Robust multi-channel crosstalk resistant noise cancellation based on SAD structure and LMMN Algorithm

Frequency-domain criterion for the speech distortion weighted multichannel Wiener filter for robust noise reduction

In last 10 years, several noise reduction (NR) algorithms have been proposed to be combined with the blind source separation techniques to separate speech and noise signals from blind noisy observations. More often, techniques use voice activity detector (VAD) systems for the optimal solution. In this paper, we propose a new backward blind source separation (BBSS) structure that uses the input correlation properties to provide: (i) high convergence rates and good tracking capabilities, since the acoustic environments imply long and time-variant noise paths, and (ii) low misalignment and robustness against different noise type variations and double-talk. The proposed algorithm has an automatic behavior to enhance noisy speech signals, and do not need any VAD systems to separate speech and noise signals. The obtained results in terms of several objective criteria show the good performance properties of the proposed algorithm in comparison with state-of-the-art algorithms.

In this paper, a variable threshold voice activity detector (VAD) is developed to control the operation of a two-sensor adaptive noise canceller (ANC). The VAD prohibits the reference input of the ANC from containing some strength of actual speech signal during adaptation periods. The novelty of this approach resides in using the residual output from the noise canceller to control the decisions made by the VAD. Thresholds of full-band energy and zero-crossing features are adjusted according to the residual output of the adaptive filter. Performance evaluation of the proposed approach is quoted in terms of signal to noise ratio improvements as well mean square error (MSE) convergence of the ANC. The new approach showed an improved noise cancellation performance when tested under several types of environmental noise. Furthermore, the computational power of the adaptive process is reduced since the output of the adaptive filter is efficiently calculated only during non-speech periods.

This dissertation presents a new model for two adaptive filtering algorithms based on fourth-order moments: the least-mean fourth (LMF) and the leastmean mixed-norm (LMMN) algorithms. The novelty of the new model is its emphasis on computing the probability of a reasonable performance of a single realization of the algorithm (in this case, convergence), instead of looking for average performance indices such as mean-square error.

When people listen to the phone in noisy environments, near-end listening enhancement (NELE) is a technology to enhance speech intelligibility against environmental noise. The complex environments in mobile communications have inspired many scholars to engage in NELE researches. Although they have proposed a lot of NELE systems, they only focus on the speech modification to enhance the intelligibility. Few scholars have attempted to further enhance the intelligibility by noise cancellation. Because traditional noise cancellation is based on adaptive filtering. If the adaptive filtering is used in the most common handset mode, the noise cancellation result will be poor because of inadequate feedback caused by the feedback microphone exposed to complex environments. With the booming of the deep neural network (DNN), DNN is able to predict noise signals for noise cancellation without the feedback microphone, especially for recurrent neural network (RNN). In this study, we propose a NELE System by RNN-based noise cancellation and speech modification (RNC-SM), which introduce a noise cancellation function after speech modification. Compared with existing NELE systems, RNC-SM system effectively improves the objective speech intelligibility index (SII) scores and the subjective listening quality.

Speech enhancement is today a growing necessity for a wide range of applications in which the noise-free speech signal is important and necessary for the processing to be continued. The main purpose of these speech enhancement techniques is on a higher level is to remove noise from the speech signal. The reverberation component in the corrupted speech signal is also removed using the auto-regressive techniques for better performances. In this paper, unscented Kalman filtering which is an adaptive algorithm is proposed that executes both denoising and dereverberation of the speech recorded in adverse conditions. The algorithm relies on the parameter such as mean and covariance of the state spaces created and updating the concerned measurements to provide the optimal denoised and dereverberated signal. This proposed algorithm is assessed with regard to quality of speech, intelligibility of speech and performance metrics like the figure of merit and cross correlation and is also compared with other denoising and dereverberation techniques. The trial outputs on executing the algorithm using the noisy reverberant speech exhibit the adequacy of the proposed adaptive enhancement algorithm.KeywordsSpeech enhancementAdaptive algorithmDenoisingDereverberationKalman filteringAuto-regressive filteringUnscented Kalman filterParameter estimationUnscented transformParameter estimation

Signal de-noising has been sparked and given a great attention by signal processing community since its applications are found in a diverse range of digital signal processing and computer vision problems. To improve signal quality, phase, frequency and power are the important features that should be preserved during the de-noising process. Adaptive filters have been widely used for this purpose due to their ability to cancel out noise signal from the corrupted one precisely. This paper presents a robust adaptive estimator for solving the problem of signal noise cancellation, based on a new adaptive algorithm derived from a new constrained optimization. Simulation results evaluated using MATLAB show that the proposed algorithm is appropriate for several forms of signals contaminated by diverse levels of noise power. The performance of the proposed algorithm is illustrated to be preferable in terms of the power signal to noise ratio, mean square error and time of speed convergence of filter parameters. It is compared to other conventional approaches such as least mean square and normalized least mean square algorithms with various values of white noise power, variance. It exhibits lower steady-state error and faster convergent time than the other implementations. Finally, an efficient performance is achieved comparable with recursive least square and affine projection algorithms.

