Musical Noise Research Articles

Spectral subtraction based on two-stage spectral estimation and modified cepstrum thresholding

This paper proposes a low-variance and adaptive-bandwidth spectral estimator for spectral subtraction, which is based on the two-stage spectral estimation (TSSE) and the modified cepstrum thresholding (MCT). In the first stage, both the raw periodogram and the noise power spectral density (NPSD) are smoothed over frequency based on the structure of the NPSD. The second stage is applied to distinguish each harmonic component for speech signals, which is based on the structure of the speech spectrum. The TSSE could provide a low-variance and adaptive-bandwidth spectral estimator for both noise and speech since the TSSE considers both the structure of the NPSD and that of the speech spectrum. Although spectral subtraction based on the TSSE (TSSE-SS) could solve the annoying musical noise problem, but the TSSE-SS could not suppress the non-stationary noise effectively, so the MCT is applied to the TSSE-SS to further reduce the non-stationary noise components. Experimental results show that the proposed algorithm has higher signal-to-noise-ratio improvement and higher PESQ scores than conventional spectral subtraction algorithms.

Optimization and evaluation of sigmoid function with a priori SNR estimate for real-time speech enhancement

In this paper, an a priori signal-to-noise ratio (SNR) estimator with a modified sigmoid gain function is proposed for real-time speech enhancement. The proposed sigmoid gain function has three parameters, which can be optimized such that they match conventional gain functions. In addition, the joint temporal dynamics between the SNR estimate and the spectral gain function is investigated to improve the performance of the speech enhancement scheme. As the widely-used decision-directed (DD) a priori SNR estimate has a well-known one-frame delay that leads to the degradation of speech quality, a modified a priori SNR estimator is proposed for the DD approach to overcome this delay. Evaluations are performed by utilizing the objective evaluation metric that measures the trade-off between the noise reduction, the speech distortion and the musical noise in the enhanced signal. The results are compared using the PESQ and the SNRseg measures as well as subjective listening tests. Simulation results show that the proposed gain function, which can flexibly model exponential distributions, is a potential alternative speech enhancement gain function.

Musical-Noise-Free Speech Enhancement Based on Optimized Iterative Spectral Subtraction

In this paper, we provide a theoretical analysis of the amount of musical noise in iterative spectral subtraction, and its optimization method for the least musical noise generation. To achieve high-quality noise reduction with low musical noise, iterative spectral subtraction, i.e., iteratively applied weak nonlinear signal processing, has been proposed. Although the effectiveness of the method has been reported experimentally, there have been no theoretical studies. Therefore, in this paper, we formulate the generation process of musical noise by tracing the change in kurtosis of noise spectra, and conduct a comparison of the amount of musical noise for different parameter settings but the same achieved level of noise attenuation. Furthermore, we theoretically derive the optimal internal parameters that generate no musical noise. It is clarified that to find a fixed point in kurtosis yields the no-musical-noise property. Comparative experiments with commonly used noise reduction methods show the proposed method's efficacy.

Stereophonic Acoustic Echo Suppression Based on Wiener Filter in the Short-Time Fourier Transform Domain

An open-loop stereophonic acoustic echo suppression (SAES) method without preprocessing is presented for teleconferencing systems, where the Wiener filter in the short-time Fourier transform (STFT) domain is employed. Instead of identifying the echo path impulse responses with adaptive filters, the proposed algorithm estimates the echo spectra from the stereo signals using two weighting functions. The spectral modification technique originally proposed for noise reduction is adopted to remove the echo from the microphone signal. Moreover, a priori signal-to-echo ratio (SER) based Wiener filter is used as the gain function to achieve a trade-off between musical noise reduction and computational load for real-time operations. Computer simulation shows the effectiveness and the robustness of the proposed method in several different scenarios.

A Two Stage Algorithm for Denoising of Speech Signal

Spectral Subtraction based speech denoising suffers the major problem of introduction of new type of noise called musical noise. In this paper we introduce a two stage algorithm for denoising of noisy speech signal by minimizing the musical noise and additive white Gaussian noise. The algorithm works in two stages. First spectral subtraction method is applied on the noisy speech signal after applying pre-emphasis filter. Since the speech is highly corrupted by white Gaussian noise (3 dB SNR), the method introduces a musical noise. The enhanced speech is the processed using Savitzky-Golay smoothing filter to reduce the musical noise. The various objective as well as subjective measures are evaluated for the two stage method and compared against the individual methods. The two stage algorithm produces enhanced speech signal having high perceptual quality and intelligibility.

