Single-channel Noise Reduction Research Articles

Deep-Learning-Assisted Sound Source Localization From a Flying Drone

Sound source localization from a flying drone is a challenging task due to the strong ego-noise from rotating motors and propellers as well as the movement of the drone and the sound sources. To address this challenge, we propose a deep-learning-based framework that integrates single-channel noise reduction and multichannel source localization. In this framework, we suppress the ego-noise and estimate a time–frequency soft ratio mask with a single-channel deep neural network (DNN). Then, we design two downstream multichannel source localization algorithms, based on steered response power (SRP-DNN) and time–frequency spatial filtering (TFS-DNN). The main novelty lies in the proposed TFS-DNN approach, which estimates the presence probability of the target sound at the individual time–frequency bins by combining the DNN-inferred soft ratio mask and the instantaneous direction of arrival (DOA) of the sound received by the microphone array. The time–frequency presence probability of the target sound is then used to design a set of spatial filters to construct a spatial likelihood map for source localization. By jointly exploiting spectral and spatial information, TFS-DNN robustly processes signals in short segments (e.g., 0.5 s) in dynamic and low signal-to-noise-ratio (SNR) scenarios (e.g., SNR −20 dB). Results on real and simulated data in a variety of scenarios (static sources, moving sources, and moving drones) indicate the advantage of TFS-DNN over competing methods, including SRP-DNN and the state-of-the-art TFS.

Supervised speech separation combined with adaptive beamforming

Microphone arrays are a powerful tool for ambient noise suppression. A multi-channel minimum mean square error (MMSE) solution can be factorized into a minimum variance distortionless response beamformer (MVDR) followed by a single-channel Wiener post-filter. MVDR beamformer, as well as its equivalent form of generalized sidelobe canceller (GSC), often does not provide sufficient noise reduction due to its limited ability to reduce diffuse noise and reverberation. Steering and calibration errors also degrade the performance of both MVDR and GSC beamformers. Post-filter can be realized by any single-channel noise reduction method. A modern and promising approach for single-channel noise reduction is formulated as a supervised speech separation (SSS) in which a supervised learning algorithm, typically a deep neural network (DNN), is trained to learn a mapping from the noisy features to a time-frequency representation of the target of interest. In this paper, we combined SSS and adaptive beamforming approaches. Adaptive beamforming is realized by simplified GSC (S-GSC) whose equivalence with MVDR beamformer is also proved in the paper. In the proposed S-GSC beamformer, the conventional beamformer is replaced by the central microphone signal. Steering towards the target speaker needs no direction of arrival (DOA) estimation. Trained DNN of the SSS module estimates ideal ratio mask (IRM) which is used for adaptation of the blocking matrix, calibration of the microphones, adaptation for the adaptive noise canceller, and the post-filtering. The proposed method was tested on 720 utterances of the TIMIT database used as target speech. The reverberant room was simulated by acoustic impulse responses recorded in the real room. Performance analysis was carried out with PESQ, STOI, and SDR measures. The test results showed that the proposed combined method outperforms the individual SSS and S-GSC methods.

Infrasound single-channel noise reduction: application to detection and localization of explosive volcanism in Alaska using backprojection and array processing

