A new efficient backward BSS crosstalk-resistant algorithm for automatic blind speech quality enhancement

Noise reduction for speech applications is often formulated as a digital filtering problem, where the clean speech estimate is obtained by passing the noisy speech through a linear filter/transform. With such a formulation, the core issue of noise reduction becomes how to design an optimal filter (based on the statistics of the speech and noise signals) that can significantly suppress noise without introducing perceptually noticeable speech distortion. The optimal filters can be designed either in the time or in a transform domain. The advantage of working in a transform space is that, if the transform is selected properly, the speech and noise signals may be better separated in that space, thereby enabling better filter estimation and noise reduction performance. Although many different transforms exist, most efforts in the field of noise reduction have been focused only on the Fourier and Karhunen-Loeve transforms. Even with these two, no formal study has been carried out to investigate which transform can outperform the other. In this paper, we reformulate the noise reduction problem into a more generalized transform domain. We will show some of the advantages of working in this generalized domain, such as 1) different transforms can be used to replace each other without any requirement to change the algorithm (optimal filter) formulation, and 2) it is easier to fairly compare different transforms for their noise reduction performance. We will also address how to design different optimal and suboptimal filters in such a generalized transform domain.

Chapter 3 - Modified least mean square adaptive filter for speech enhancement

Hybrid PSO-NLMS (HPSO-NLMS) algorithm for blind speech quality enhancement in time domain

In this paper a robust and reliable adaptive noise cancellation algorithm is proposed to enhance noise carrying speech signals. The algorithm deals with situations where the microphones should locate in close proximity, as they cancel out the crosstalk effects. In other words, the source of the noise signal is not available separately and is independent of the desired speech signal. This is the case in mobile phones and hands-free systems, where the smallness of the dimension of the applied speech enhancement system is desirable. In our proposed method, Symmetric Adaptive Decorrelation (SAD) structure and Least Mean Mixed-Norm (LMMN) adaptation algorithm are applied to make a considerable improvement in the performance of the noise cancellation procedure, where the Voice Activity Detection (VAD) is used to reduce the computational costs. However in this paper, there was an utmost attempt to consider all of the practical problems, while the minimum simplifying assumptions are made. The simulation results have proven the robustness of this algorithm compared with commonly used algorithms, in the sense of SNR and MSE improvement and speech intelligibility.

Pervasive computing with voice application service must be providing unprecedented levels of technology in order to adapt the continually changed computing resources in ubiquitous environment. A number of conditions differ from the speech signal over IP network with heterogeneous devices while voice traffic is transported by packet-based in ubiquitous environment. In this paper, voice signals in packet loss and noise is controlled for the guaranteed quality of service (QoS) on the IP network which are responsible for digitizing, encoding, decoding, and playing out the speech signal. We implement the algorithms of packet loss concealment (PLC) and noise reduction, which is called PEA (pervasive environment adaptation)-PLC, PEA-noise reduction together, for providing high quality voice-based pervasive computing services. The proposed PEA-PLC algorithms are applied as post-processing after decoding on devices, and then PEA-noise reduction method is applied as pre-processing before speech coding. The proposed PEA-algorithms for packet loss and noise are applied in G.729A codec optimally using mobile-softphone and a 1-port gateway PSTN phone system. The performances of the proposed methods are evaluated by the PESQ measurement in various packet loss and noisy conditions

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

In this paper, we address the problem of speech enhancement by adaptive filtering algorithm. A particular attention has been made for the backward blind source separation (BBSS) algorithm and its use in speech enhancement application. In this paper, we propose to implement the BBSS algorithm in the Wavelet-domain. The proposed backward wavelet BBSS (WBBSS) algorithm is then used in speech enhancement test. The new WBBSS shows better performances in terms of convergence speed and steady state in comparison with BBSS one. The obtained results have been evaluated in terms of segmental SNR and cepstral distance criteria and confirm the best performance of the proposed WBBSS algorithm.

Multiple noise sources in a realistic environment severely degrade the quality and intelligibility of the desired speech signal, thus posing a severe problem for many speech applications. Several noise reduction algorithms have been proposed with a main goal to solve this problem. However, the good performances of such algorithms are severely impaired in realistic environment under multi-noise sources condition. In this paper, the author treats the noise cancellation system as a multiple-input multiple-output (MIMO) beamformer system. The proposed approach consists of two steps. First, the noise signals are generated by applying the white noise sources to a MIMO AR system. Then, the noisy microphone signals are sequentially processed by employing multi-channel linear prediction error filters (MCLPEFs) and multi-channel adaptive noise estimation filters (MCANEFs) in the lower path of the proposed beamformer. The MCLPEFs are used to whiten the input signals, while the MCANEFs are used as a MIMO system identification to perform the modeling process of the noise signals. Finally, the noise signals in the upper path are subtracted from the estimated noises in the lower path to recover an enhanced speech signal. Moreover, the performance of the proposed MIMO approach was validated under a realistic environment with real noise sources.

This paper presents a new method for voice activity detection (VAD) based on the autoregressive-generalized autoregressive conditional heteroscedasticity (AR-GARCH) model. The speech signal is modeled as an AR-GARCH process in the time domain, and the likelihood ratio is computed and compared to a threshold. The time-varying variance of the speech signal needed for computing the likelihood function under speech presence hypothesis, is estimated using the AR-GARCH model. The model parameters are estimated using a novel technique based on the recursive maximum likelihood (RML) estimation. The variance of the additive noise, a critical issue in designing a VAD, is estimated using the improved minima controlled recursive averaging (IMCRA) method, which is properly modified to be applicable to noise variance estimation in the time domain. The performances of the VAD and the parameter estimation method are examined under several conditions. Experimental results indicate the robustness of the AR-GARCH based VAD both to noise variations and low signal-to-noise ratio (SNR) conditions.

