Singing voice separation (SVS) can be defined as the process of extracting the vocal element from a given song recording. The impetus for research in this area is mainly that of facilitating certain important applications of music information retrieval (MIR) such as lyrics recognition, singer identification, and melody extraction. To date, the research in the field of SVS has been relatively limited, and mainly focused on the extraction of vocals from monophonic sources. The general approach in this scenario has been one of considering SVS as a blind source separation (BSS) problem. Given the inherent diversity of music, such an approach is motivated by the quest for a generic solution. However, it does not allow the exploitation of prior information, regarding the way in which commercial music is produced. To this end, investigations are conducted into effective methods for unsupervised separation of singing voice from stereophonic studio recordings. The work involves extensive literature review of existing methods that relate to SVS, as well as commercial approaches. Following the identification of shortcomings of the conventional methods, two novel approaches are developed for the purpose of SVS. These approaches, termed SEMANICS and SEMANTICS draw their motivation from statistical as well as spectral properties of the target signal and focus on the separation of voice in the frequency domain. In addition, a third method, named Hybrid SEMANTICS, is introduced that addresses time‐, as well as frequency‐domain separation. As there is lack of a concrete standardised music database that includes a large number of songs, a dataset is created using conventional stereophonic mixing methods. Using this database, and based on widely adopted objective metrics, the effectiveness of the proposed methods has been evaluated through thorough experimental investigations.

While humans can separate a sound of interest amidst a cacophony of contending sounds in an echoic environment, machine-based methods lag behind in solving this task. This thesis thus aims at improving performance of audio separation algorithms when they are informed i.e. have access to source location information. These locations are assumed to be known a priori in this work, for example by video processing. Initially, a multi-microphone array based method combined with binary time-frequency masking is proposed. A robust least squares frequency invariant data independent beamformer designed with the location information is utilized to estimate the sources. To further enhance the estimated sources, binary time-frequency masking based post-processing is used but cepstral domain smoothing is required to mitigate musical noise. To tackle the under-determined case and further improve separation performance at higher reverberation times, a two-microphone based method which is inspired by human auditory processing and generates soft time-frequency masks is described. In this approach interaural level difference, interaural phase difference and mixing vectors are probabilistically modeled in the time-frequency domain and the model parameters are learned through the expectation-maximization (EM) algorithm. A direction vector is estimated for each source, using the location information, which is used as the mean parameter of the mixing vector model. Soft time-frequency masks are used to reconstruct the sources. A spatial covariance model is then integrated into the probabilistic model framework that encodes the spatial characteristics of the enclosure and further improves the separation performance in challenging scenarios i.e. when sources are in close proximity and when the level of reverberation is high. Finally, new dereverberation based pre-processing is proposed based on the cascade of three dereverberation stages where each enhances the twomicrophone reverberant mixture. The dereverberation stages are based on amplitude spectral subtraction, where the late reverberation is estimated and suppressed. The combination of such dereverberation based pre-processing and use of soft mask separation yields the best separation performance. All methods are evaluated with real and synthetic mixtures formed for example from speech signals from the TIMIT database and measured room impulse responses.

In an exploratory approach to data analysis, it is often useful to consider the observations as generated from a set of latent generators or 'sources' via a generally unknown mapping. Reconstructing sources from their mixtures is an extremely ill-posed problem in general. However, solutions to such inverse problems can, in many cases, be achieved by incorporating prior knowledge about the problem, captured in the form of constraints. This setting is a natural candidate for the application of the Bayesian method- ology, allowing us to incorporate "soft" constraints in a natural manner. This Thesis proposes the use of sparse statistical decomposition methods for ex- ploratory analysis of datasets. We make use of the fact that many natural signals have a sparse representation in appropriate signal dictionaries. The work described in this Thesis is mainly driven by problems in the analysis of large datasets, such as those from functional magnetic resonance imaging of the brain for the neuro-scientific goal of extracting relevant 'maps' from the data. We first propose Bayesian Iterative Thresholding, a general method for solv- ing blind linear inverse problems under sparsity constraints, and we apply it to the problem of blind source separation. The algorithm is derived by maximiz- ing a variational lower-bound on the likelihood. The algorithm generalizes the recently proposed method of Iterative Thresholding. The probabilistic view en- ables us to automatically estimate various hyperparameters, such as those that control the shape of the prior and the threshold, in a principled manner. We then derive an efficient fully Bayesian sparse matrix factorization model for exploratory analysis and modelling of spatio-temporal data such as fMRI. We view sparse representation as a problem in Bayesian inference, following a ma- chine learning approach, and construct a structured generative latent-variable model employing adaptive sparsity-inducing priors. The construction allows for automatic complexity control and regularization as well as denoising. The performance and utility of the proposed algorithms is demonstrated on a variety of experiments using both simulated and real datasets. Experimental results with benchmark datasets show that the proposed algorithms outper- form state-of-the-art tools for model-free decompositions such as independent component analysis.