The persistent improvement of the hybrid adaptive algorithms and the swift growth of signal processing chip enhanced the performance of signal processing technique exalted mobile telecommunication systems. The proposed Artificial Neural Network Hybrid Back Propagation Adaptive Algorithm (ANNHBPAA) for mobile applications exploits relationship among the pure speech signal and noise corrupted signal in order to estimate of the noise. An adaptive linear system responds for changes in its environment as it is operating. Linear networks are gets adjusted at each time step based on new input and target vectors can find weights and biases that minimize the networks sum squared error for recent input and target vectors. Networks of this kind are quite oftenly used for error cancellation, speech signal processing and control systems. Noise in an audio signal has become major problem and hence mobile communication systems are demanding noise-free signal. In order to achieve noise-free signal various research communities have provided significant techniques. Adaptive noise cancellation (ANC) is a kind of technique which helps in estimation of un-wanted signal and removes them from corrupted signal. This paper introduces an Adaptive Filter Based Noise Cancellation System (AFNCS) that incorporates a hybrid back propagation learning for the adaptive noise cancellation in mobile applications. An extensive study has been made to explore the effects of different parameters, such as number of samples, number of filter coefficients, step size and noise level at the input on the performance of the adaptive noise cancelling system. The proposed hybrid algorithm consists all the significant features of Gradient Adaptive Lattice (GAL) and Least Mean Square (LMS) algorithms. The performance analysis of the method is performed by considering convergence complexity and bit error rate (BER) parameters along with performance analyzed with varying some parameters such as number of filter coefficients, step size, number of samples and input noise level. The outcomes suggest the errors are reduced significantly when the numbers of epochs are increased. Also incorporation of less hidden layers resulted in negligible computational delay along with effective utilization of memory. All the results have been obtained using computer simulations built on MATLAB platform

One of the most important applications of adaptive filter is Interference or noise cancellation. The objective of adaptive interference cancellation is to obtain an estimate of the interfering signal and to subtract it from the corrupted signal and hence obtain a noise free signal. The tracking performances of the LMS and NLMS algorithms are compared when the input of the adaptive filter is no stationary For this purpose, the filter uses an adaptive algorithm to change the value of the filter coefficients, so that it acquires a better approximation of the signal after each iteration. The LMS (Least Mean Square), and its variant the NLMS (Normalized LMS) are two of the adaptive algorithms widely in use. This paper presents a comparative analysis of the LMS and the NLMS in case of interference cancellation from speech signals. For each algorithm, the effects of two parameters-filter length and step size have been analyzed. Finally, the performances of the two algorithms in different cases have been compared.

Acoustic communication with autonomous underwater vehicles (AUVs) is complicated by the vehicles self-noise, particularly on AUVs with on-board active acoustic systems. Fortunately, active emissions generated on-board the AUV provide a convenient reference signal characterizing the self-noise. This paper addresses the use of such reference signals to improve the performance of acoustic communication systems. To this end, two methods of noise cancellation are examined both theoretically and experimentally. The first, called the cascade method, makes use of a conventional adaptive noise cancellation algorithm followed by a receiver with adaptive equalization. The second, called the noise cancellation method, involves a recently-proposed multi-channel adaptive receiver algorithm. It is shown that the multi-channel receiver can operate reliably in a signal-to-noise ratio of -3 dB or worse, a gain of about 10 dB over a conventional receiver for the codes and data-rates used. The cascade method, however, has a performance inferior to that of the multi-channel receiver due to adaptation noise.

