- 期刊
- OpenAccess
Blind Audio Source Separation with Sparse Nonnegative Matrix Factorization
並列摘要
In this study, a new technique in source separation using Two-Dimensional Nonnegative Matrix Factorization (NMF2D) with the Beta-divergence is proposed. The Time-Frequency (TF) profile of each source is modeled as two-dimensional convolution of the temporal code and the spectral basis. In addition, adaptive sparsity constraint was imposed to reduce the ambiguity and provide uniqueness to the solution. The proposed model used Beta-divergence as a cost function and updated by maximizing the joint probability of the mixing spectral basis and temporal codes using the multiplicative update rules. Experimental tests have been conducted in audio application to blindly separate the source in musical mixture. Results have shown the effectiveness of the algorithm in separating the audio sources from single channel mixture.
延伸閱讀
- 郭家維（2015）。Music Source Separation Based on Total Harmonic to Signal Ratio〔碩士論文，國立中央大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0031-0412201512035425
- Lin, C. C. (2008). S-transform with the Special Varying Window and Clustering Method for Blind Audio Separation [master's thesis, National Taiwan University]. Airiti Library. https://doi.org/10.6342/NTU.2008.00331
- 黃俊凱（2007）。Applying Hidden Markov Model and Observable Markov Model for Audio Content Identification〔碩士論文，國立臺北科技大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0006-0808200715024600
- Hou, Y. T. (2020). Blind Channel Estimation via MUSIC Algorithm and Compressed Sensing [master's thesis, National Taiwan University]. Airiti Library. https://doi.org/10.6342/NTU202000601
- 蕭昀軒（2011）。Multiple Source Tracking and Separation Using MUSIC Algorithm〔碩士論文，國立交通大學〕。華藝線上圖書館。https://doi.org/10.6842/NCTU.2011.00871