由於在盲語音分離的應用，線性準平穩源訊號混合的盲識別獲得了巨大的研究興趣。在這個問題上，我們利用準穩態源訊號的時變特性來識別未知的混合系統系數。傳統的方法有二：i)基於張量分解的平行因子分析(PARAFAC);ii)基於對多個矩陣的聯合對角化的聯合對角化算法(JD)。一般來說，PARAFAC和JD 都採用了源聯合的提取方法；即是說，對應所有訊號源的系統係數在升法上是用時進行識別的。 / 在這篇論文中，我利用Khati-Rao(KR)子空間來設計一種新的盲識別算法。在我設計的算法中提出一種與傳統的方法不同的提法。在我設計的算法中，盲識別問題被分解成數個結構上相對簡單的子問題，分別對應不同的源。在超定混合模型，我們提出了一個專門的交替投影算法(AP)。由此產生的算法，不但能從經驗發現是非常有競爭力的，而且更有理論上的利落收斂保證。另外，作為一個有趣的延伸，該算法可循一個簡單的方式應用於欠混合模型。對於欠定混合模型，我們提出啟發式的秩最小化算法從而提高算法的速度。 / Blind identification of linear instantaneous mixtures of quasi-stationary sources (BI-QSS) has received great research interest over the past few decades, motivated by its application in blind speech separation. In this problem, we identify the unknown mixing system coefcients by exploiting the time-varying characteristics of quasi-stationary sources. Traditional BI-QSS methods fall into two main categories: i) Parallel Factor Analysis (PARAFAC), which is based on tensor decomposition; ii) Joint Diagonalization (JD), which is based on approximate joint diagonalization of multiple matrices. In both PARAFAC and JD, the joint-source formulation is used in general; i.e., the algorithms are designed to identify the whole mixing system simultaneously. / In this thesis, I devise a novel blind identification framework using a Khatri-Rao (KR) subspace formulation. The proposed formulation is different from the traditional formulations in that it decomposes the blind identication problem into a number of per-source, structurally less complex subproblems. For the over determined mixing models, a specialized alternating projections algorithm is proposed for the KR subspace for¬mulation. The resulting algorithm is not only empirically found to be very competitive, but also has a theoretically neat convergence guarantee. Even better, the proposed algorithm can be applied to the underdetermined mixing models in a straightforward manner. Rank minimization heuristics are proposed to speed up the algorithm for the underdetermined mixing model. The advantages on employing the rank minimization heuristics are demonstrated by simulations. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Lee, Ka Kit. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2012. / Includes bibliographical references (leaves 72-76). / Abstracts also in Chinese. / Abstract --- p.i / Acknowledgement --- p.ii / Chapter 1 --- Introduction --- p.1 / Chapter 2 --- Settings of Quasi-Stationary Signals based Blind Identification --- p.4 / Chapter 2.1 --- Signal Model --- p.4 / Chapter 2.2 --- Assumptions --- p.5 / Chapter 2.3 --- Local Covariance Model --- p.7 / Chapter 2.4 --- Noise Covariance Removal --- p.8 / Chapter 2.5 --- Prewhitening --- p.9 / Chapter 2.6 --- Summary --- p.10 / Chapter 3 --- Review on Some Existing BI-QSS Algorithms --- p.11 / Chapter 3.1 --- Joint Diagonalization --- p.11 / Chapter 3.1.1 --- Fast Frobenius Diagonalization [4] --- p.12 / Chapter 3.1.2 --- Pham’s JD [5, 6] --- p.14 / Chapter 3.2 --- Parallel Factor Analysis --- p.16 / Chapter 3.2.1 --- Tensor Decomposition [37] --- p.17 / Chapter 3.2.2 --- Alternating-Columns Diagonal-Centers [12] --- p.21 / Chapter 3.2.3 --- Trilinear Alternating Least-Squares [10, 11] --- p.23 / Chapter 3.3 --- Summary --- p.25 / Chapter 4 --- Proposed Algorithms --- p.26 / Chapter 4.1 --- KR Subspace Criterion --- p.27 / Chapter 4.2 --- Blind Identification using Alternating Projections --- p.29 / Chapter 4.2.1 --- All-Columns Identification --- p.31 / Chapter 4.3 --- Overdetermined Mixing Models (N > K): Prewhitened Alternating Projection Algorithm (PAPA) --- p.32 / Chapter 4.4 --- Underdetermined Mixing Models (N <K) --- p.34 / Chapter 4.4.1 --- Rank Minimization Heuristic --- p.34 / Chapter 4.4.2 --- Alternating Projections Algorithm with Huber Function Regularization --- p.37 / Chapter 4.5 --- Robust KR Subspace Extraction --- p.40 / Chapter 4.6 --- Summary --- p.44 / Chapter 5 --- Simulation Results --- p.47 / Chapter 5.1 --- General Settings --- p.47 / Chapter 5.2 --- Overdetermined Mixing Models --- p.49 / Chapter 5.2.1 --- Simulation 1 - Performance w.r.t. SNR --- p.49 / Chapter 5.2.2 --- Simulation 2 - Performance w.r.t. the Number of Available Frames M --- p.49 / Chapter 5.2.3 --- Simulation 3 - Performance w.r.t. the Number of Sources K --- p.50 / Chapter 5.3 --- Underdetermined Mixing Models --- p.52 / Chapter 5.3.1 --- Simulation 1 - Success Rate of KR Huber --- p.53 / Chapter 5.3.2 --- Simulation 2 - Performance w.r.t. SNR --- p.54 / Chapter 5.3.3 --- Simulation 3 - Performance w.r.t. M --- p.54 / Chapter 5.3.4 --- Simulation 4 - Performance w.r.t. N --- p.56 / Chapter 5.4 --- Summary --- p.56 / Chapter 6 --- Conclusion and Future Works --- p.58 / Chapter A --- Convolutive Mixing Model --- p.60 / Chapter B --- Proofs --- p.63 / Chapter B.1 --- Proof of Theorem 4.1 --- p.63 / Chapter B.2 --- Proof of Theorem 4.2 --- p.65 / Chapter B.3 --- Proof of Observation 4.1 --- p.65 / Chapter B.4 --- Proof of Proposition 4.1 --- p.66 / Chapter C --- Singular Value Thresholding --- p.67 / Chapter D --- Categories of Speech Sounds and Their Impact on SOSs-based BI-QSS Algorithms --- p.69 / Chapter D.1 --- Vowels --- p.69 / Chapter D.2 --- Consonants --- p.69 / Chapter D.1 --- Silent Pauses --- p.70 / Bibliography --- p.72