A NEW SPEECH ENHANCEMENT TECHNIQUE USING PERCEPTUAL CONSTRAINED SPECTRAL WEIGHTING FACTORS

This paper deals with musical noise result from perceptual speech enhancement type algorithms and especially wiener filtering. Although perceptual speech enhancement methods perform better than the non perceptual methods, most of them still return annoying residual musical noise. This is due to the fact that if only noise above the noise masking threshold is filtered then noise below the noise masking threshold can become audible if its maskers are filtered. It can affect the performance of perceptual speech enhancement method that process audible noise only. In order to overcome this drawback here proposed a new speech enhancement technique. It aims to improve the quality of the enhanced speech signal provided by perceptual wiener filtering by controlling the latter via a second filter regarded as a psychoacoustically motivated weighting factor. The simulation results shows that the performance is improved compared to other perceptual speech enhancement methods

A convex model and L1 minimization for musical noise reduction in blind source separation

Blind source separation (BSS) methods are useful tools to recover or enhance in- dividual speech sources from their mixtures in a multi-talker environment. A class of efficient BSS methods are based on the mutual exclusion hypothesis of the source signal Fourier spectra on the time- frequency (TF) domain, and subsequent data clustering and classification. Though such methodology is simple, the discontinuous decisions in the TF domain for classification often cause defects in the recovered signals in the time domain. The defects are perceived as unpleasant ringing sounds, the so called musical noise. Post-processing is desired for further quality enhancement. In this paper, an efficient musical noise reduction method is presented based on a convex model of time-domain sparse filters. The sparse filters are intended to cancel out the interference due to major sparse peaks in the mixing coefficients or physically the early arrival and high energy portion of the room impulse responses. This strategy is efficiently carried out by l1 regularization and the split Bregman method. Evaluations by both synthetic and room recorded speech and music data show that our method outperforms existing musical noise reduction methods in terms of objective and subjective measures. Our method can be used as a post-processing tool for more general and recent versions of TF domain BSS methods as well.

Application of Perceptual Filtering Models to Noisy Speech Signals Enhancement

This paper describes a new speech enhancement approach using perceptually based noise reduction. The proposed approach is based on the application of two perceptual filtering models to noisy speech signals: the gammatone and the gammachirp filter banks with nonlinear resolution according to the equivalent rectangular bandwidth (ERB) scale. The perceptual filtering gives a number of subbands that are individually spectral weighted and modified according to two different noise suppression rules. The importance of an accurate noise estimate is related to the reduction of the musical noise artifacts in the processed speech that appears after classic subtractive process. In this context, we use continuous noise estimation algorithms. The performance of the proposed approach is evaluated on speech signals corrupted by real‐world noises. Using objective tests based on the perceptual quality PESQ score and the quality rating of signal distortion (SIG), noise distortion (BAK) and overall quality (OVRL), and subjective test based on the quality rating of automatic speech recognition (ASR), we demonstrate that our speech enhancement approach using filter banks modeling the human auditory system outperforms the conventional spectral modification algorithms to improve quality and intelligibility of the enhanced speech signal.

Theoretical Analysis of Amounts of Musical Noise and Speech Distortion in Structure-Generalized Parametric Blind Spatial Subtraction Array

SUMMARY We propose a structure-generalized blind spatial subtraction array (BSSA), and the theoretical analysis of the amounts of musical noise and speech distortion. The structure of BSSA should be selected according to the application, i.e., a channelwise BSSA is recommended for listening but a conventional BSSA is suitable for speech recognition.

Speech enhancement using a minimum mean-square error short-time spectral modulation magnitude estimator

In this paper we investigate the enhancement of speech by applying MMSE short-time spectral magnitude estimation in the modulation domain. For this purpose, the traditional analysis-modification-synthesis framework is extended to include modulation domain processing. We compensate the noisy modulation spectrum for additive noise distortion by applying the MMSE short-time spectral magnitude estimation algorithm in the modulation domain. A number of subjective experiments were conducted. Initially, we determine the parameter values that maximise the subjective quality of stimuli enhanced using the MMSE modulation magnitude estimator. Next, we compare the quality of stimuli processed by the MMSE modulation magnitude estimator to those processed using the MMSE acoustic magnitude estimator and the modulation spectral subtraction method, and show that good improvement in speech quality is achieved through use of the proposed approach. Then we evaluate the effect of including speech presence uncertainty and log-domain processing on the quality of enhanced speech, and find that this method works better with speech uncertainty. Finally we compare the quality of speech enhanced using the MMSE modulation magnitude estimator (when used with speech presence uncertainty) with that enhanced using different acoustic domain MMSE magnitude estimator formulations, and those enhanced using different modulation domain based enhancement algorithms. Results of these tests show that the MMSE modulation magnitude estimator improves the quality of processed stimuli, without introducing musical noise or spectral smearing distortion. The proposed method is shown to have better noise suppression than MMSE acoustic magnitude estimation, and improved speech quality compared to other modulation domain based enhancement methods considered.