The complexity of the speech enhancement problem has motivated many different solutions. However, most techniques address situations in which the target speech is fully intelligible and the background noise energy is low in comparison with that of the speech. Thus while current enhancement algorithms can improve the perceived quality, the intelligibility of the speech is not increased significantly and may even be reduced. Recent research shows that intelligibility of very noisy speech can be improved by the use of a binary mask, in which a binary weight is applied to each time-frequency bin of the input spectrogram. There are several alternative goals for the binary mask estimator, based either on the Signalto-Noise Ratio (SNR) of each time-frequency bin or on the speech signal characteristics alone. Our approach to the binary mask estimation problem aims to preserve the important speech cues independently of the noise present by identifying time-frequency regions that contain significant speech energy. The speech power spectrum varies greatly for different types of speech sound. The energy of voiced speech sounds is concentrated in the harmonics of the fundamental frequency while that of unvoiced sounds is, in contrast, distributed across a broad range of frequencies. To identify the presence of speech energy in a noisy speech signal we have therefore developed two detection algorithms. The first is a robust algorithm that identifies voiced speech segments and estimates their fundamental frequency. The second detects the presence of sibilants and estimates their energy distribution. In addition, we have developed a robust algorithm to estimate the active level of the speech. The outputs of these algorithms are combined with other features estimated from the noisy speech to form the input to a classifier which estimates a mask that accurately reflects the time-frequency distribution of speech energy even at low SNR levels. We evaluate a mask-based speech enhancer on a range of speech and noise signals and demonstrate a consistent increase in an objective intelligibility measure with respect to noisy speech.

Analysis of derivative of instantaneous frequency and its application to voice activity detection

<p>A combined three-microphone voice activity detector (VAD) and noise-canceling system is studied to enhance speech recognition in an automobile environment. A previous experiment clearly shows the ability of the composite system to cancel a single noise source outside of a defined zone. This paper investigates the performance of the composite system when there are frequently moving noise sources (noise sources are coming from different locations but are not always presented at the same time) e.g. there is other passenger speech or speech from a radio when a desired speech is presented. To work in a frequently moving noise sources environment, whilst a three-microphone voice activity detector (VAD) detects voice from a “VAD valid zone", the 3-microphone noise canceller uses a “noise canceller valid zone" defined in freespace around the users head. Therefore, a desired voice should be in the intersection of the noise canceller valid zone and VAD valid zone. Thus all noise is suppressed outside this intersection of area. Experiments are shown for a real environment e.g. all results were recorded in a car by omni-directional electret condenser microphones.</p>

In recent times noise cancellation is a vital issue, as it is responsible for reducing undesired disturbances in the process of communication. Active Noise Cancellation (ANC) is the most effective technique to cancel noise. ANC has been an active area of research and various adaptive methodologies have been employed to achieve a better ANC scheme. In the ANC technique, the aim is to minimize the noise interference that corrupts the original signal. ANC has a broad variety of applications in common commercial products, industrial uses and other machinery. An analysis of the prevailing adaptive methodologies is necessary for future research as we need to know the demerits of all the prevailing adaptive methodologies in ANC. This paper provides an analysis of various adaptive algorithms for noise cancellation and a comparison is made between them. The strengths, weaknesses and practical effectiveness of all the algorithms have been discussed. This paper deals with cancellation of noise on speech signal using three existing algorithms- Least Mean Square (LMS) algorithm, Normalized Least Mean Square (NLMS) algorithm and Recursive Least Square (RLS) algorithm and a proposed algorithm- advanced Block Least Mean Square (BLMS) algorithm. The algorithms are simulated in MATLAB platform and a tabular comparison of the algorithms is drawn. Finally, conclusions have been drawn by choosing the algorithms that provide efficient performance with less computational complexity.

Internet of things (IoT) has made possible the sharing of data wirelessly among everyday objects and physical devices. However, distortion and noise can significantly degrade the signaling performance of IoT. Adaptive Noise Cancellation (ACN) technology plays an important role not only in IoT but also in other numerous fields such as system identification, noise cancellation, communication, signal control, etc. At present, IoT is an advance trend and need to deal with upcoming issues in a way to deal properly in enjoying this technology. Due to the presence of noise in the signal, the quality of the signal is reduced, so the quality of communication (signal) can be improved by adaptive noise (acoustic) cancellation technology. This paper aims to provide the performance analysis of adaptive algorithm and the application of adaptive noise cancellation technique. Further, use cases of three-noise cancellation filter in IoT are introduced. In wireless communication, quality of services relies on signal quality and thus noise cancellation plays a vital role in the signaling of IoT. At the end of this paper performance analysis of different ANC algorithms including least mean square (LMS), Normalized least mean square (NLMS), recursive least square (RLS) will be done based on simulation results.