Thesis (Ph. D.)--University of Hong Kong, 2006. / Title proper from title frame. Also available in printed format.

Thesis (Master)--İzmir Institute of Technology, İzmir, 2004 / Includes bibliographical references (leaves. 56).

Music recordings contain the mixed contribution of multiple overlapping instruments. In order to better understand the music, it would be beneficial to understand each instrument independently. This thesis focuses on separating the individual instrument recordings within a song. In particular, we propose novel algorithms for separating instrument recordings given only their mixture. When the number of source signals does not exceed the number of mixture signals, we focus on a subclass of source separation algorithms based on joint diagonalization. Each approach leverages a different form of source structure. We introduce repetitive structure as an alternative that leverages unique repetition patterns in music and compare its performance against the other techniques. When the number of source signals exceeds the number of mixtures (i.e. the underdetermined problem), we focus on spectrogram factorization techniques for source separation. We extend single-channel techniques to utilize the additional spatial information in multichannel recordings, and use phase information to improve the estimation of the underlying components.

The objective of the dissertation research is to investigate noise reduction methods for binaural hearing aids based on array and statistical signal processing and inspired by a human auditory model. In digital hearing aids, wide dynamic range compression (WDRC) is the most successful technique to deal with monaural hearing losses. This WDRC processing is usually performed after a monaural noise reduction algorithm. When hearing losses are present in both ears, i.e., a binaural hearing loss, independent monaural hearing aids have been shown not to be comfortable for most users, preferring a processing that involves synchronization between both hearing devices. In addition, psycho-acoustical studies have identified that under hostile environments, e.g., babble noise at very low SNR conditions, users prefer to use linear amplification rather than WDRC. In this sense, the noise reduction algorithm becomes an important component of a digital hearing aid to provide improvement in speech intelligibility and user comfort. Including a wireless link between both hearing aids offers new ways to implement more efficient methods to reduce the background noise and coordinate processing for the two ears. This approach, called binaural hearing aid, has been recently introduced in some commercial products but using very simple processing strategies. This research analyzes the existing binaural noise-reduction techniques, proposes novel perceptually-inspired methods based on blind source separation (BSS) and multichannel Wiener filter (MWF), and identifies different strategies for the real-time implementation of these methods. The proposed methods perform efficient spatial filtering, improve SNR and speech intelligibility, minimize block processing artifacts, and can be implemented in low-power architectures.

We live in a dynamic world full of noises and interferences. The conventional acoustic echo cancellation (AEC) framework based on the least mean square (LMS) algorithm by itself lacks the ability to handle many secondary signals that interfere with the adaptive filtering process, e.g., local speech and background noise. In this dissertation, we build a foundation for what we refer to as the system approach to signal enhancement as we focus on the AEC problem. We first propose the residual echo enhancement (REE) technique that utilizes the error recovery nonlinearity (ERN) to "enhances" the filter estimation error prior to the filter adaptation. The single-channel AEC problem can be viewed as a special case of semi-blind source separation (SBSS) where one of the source signals is partially known, i.e., the far-end microphone signal that generates the near-end acoustic echo. SBSS optimized via independent component analysis (ICA) leads to the system combination of the LMS algorithm with the ERN that allows for continuous and stable adaptation even during double talk. Second, we extend the system perspective to the decorrelation problem for AEC, where we show that the REE procedure can be applied effectively in a multi-channel AEC (MCAEC) setting to indirectly assist the recovery of lost AEC performance due to inter-channel correlation, known generally as the "non-uniqueness" problem. We develop a novel, computationally efficient technique of frequency-domain resampling (FDR) that effectively alleviates the non-uniqueness problem directly while introducing minimal distortion to signal quality and statistics. We also apply the system approach to the multi-delay filter (MDF) that suffers from the inter-block correlation problem. Finally, we generalize the MCAEC problem in the SBSS framework and discuss many issues related to the implementation of an SBSS system. We propose a constrained batch-online implementation of SBSS that stabilizes the convergence behavior even in the worst case scenario of a single far-end talker along with the non-uniqueness condition on the far-end mixing system. The proposed techniques are developed from a pragmatic standpoint, motivated by real-world problems in acoustic and audio signal processing. Generalization of the orthogonality principle to the system level of an AEC problem allows us to relate AEC to source separation that seeks to maximize the independence, hence implicitly the orthogonality, not only between the error signal and the far-end signal, but rather, among all signals involved. The system approach, for which the REE paradigm is just one realization, enables the encompassing of many traditional signal enhancement techniques in analytically consistent yet practically effective manner for solving the enhancement problem in a very noisy and disruptive acoustic mixing environment.

M. Tech. Electrical Engineering. / Discusses node location as a major problem when considering several areas of application based on wireless sensor networks. Many localisation algorithms have been proposed in the literature to solve the problem of locating sensor nodes in WSN. However, most of these algorithms have poor localisation accuracy and high computational cost. Due to these limitations, this research study considers the modelling of an efficient and robust localisation scheme to determine the location of individual sensor nodes in WSN. To successfully solve this task, this research study focuses on the aspect of improving the position accuracy of wireless sensor nodes in WSN. The study considers a distance based cooperative localisation algorithm called Curvilinear Component Analysis Mapping (CCA-MAP) to accurately localise the sensor nodes in WSN. CCA-MAP is used because it delivers improved position accuracy and computational efficiency.