SUMMARY Infrasound sensors are deployed in a variety of spatial configurations and scales for geophysical monitoring, including networks of single sensors and networks of multisensor infrasound arrays. Infrasound signal detection strategies exploiting these data commonly make use of intersensor correlation and coherence (array processing, multichannel correlation); network-based tracking of signal features (e.g. reverse time migration); or a combination of these such as backazimuth cross-bearings for multiple arrays. Single-sensor trace-based denoising techniques offer significant potential to improve all of these various infrasound data processing strategies, but have not previously been investigated in detail. Single-sensor denoising represents a pre-processing step that could reduce the effects of ambient infrasound and wind noise in infrasound signal association and location workflows. We systematically investigate the utility of a range of single-sensor denoising methods for infrasound data processing, including noise gating, non-negative matrix factorization, and data-adaptive Wiener filtering. For the data testbed, we use the relatively dense regional infrasound network in Alaska, which records a high rate of volcanic eruptions with signals varying in power, duration, and waveform and spectral character. We primarily use data from the 2016–2017 Bogoslof volcanic eruption, which included multiple explosions, and synthetics. The Bogoslof volcanic sequence provides an opportunity to investigate regional infrasound detection, association, and location for a set of real sources with varying source spectra subject to anisotropic atmospheric propagation and varying noise levels (both incoherent wind noise and coherent ambient infrasound, primarily microbaroms). We illustrate the advantages and disadvantages of the different denoising methods in categories such as event detection, waveform distortion, the need for manual data labelling, and computational cost. For all approaches, denoising generally performs better for signals with higher signal-to-noise ratios and with less spectral and temporal overlap between signals and noise. Microbaroms are the most globally pervasive and repetitive coherent ambient infrasound noise source, with such noise often referred to as clutter or interference. We find that denoising offers significant potential for microbarom clutter reduction. Single-channel denoising of microbaroms prior to standard array processing enhances both the quantity and bandwidth of detectable volcanic events. We find that reduction of incoherent wind noise is more challenging using the denoising methods we investigate; thus, station hardware (wind noise reduction systems) and site selection remain critical and cannot be replaced by currently available digital denoising methodologies. Overall, we find that adding single-channel denoising as a component in the processing workflow can benefit a variety of infrasound signal detection, association, and location schemes. The denoising methods can also isolate the noise itself, with utility in statistically characterizing ambient infrasound noise.

A New Approach for Noise Suppression in Cochlear Implants: A Single-Channel Noise Reduction Algorithm.

The cochlea “translates” the in-air vibrational acoustic “language” into the spikes of neural “language” that are then transmitted to the brain for auditory understanding and/or perception. During this intracochlear “translation” process, high resolution in time–frequency–intensity domains guarantees the high quality of the input neural information for the brain, which is vital for our outstanding hearing abilities. However, cochlear implants (CIs) have coarse artificial coding and interfaces, and CI users experience more challenges in common acoustic environments than their normal-hearing (NH) peers. Noise from sound sources that a listener has no interest in may be neglected by NH listeners, but they may distract a CI user. We discuss the CI noise-suppression techniques and introduce noise management for a new implant system. The monaural signal-to-noise ratio estimation-based noise suppression algorithm “eVoice,” which is incorporated in the processors of Nurotron® EnduroTM, was evaluated in two speech perception experiments. The results show that speech intelligibility in stationary speech-shaped noise can be significantly improved with eVoice. Similar results have been observed in other CI devices with single-channel noise reduction techniques. Specifically, the mean speech reception threshold decrease in the present study was 2.2 dB. The Nurotron society already has more than 10,000 users, and eVoice is a start for noise management in the new system. Future steps on non-stationary-noise suppression, spatial-source separation, bilateral hearing, microphone configuration, and environment specification are warranted. The existing evidence, including our research, suggests that noise-suppression techniques should be applied in CI systems. The artificial hearing of CI listeners requires more advanced signal processing techniques to reduce brain effort and increase intelligibility in noisy settings.

Quadratic approach for single-channel noise reduction

In this paper, we introduce a quadratic approach for single-channel noise reduction. The desired signal magnitude is estimated by applying a linear filter to a modified version of the observations’ vector. The modified version is constructed from a Kronecker product of the observations’ vector with its complex conjugate. The estimated signal magnitude is multiplied by a complex exponential whose phase is obtained using a conventional linear filtering approach. We focus on the linear and quadratic maximum signal-to-noise ratio (SNR) filters and demonstrate that the quadratic filter is superior in terms of subband SNR gains. In addition, in the context of speech enhancement, we show that the quadratic filter is ideally preferable in terms of perceptual evaluation of speech quality (PESQ) and short-time objective intelligibility (STOI) scores. The advantages, compared to the conventional linear filtering approach, are particularly significant for low input SNRs, at the expanse of a higher computational complexity. The results are verified in practical scenarios with nonstationary noise and in comparison to well-known speech enhancement methods. We demonstrate that the quadratic maximum SNR filter may be superior, depending on the nonstationary noise type.

A speech-based computational auditory signal processing and perception model.