Recently, the deep-belief-networks (DBN) based voice activity detection (VAD) has been proposed. It is powerful in fusing the advantages of multiple features, and achieves the state-of-the-art performance. However, the deep layers of the DBN-based VAD do not show an apparent superiority to the shallower layers. In this paper, we propose a denoising-deep-neural-network (DDNN) based VAD to address the aforementioned problem. Specifically, we pre-train a deep neural network in a special unsupervised denoising greedy layer-wise mode, and then fine-tune the whole network in a supervised way by the common back-propagation algorithm. In the pre-training phase, we take the noisy speech signals as the visible layer and try to extract a new feature that minimizes the reconstruction cross-entropy loss between the noisy speech signals and its corresponding clean speech signals. Experimental results show that the proposed DDNN-based VAD not only outperforms the DBN-based VAD but also shows an apparent performance improvement of the deep layers over shallower layers.

In normal human communication, people face the speaker when listening and usually pay attention to the speaker' face. Therefore, in robot audition, the recognition of the front talker is critical for smooth interactions. This paper presents an enhanced speech detection method for a humanoid robot that can separate and recognize speech signals originating from the front even in noisy home environments. The robot audition system consists of a new type of voice activity detection (VAD) based on the complex spectrum circle centroid (CSCC) method and a maximum signal-to-noise (Max-SNR) beamformer. This VAD based on CSCC can classify speech signals that are retrieved at the frontal region of two microphones embedded on the robot. The system works in real-time without needing training filter coefficients given in advance even in a noisy environment (SNR ≫ 0 dB). It can cope with speech noise generated from televisions and audio devices that does not originate from the center. Experiments using a humanoid robot, SIG2, with two microphones showed that our system enhanced extracted target speech signals more than 12 dB (SNR) and the success rate of automatic speech recognition for Japanese words was increased about 17 points.

The speech signal in general is corrupted by noise and the noise signal does not affect the speech signal uniformly over the entire spectrum. An improved Wiener filtering method is proposed in this paper for reducing background noise from speech signal in colored noise environments. In view of nonlinear variation of human ear sensibility in frequency spectrum, nonlinear multi-band Bark scale frequency spacing approach is used. The cross-correlation between the speech and noise signal is considered in the proposed method to reduce colored noise. To overcome harmonic distortion introduced in enhanced speech, in the proposed method regenerate the suppressed harmonics are regenerated. Objective and subjective tests were carried out to demonstrate improvement in the perceptual quality of speeches by the proposed technique.

The presence of acoustic noise in audio recordings is an ongoing issue that plagues many applications. This ambient background noise is difficult to reduce due to its unpredictable nature. Many single channel noise reduction techniques exist but are limited in that they may distort the desired speech signal due to overlapping spectral content of the speech and noise. It is therefore of interest to investigate the use of multichannel noise reduction algorithms to further attenuate noise while attempting to preserve the speech signal of interest. Specifically, this thesis looks to investigate the use of microphone arrays in conjunction with multichannel noise reduction algorithms to aid aiding in speaker identification. Recording a speaker in the presence of acoustic background noise ultimately limits the performance and confidence of speaker identification algorithms. In situations where it is impossible to control the noise environment where the speech sample is taken, noise reduction algorithms must be developed and applied to clean the speech signal in order to give speaker identification software a chance at a positive identification. Due to the limitations of single channel techniques, it is of interest to see if spatial information provided by microphone arrays can be exploited to aid in speaker identification. This thesis provides an exploration of several time domain multichannel noise reduction techniques including delay sum beamforming, multi-channel Wiener filtering, and Spatial-Temporal Prediction filtering. Each algorithm is prototyped and filter performance is evaluated using various simulations and experiments. A three-dimensional noise model is developed to simulate and compare the performance of the above methods and experimental results of three data collections are presented and analyzed. The algorithms are compared and recommendations are given for the use of each technique. Finally, ideas for future work are discussed to improve performance and implementation of these multichannel algorithms. Possible applications for this technology include audio surveillance, identity verification, video chatting, conference calling and sound source localization.

Speech enhancement technology has been applied to many research fields of speech signal processing, such as speech coding, speech recognition, noise suppression and so on. However, in practical applications, there is always noise in the environment, voice in the transmission, to avoid noise interference, and the quality of the environment noise has different effects on the quality of the speech signal. Therefore, speech enhancement algorithm can realize the original clean speech from noisy speech signal. The traditional single microphone in the process of dealing with environmental noise, the received target signal and noise signals are often overlapping in the time and frequency spectrum. In view of the traditional method, using microphone array can improve the quality of voice communication, and can make full use of the time and space information of speech signal. Therefore, the use of a microphone array can overcome the shortcomings of a single microphone.

The real-time implementation of the existing multi-channel Wiener filter (MWF) algorithms suffer from performance degradation due to the lack of robustness against estimation errors of the second-order statistics. The reasons are twofold: one, the estimation of the statistics relies on real voice activity detector (VAD), which often fails in adverse environments. Second, the MWF solutions involve estimation of the second order clean speech statistics, which also exaggerates the errors. This paper presents an MWF algorithm that requires neither VAD nor clean speech statistics. Performance evaluation under real scenarios shows that the proposed method outperforms the conventional MWF solution in terms of the trade-off between noise reduction and speech distortion.