data is of great significance.Because of the sparse qualification in audio signals,we employ Bayesian K-SVD algorithm in the enhancement process,with a comparison to other classic algorithms.Experimental results show that the proposed method can significantly remove white noise and has better robustness.