A new speech intelligibility prediction model is presented which is based on the Computational Auditory Signal Processing and Perception model (CASP) of Jepsen, Ewert, and Dau [(2008). J. Acoust. Soc. Am. 124(1), 422-438]. The model combines a non-linear auditory-inspired preprocessing with a backend based on the cross-correlation between the clean and the degraded speech representations in the modulation envelope domain. Several speech degradation and speech enhancement algorithms were considered to study the ability of the model to predict data from normal-hearing listeners. Degradations of speech intelligibility due to additive noise, phase-jitter distortion, and single-channel noise reduction as well as improved speech intelligibility due to ideal binary mask processing are shown to be successfully accounted for by the model. Furthermore, the model reflects stimulus-level dependent effects of auditory perception, including audibility limitations at low levels and degraded speech intelligibility at high levels. Given its realistic non-linear auditory processing frontend, the speech-based computational auditory signal processing and perception model may provide a valuable computational framework for studying the effects of sensorineural hearing impairment on speech intelligibility.

Predicting the impact of hearing aid processing on speech intelligibility

Speech intelligibility is usually assessed using subjective tests, but several objective measures of speech intelligibility have also been proposed. This study compares intelligibility of IEEE sentences in collocated speech-shaped noise, recorded through a low-cost amplifier and three current-issue NHS hearing aids with single-channel noise reduction settings switched on and off. Results from twenty-one normal hearing listeners indicates that single-channel noise reduction algorithms can significantly improve intelligibility of noisy speech (p < 0.05), but no differences can be seen between different hearing aid models currently available on the NHS. The low-cost, off-the shelf hearing amplifier performed significantly worse than the NHS prescribed devices. Three objective speech intelligibility measures were applied to the recordings. None of the objective metrics was found to accurately predict the outcomes seen in this study for all hearing aid conditions, though general trends can be predicted (r = 0.83 overall, but for any discrete signal-to-noise ratio across all conditions, maximum r = 0.64). Machine-learning approaches were used to refit parameters from the best performing of these metrics. Although performance was improved, problems in predicting particular conditions remain, suggesting that objective measures do not encompass all relevant factors. [Action on Hearing Loss (England/Wales Registered Charity Number 207720).]Speech intelligibility is usually assessed using subjective tests, but several objective measures of speech intelligibility have also been proposed. This study compares intelligibility of IEEE sentences in collocated speech-shaped noise, recorded through a low-cost amplifier and three current-issue NHS hearing aids with single-channel noise reduction settings switched on and off. Results from twenty-one normal hearing listeners indicates that single-channel noise reduction algorithms can significantly improve intelligibility of noisy speech (p < 0.05), but no differences can be seen between different hearing aid models currently available on the NHS. The low-cost, off-the shelf hearing amplifier performed significantly worse than the NHS prescribed devices. Three objective speech intelligibility measures were applied to the recordings. None of the objective metrics was found to accurately predict the outcomes seen in this study for all hearing aid conditions, though general trends can be predicted (r = 0....

Beamforming based on null-steering with small spacing linear microphone arrays.

This paper develops an approach to beamforming with small spacing uniform linear microphone arrays based on the null-steering (NS) principle. It first formulates the beamforming problem from the conventional mean-squared error (MSE) criterion and its normalized version. Several NS algorithms are then derived for beamforming with the constraint of placing nulls to either a single direction or multiple angles. The difference and relationships between different algorithms are discussed and their performances are evaluated. These algorithms can be used to design either fixed or adaptive beamformers. In the former case, the resulting beamformers work as differential microphone arrays (DMAs) since they exhibit frequency-invariant beampatterns and their directivity factors are high with a given number of sensors. In the latter case, the resulting beamformers can be viewed as a combination of DMAs and single-channel noise reduction since they do not only exhibit frequency-invariant beampatterns but also can achieve noise reduction based on the noise statistics.