Impact of SNR and gain-function over- and under-estimation on speech intelligibility

Most noise reduction algorithms rely on obtaining reliable estimates of the SNR of each frequency bin. For that reason, much work has been done in analyzing the behavior and performance of SNR estimation algorithms in the context of improving speech quality and reducing speech distortions (e.g., musical noise). Comparatively little work has been reported, however, regarding the analysis and investigation of the effect of errors in SNR estimation on speech intelligibility. It is not known, for instance, whether it is the errors in SNR overestimation, errors in SNR underestimation, or both that are harmful to speech intelligibility. Errors in SNR estimation produce concomitant errors in the computation of the gain (suppression) function, and the impact of gain estimation errors on speech intelligibility is unclear. The present study assesses the effect of SNR estimation errors on gain function estimation via sensitivity analysis. Intelligibility listening studies were conducted to validate the sensitivity analysis. Results indicated that speech intelligibility is severely compromised when SNR and gain over-estimation errors are introduced in spectral components with negative SNR. A theoretical upper bound on the gain function is derived that can be used to constrain the values of the gain function so as to ensure that SNR overestimation errors are minimized. Speech enhancement algorithms that can limit the values of the gain function to fall within this upper bound can improve speech intelligibility.

Musical Noise Controllable Algorithm of Channelwise Spectral Subtraction and Adaptive Beamforming Based on Higher Order Statistics

In this paper, we propose a musical-noise-controllable algorithm for array signal processing with the aim for high-performance and high-quality noise reduction. Recently, many methods of integrating linear microphone array signal processing and nonlinear signal processing for noise reduction have been studied, but these methods often suffer from the problem of musical noise. In the proposed algorithm, channelwise spectral subtraction is applied before adaptive array signal processing. We also introduce a new automatic control algorithm to obtain the subtraction strength parameter used in the spectral subtraction, which depends on the amount of generated musical noise, measured by higher order statistics. We confirm the effectiveness of the proposed algorithm via objective and subjective evaluations.

Theoretical Analysis of Musical Noise in Generalized Spectral Subtraction Based on Higher Order Statistics

In this paper, we provide a new theoretical analysis of the amount of musical noise generated via generalized spectral subtraction based on higher order statistics. Power spectral subtraction is the most commonly used spectral subtraction method, and in our previous study a musical noise assessment theory limited to the power spectral domain was proposed. In this paper, we propose a generalization of our previous theory on spectral subtraction for arbitrary exponent parameters. We can thus compare the amount of musical noise between any exponent domains from the results of our analysis. We also clarify that less musical noise is generated when we choose a lower exponent spectral domain; this implies that there is no theoretical justification for using power/amplitude spectral subtraction.

Speech Enhancement Algorithm Using Sub band Two Step Decision Directed Approach with Adaptive Weighting factor and Noise Masking Threshold

Problem statement: Speech Enhancement plays an important role in any of the speech processing systems like speech recognition, mobile communication, hearing aid. Approach: In this work, human perceptual auditory masking effect is incorporated into the single channel speech enhancement algorithm. The algorithm is based on a criterion by which the audible noise may be masked rather than being attenuated and thereby reducing the chance of distortion to speech. The basic decision directed approach is for efficient reduction of musical noise, that includes the estimation of the a priori SNR which is a crucial parameter of the spectral gain, follows the a posteriori SNR with a delay of one frame in speech frames. In this work a simple adaptive speech enhancement technique, using an adaptive sigmoid type function to determine the weighting factor of the TSDD algorithm is employed based on a sub band approach. In turn the spectral estimate is used to obtain a perceptual gain factor. Results: Objective and subjective measures like SNR, MSE, IS distance and were obtained, which shows the ability of the proposed method for efficient enhancement of noisy speech Conclusion/Recommendations: Performance assessment shows that our proposal can achieve a more significant noise reduction and a better spectral estimation of weak speech spectral components from a noisy signal as compared to the conventional speech enhancement algorithm.

Noise Reduction Using mel-Scale Spectral Subtraction with Perceptually Defined Subtraction Parameters- A New Scheme

The noise signal does not affect uniformly the speech signal over the whole spectrum isn the case of colored noise. In order to deal with speech improvement in such situations a new spectral subtraction algorithm is proposed for reducing colored noise from noise corrupted speech. The spectrum is divided into frequency sub-bands based on a nonlinear multiband bark scale. For each sub-band, the noise corrupted speech power in past and present time frames is compared to statistics of the noise power to improve the determination of the presence or absence of speech. During the subtraction process, a larger proportion of noise is removed from sub-bands that do not contain speech. For sub-bands that contain speech, a function is developed which allows for the removal of less noise during relatively low amplitude speech and more noise during relatively high amplitude speech .Further the performance of the spectral subtraction is improved by formulating process without neglecting the cross correlation between the speech signal and background noise. Residual noise can be masked by exploiting the masking properties of the human auditory system. In the proposed method subtraction parameters are adaptively adjusted using noise masking threshold. A psychoacoustically motivated weighting filter was included to eliminate residual musical noise. Experimental results show that the algorithm removes more colored noise without removing the relatively low amplitude speech at the beginning and ending of words.