Hearing aid noise suppression and working memory function

Objective: Research findings concerning the relation between benefit from hearing aid (HA) noise suppression and working memory function are inconsistent. The current study thus investigated the effects of three noise suppression algorithms on auditory working memory and the relation with reading span. Design: Using a computer simulation of bilaterally fitted HAs, four settings were tested: (1) unprocessed, (2) directional microphones, (3) single-channel noise reduction, and (4) binaural coherence-based noise reduction. Settings 2–4 were matched in terms of the speech-weighted signal-to-noise ratio (SNR) improvement. Auditory working memory was assessed at +6 dB SNR using listening span and N-back paradigms. Study sample: Twenty experienced HA users aged 55–80 years with large differences in reading span. Results: For the listening span measurements, there was an influence of HA setting on sentence-final word recognition and recall, with the directional microphones leading to ∼6% better performance than the single-channel noise reduction. For the N-back measurements, there was substantial test-retest variability and no influence of HA setting. No interactions with reading span were found. Conclusion: HA noise suppression may affect the recognition and recall of speech at positive SNRs, irrespective of individual reading span. Future work should improve the reliability of the auditory working memory measurements.

Robust Speech-Distortion Weighted Interframe Wiener Filters for Single-Channel Noise Reduction

In this paper, speech-distortion weighted (SDW) interframe Wiener filters (IFWFs) are investigated for single-channel noise reduction in a filter bank structure. The filters utilize a parameter $\mu$ that explicitly sets a tradeoff between noise reduction and speech distortion and have traditionally been used in multichannel applications under the term SDW multichannel Wiener filter. The application of these SDW-IFWFs relies on the estimation of interframe correlation (IFC) coefficients, and it is shown that the IFC coefficients can be more robustly estimated using a secondary higher resolution filter bank (HRFB). It is then shown how real-valued scalar gains, which are optimal in the primary filter bank, can be applied directly in the HRFB instead of the interframe filtering in the primary filter bank, which leads to a more robust noise reduction performance for any value of $\mu$ . Computing these gains is also cheaper since matrix inversions are avoided and the primary filter bank is not needed in the actual implementation. Experimental results are given that support the claims, where the proposed methods are compared to relevant reference methods using measures such as the segmental SNR and the objective PESQ.

On single-channel noise reduction with rank-deficient noise correlation matrix

The widely studied subspace and linear filtering methods for noise reduction require the noise correlation matrix to be invertible. In certain application scenarios, however, this matrix is either rank deficient or very ill conditioned, so this requirement cannot be fulfilled. In this paper, we investigate possible solutions to this important problem based on subspace techniques for single-channel time-domain noise reduction. The eigenvalue decomposition is applied to both the speech and noise correlation matrices to separate the null and nonnull subspaces. Then, a set of optimal and suboptimal filters are derived from the nullspace of the noise signal. Through simulations, we observe that the proposed filters are able to significantly reduce noise without introducing much distortion to the desired signal. In comparison with the conventional Wiener approach, the developed filters perform significantly better in improving both the signal-to-noise ratio (SNR) and the perceptual evaluation of speech quality (PESQ) score when the noise correlation matrix is rank deficient.

Evaluation of combined dynamic compression and single channel noise reduction for hearing aid applications

Objective: Single-channel noise reduction (SCNR) and dynamic range compression (DRC) are important elements in hearing aids. Only relatively few studies have addressed interaction effects and typically used real hearing aids with limited knowledge about the integrated algorithms. Here the potential benefit of different combinations and integration of SCNR and DRC was systematically assessed. Design: Ten different systems combining SCNR and DRC were implemented, including five serial arrangements, a parallel and two multiplicative approaches. In an instrumental evaluation, signal-to-noise ratio (SNR) improvement and spectral contrast enhancement (SCE) were assessed. Quality ratings at 0 and +6 dB SNR, and speech reception thresholds (SRTs) in noise were measured using stationary and babble noise. Study sample: Thirteen young normal-hearing (NH) listeners and 12 hearing-impaired (HI) listeners participated. Results: In line with an increased segmental SNR and spectral contrast compared to a serial concatenation, the parallel approach significantly reduced the perceived noise annoyance for both subject groups. The proposed multiplicative approaches could partly counteract increased speech distortions introduced by DRC and achieved the best overall quality for the HI listeners. Conclusions: For high SNRs well above the individual SRT, the specific combination of SCNR and DRC is perceptually relevant and the integrative approaches were preferred.

Single channel noise reduction system in low SNR

We propose a two stage noise reduction system for reducing background noise using single-microphone recordings in very low signal-to-noise ratio (SNR) based on Wiener filtering and ideal binary masking. The proposed system contains two stages. In first stage, the Wiener filtering with improved a priori SNR is applied to noisy speech for background noise reduction. In second stage, the ideal binary mask is estimated at every time---frequency channel by using pre-processed first stage speech and comparing the time---frequency channels against a pre-selected threshold T to reduce the residual noise. The time---frequency channels satisfying the threshold are preserved whereas all other time---frequency channels are attenuated. The results revealed substantial improvements in speech intelligibility and quality over that accomplished with the traditional noise reduction algorithms and unprocessed speech.

A Frequency-Domain Adaptive Line Enhancer With Step-Size Control Based on Mutual Information for Harmonic Noise Reduction

We propose an adaptive line enhancer with a frequency-dependent step-size. The proposed frequency-domain adaptive line enhancer is used as a single-channel noise reduction system for removing harmonic noise from noisy speech. Our main contribution is to exploit the temporal dependence in the log-magnitude and phase spectra of the noisy speech using mutual information, and to derive a frequency-dependent step-size which detects the presence of harmonic noise in different frequency bins. Our proposed step-size control allows the suppression of harmonic noise and the preservation of speech components. The experiments are performed with different real-life acoustic noises which contain harmonic components. Using instrumental speech intelligibility and quality measures, we demonstrate that the proposed approach can outperform the conventional frequency-domain adaptive line enhancer with a fixed step-size for harmonic noise reduction.

The Voice Track multiband single-channel modified Wiener-filter noise reduction system for cochlear implants: patients' outcomes and subjective appraisal

Objective: The aim of this study was to evaluate the potential improvements of speech perception and sound quality provided by a multiband single channel noise-reduction algorithm based on the modified Wiener-filter adapted to cochlear implant sound processing. Design: This study was a longitudinal trial with a repeated-measures design. Outcome measures were performed on the first day when the noise reduction feature was provided and after a one month habituation period. Objective measures included pure-tone thresholds and vocal audiometry assessments. Speech perception was measured in quiet and in the presence of two types of noise: a stationary speech shaped noise and a two-talker cocktail noise. Subjective sound quality was assessed using a ten item questionnaire. Study Sample: Thirteen post-lingual deaf adults, experienced users of a cochlear implant system, took part in this study. Results: The noise-reduction algorithm provided a benefit for the perception of speech presented in a stationary speech shaped noise and an overall improvement in subjective sound quality ratings. Conclusions: It was shown that a single channel noise reduction system based on a modified Wiener-filter approach can improve speech in noise perception performance and subjective sound quality in cochlear implant patients.

Comprehensive evaluation of binaural hearing aid pre-processing strategies—Speech intelligibility in realistic noise scenarios

For the development of new signal processing approaches, e.g., for hearing aids, a final stage of evaluation is crucial. However, the evaluation procedures for assessing performance and benefit often represent a vast and not unified realm. Thus, we suggest a comprehensive evaluation to describe the efficacy, i.e., the anticipated real-world benefit as precise as possible. This comprises physical, instrumental, and perceptual (human) measurements of objective measures, as well as subjective attributes. Eight signal pre-processing strategies were evaluated following this concept, including directional microphones, coherence filters, single-channel noise reduction, binaural beamformers, and their combinations. Speech reception thresholds (SRTs) were measured with normal-hearing and hearing-impaired listeners in three realistic noise scenarios and compared with predictions of common instrumental measures. Although hearing-impaired listeners required a better signal-to-noise ratio to obtain 50% intelligibility...