Subjective and Objective Quality Assessment of Audio Source Separation

We aim to assess the perceived quality of estimated source signals in the context of audio source separation. These signals may involve one or more kinds of distortions, including distortion of the target source, interference from the other sources or musical noise artifacts. We propose a subjective test protocol to assess the perceived quality with respect to each kind of distortion and collect the scores of 20 subjects over 80 sounds. We then propose a family of objective measures aiming to predict these subjective scores based on the decomposition of the estimation error into several distortion components and on the use of the PEMO-Q perceptual salience measure to provide multiple features that are then combined. These measures increase correlation with subjective scores up to 0.5 compared to nonlinear mapping of individual state-of-the-art source separation measures. Finally, we released the data and code presented in this paper in a freely available toolkit called PEASS.

Analysis of the Decision-Directed SNR Estimator for Speech Enhancement With Respect to Low-SNR and Transient Conditions

Because of their many applications and their relative ease of implementation, single-channel speech enhancement algorithms have received much attention. As a consequence, a vast amount of publications on estimation procedures and their implementation in noise reduction systems exists. However, there has been little systematic research on the theoretic performance of such estimators. In this paper, we provide a systematic analysis of the performance of noise reduction algorithms in low signal-to-noise ratio (SNR) and transient conditions, where we consider approaches using the well-known decision-directed SNR estimator. We show that the smoothing properties of the decision-directed SNR estimator in low SNR conditions can be analytically described and that the limits of noise reduction for widely used spectral speech estimators based on the decision-directed approach can be predicted. We also illustrate that achieving both a good preservation of speech onsets in transient conditions on one side and the suppression of musical noise on the other can be especially problematic when the decision-directed SNR estimation is used.

A multistage approach to blind separation of convolutive speech mixtures

We propose a novel algorithm for the separation of convolutive speech mixtures using two-microphone recordings, based on the combination of independent component analysis (ICA) and ideal binary mask (IBM), together with a post-filtering process in the cepstral domain. The proposed algorithm consists of three steps. First, a constrained convolutive ICA algorithm is applied to separate the source signals from two-microphone recordings. In the second step, we estimate the IBM by comparing the energy of corresponding time–frequency (T–F) units from the separated sources obtained with the convolutive ICA algorithm. The last step is to reduce musical noise caused by T–F masking using cepstral smoothing. The performance of the proposed approach is evaluated using both reverberant mixtures generated using a simulated room model and real recordings in terms of signal to noise ratio measurement. The proposed algorithm offers considerably higher efficiency and improved speech quality while producing similar separation performance compared with a recent approach.

Over-Attenuated Components Regeneration for Speech Enhancement

Despite the quality improvement of the speech signal with most traditional noise reduction (TNR) algorithms, the output is always distorted to some extent due to the over-attenuation of speech components. Weak speech components are usually regarded as noise in noise reduction processing and are therefore highly suppressed. In this paper, we propose a postprocessing technique which is based on the regeneration of both the voiced and unvoiced speech in the entire frequency domain to reduce this problem. A nonlinear transform is first applied to obtain the excitation signal, and a smooth envelope is then estimated. To utilize the information of the clean speech contained in the envelope, we combine the original TNR filter output with a weighted product of the excitation signal and the estimated envelope to generate the final synthesized speech. The synthesized speech is quite close to the clean speech and is more natural-sounding. Moreover, our algorithm can mask the residual musical noise effectively with the regenerated speech components. Experimental results demonstrate the excellent performance of our algorithm. In addition, we introduce two novel objective measures and further show the efficiency of our algorithm in maintaining the clean speech while reducing the noise as much as possible.

Reduction of musical residual noise using perceptual tools with classic speech denoising techniques

Single channel enhancement techniques based on short-time spectral amplitude (STSA) estimation have the major drawback of generating an artificial and annoying residual noise with musical character, due mainly to the unwanted peaks in the denoised signal spectrum. The detection and reduction of spectral peaks which have a musical characteristic are the main objectives of this paper. The proposed perceptual technique to reduce musical residual noise operates as a post-processing. Based on human auditory properties, the perceptual post-processing is established in a number of steps. First, we detect musical peaks by comparing tonality coefficients in each critical band of both denoised signal and reference signal. Detected musical peaks are audible only if they exceed the clean speech masking threshold (MT). However, the clean MT is not available. It is estimated by modifying the Johnston model. Secondly, we reduce the musical residual noise by removing only audible musical peaks which exceed the estimated MT. The proposed method is tested and compared with classic STSA technique and perceptual techniques at various levels of white and colored noise. Results show the validity of the proposed technique.