Single-channel noise reduction via semi-orthogonal transformations and reduced-rank filtering

This paper investigates the problem of single-channel noise reduction in the time domain. The objective is to find a lower dimensional filter that can yield a noise reduction performance as close as possible to or even better than that obtained by the full-rank solution. This is achieved in three steps. First, we transform the observation signal vector sequence, through a semi-orthogonal matrix, into a sequence of transformed signal vectors with a reduced dimension. Second, a reduced-rank filter is applied to get an estimate of the clean speech in the transformed domain. Third, the estimate of the clean speech in the time domain is obtained by an inverse semi-orthogonal transformation. The focus of this paper is on the derivation of semi-orthogonal transformations under certain estimation criteria in the first step and the design of the reduced-rank optimal filters that can be used in the second step. We show how noise reduction using the principle of rank reduction can be cast as an optimal filtering problem, and how different semi-orthogonal transformations affect the noise reduction performance. Simulations are performed under various conditions to validate the deduced filters for noise reduction.

Speech quality evaluation of a sparse coding shrinkage noise reduction algorithm with normal hearing and hearing impaired listeners

Although there are numerous papers describing single-channel noise reduction strategies to improve speech perception in a noisy environment, few studies have comprehensively evaluated the effects of noise reduction algorithms on speech quality for hearing impaired (HI). A model-based sparse coding shrinkage (SCS) algorithm has been developed, and has shown previously (Sang et al., 2014) that it is as competitive as a state-of-the-art Wiener filter approach in speech intelligibility. Here, the analysis is extended to include subjective quality ratings and a method called Interpolated Paired Comparison Rating (IPCR) is adopted to quantitatively link the benefit of speech intelligibility and speech quality.The subjective quality tests are performed through IPCR to efficiently quantify noise reduction effects on speech quality. Objective measures including frequency-weighted segmental signal-to-noise ratio (fwsegSNR), perceptual evaluation of speech quality (PESQ) and hearing aid speech quality index (HASQI) are adopted to predict the noise reduction effects.Results show little difference in speech quality between the SCS and the Wiener filter algorithm but a difference in quality rating between the HI and NH listeners. HI listeners generally gave better quality ratings of noise reduction algorithms than NH listeners. However, SCS reduced the noise more efficiently at the cost of higher distortions that were detected by NH but not by the HI.SCS is a promising candidate for noise reduction algorithms for HI. In general, care needs to be taken when adopting algorithms that were originally developed for NH participants into hearing aid applications. An algorithm that is evaluated negatively with NH might still bring benefits for HI participants.

Study of the maximum signal-to-noise-ratio filters for single- and multi-channel noise reduction

Noise reduction is a problem of recovering a speech signal of interest from its noisy observations. Since the objective of the problem is to reduce noise, thereby improving the signal-to-noise ratio (SNR), it is natural to consider the use of maximum SNR filters. However, the maximum SNR filters, if not designed properly, may introduce significant speech distortion, leading to speech quality degradation instead of improvement. This paper investigates the design of maximum SNR filters for noise reduction with minimum speech distortion. It covers the following design cases: (1) single-channel noise reduction in the time domain; (2) multichannel noise reduction in the time domain; (3) single-channel noise reduction in the short-time-Fourier-transform (STFT) domain with or without interframe information; (4) multichannel noise reduction in the STFT domain with or without interframe information. A large number of experiments are performed to illustrate the properties of these maximum SNR filters.

Two-Stage Filter-Bank System for Improved Single-Channel Noise Reduction in Hearing Aids

The filter-bank system implemented in hearing aids has to fulfill various constraints such as low latency and high stop-band attenuation, usually at the cost of low frequency resolution. In the context of frequency-domain noise-reduction algorithms, insufficient frequency resolution may lead to annoying residual noise artifacts since the spectral harmonics of the speech cannot properly be resolved. Especially in case of female speech signals, the noise between the spectral harmonics causes a distinct roughness of the processed signals. Therefore, this work proposes a two-stage filter-bank system, such that the frequency resolution can be improved for the purpose of noise reduction, while the original first-stage hearing-aid filter-bank system can still be used for compression and amplification. We also propose methods to implement the second filter-bank stage with little additional algorithmic delay. Furthermore, the computational complexity is an important design criterion. This finally leads to an application of the second filter-bank stage to lower frequency bands only, resulting in the ability to resolve the harmonics of speech. The paper presents a systematic description of the second filter-bank stage, discusses its influence on the processed signals in detail and further presents the results of a listening test which indicates the improved performance compared to the original single-stage filter-bank system.