Short-time Spectral Amplitude Research Articles

Auditory-Based Spectral Amplitude Estimators for Speech Enhancement

We propose a new family of Bayesian estimators for speech enhancement where the cost function includes both a power law and a weighting factor. The parameters of the cost function, and therefore of the corresponding estimator gain, are chosen based on characteristics of the human auditory system, namely, the compressive nonlinearities of the cochlea, the perceived loudness and the ear's masking properties. It is found that choosing the parameters in this way results in a decrease of the estimator gain at high frequencies. This frequency dependence of the gain improves the noise reduction while limiting the speech distortion. Experimental results show that the new estimators achieve better enhancement performance than existing Bayesian estimators such as those based on the minimum mean-square error (MMSE) of the short-time spectral amplitude (STSA), the MMSE of the logarithm of the STSA (LSA) or the weighted euclidien (WE) error, both in terms of objective and subjective measures.

Speech spectral amplitude estimators using optimally shaped Gamma and Chi priors

In this paper, four STFT based speech enhancement algorithms are proposed. The algorithms enhance speech by estimating its short time spectral amplitude and are combinations of two estimators (MMSE and MAP) with two speech spectral amplitude priors (Gamma and Chi). The proposed priors have a shape parameter a, whose effect on the quality of speech is a focal point of our investigation. Rather than using a priori estimated values of a, we seek those values that maximise the quality of the enhanced speech, in an a posteriori fashion. The performance of the algorithms is first evaluated as a function of the shape parameter a and optimal values are then sought by means of a formal subjective listening test. Finally, the parallel examination of four speech enhancement algorithms offers an insight into the relative importance of the employed priors and estimators, as the proposed algorithms are only different with respect to these two elements.

Silence–breathing–snore classification from snore-related sounds

Obstructive sleep apnea (OSA) is a highly prevalent disease in which upper airways are collapsed during sleep, leading to serious consequences. Snoring is the earliest symptom of OSA, but its potential in clinical diagnosis is not fully recognized yet. The first task in the automatic analysis of snore-related sounds (SRS) is to segment the SRS data as accurately as possible into three main classes: snoring (voiced non-silence), breathing (unvoiced non-silence) and silence. SRS data are generally contaminated with background noise. In this paper, we present classification performance of a new segmentation algorithm based on pattern recognition. We considered four features derived from SRS to classify samples of SRS into three classes. The features—number of zero crossings, energy of the signal, normalized autocorrelation coefficient at 1 ms delay and the first predictor coefficient of linear predictive coding (LPC) analysis—in combination were able to achieve a classification accuracy of 90.74% in classifying a set of test data. We also investigated the performance of the algorithm when three commonly used noise reduction (NR) techniques in speech processing—amplitude spectral subtraction (ASS), power spectral subtraction (PSS) and short time spectral amplitude (STSA) estimation—are used for noise reduction. We found that noise reduction together with a proper choice of features could improve the classification accuracy to 96.78%, making the automated analysis a possibility.

Speech enhancement based on undecimated wavelet packet-perceptual filterbanks and MMSE–STSA estimation in various noise environments

In this paper, we proposed a new speech enhancement system, which integrates a perceptual filterbank and minimum mean square error–short time spectral amplitude (MMSE–STSA) estimation, modified according to speech presence uncertainty. The perceptual filterbank was designed by adjusting undecimated wavelet packet decomposition (UWPD) tree, according to critical bands of psycho-acoustic model of human auditory system. The MMSE–STSA estimation (modified according to speech presence uncertainty) was used for estimation of speech in undecimated wavelet packet domain. The perceptual filterbank provides a good auditory representation (sufficient frequency resolution), good perceptual quality of speech and low computational load. The MMSE–STSA estimator is based on a priori SNR estimation. A priori SNR estimation, which is a key parameter in MMSE–STSA estimator, was performed by using “decision directed method.” The “decision directed method” provides a trade off between noise reduction and signal distortion when correctly tuned. The experiments were conducted for various noise types. The results of proposed method were compared with those of other popular methods, Wiener estimation and MMSE–log spectral amplitude (MMSE–LSA) estimation in frequency domain. To test the performance of the proposed speech enhancement system, three objective quality measurement tests (SNR, segSNR and Itakura–Saito distance (ISd)) were conducted for various noise types and SNRs. Experimental results and objective quality measurement test results proved the performance of proposed speech enhancement system. The proposed speech enhancement system provided sufficient noise reduction and good intelligibility and perceptual quality, without causing considerable signal distortion and musical background noise.

Speech enhancement based on perceptually motivated bayesian estimators of the magnitude spectrum

The traditional minimum mean-square error (MMSE) estimator of the short-time spectral amplitude is based on the minimization of the Bayesian squared-error cost function. The squared-error cost function, however, is not subjectively meaningful in that it does not necessarily produce estimators that emphasize spectral peak (formants) information or estimators which take into account auditory masking effects. To overcome the shortcomings of the MMSE estimator, we propose in this paper Bayesian estimators of the short-time spectral magnitude of speech based on perceptually motivated cost functions. In particular, we use variants of speech distortion measures, such as the Itakura-Saito and weighted likelihood-ratio distortion measures, which have been used successfully in speech recognition. Three classes of Bayesian estimators of the speech magnitude spectrum are derived. The first class of estimators emphasizes spectral peak information, the second class uses a weighted-Euclidean cost function that implicitly takes into account auditory masking effects, and the third class of estimators is designed to penalize spectral attenuation. Of the three classes of Bayesian estimators, the estimators that implicitly take into account auditory masking effect performed the best in terms of having less residual noise and better speech quality.

Speech enhancement based on minimum mean-square error estimation and supergaussian priors

This paper presents a class of minimum mean-square error (MMSE) estimators for enhancing short-time spectral coefficients of a noisy speech signal. In contrast to most of the presently used methods, we do not assume that the spectral coefficients of the noise or of the clean speech signal obey a (complex) Gaussian probability density. We derive analytical solutions to the problem of estimating discrete Fourier transform (DFT) coefficients in the MMSE sense when the prior probability density function of the clean speech DFT coefficients can be modeled by a complex Laplace or by a complex bilateral Gamma density. The probability density function of the noise DFT coefficients may be modeled either by a complex Gaussian or by a complex Laplacian density. Compared to algorithms based on the Gaussian assumption, such as the Wiener filter or the Ephraim and Malah (1984) MMSE short-time spectral amplitude estimator, the estimators based on these supergaussian densities deliver an improved signal-to-noise ratio.

Modified Wiener filtering

In this paper, a new Wiener filtering-based method for speech enhancement is described. Standard Wiener filtering formulation requires an iterative estimation of the clean speech spectrum. However, the proposed algorithm is non-iterative, therefore the computation is much faster. In addition, it employs a time-varying noise suppression factor which is based on the frame-by-frame SNR. This feature gives us the ability to suppress those parts of the degraded signal, where speech is not likely to be present and not to suppress, and hence not to distort the speech segments as much. The algorithm is tested under simulated and actual car noise conditions and is shown to perform substantially better than the well-known spectral subtraction method in both subjective and objective speech quality evaluations. Proposed method also outperformed well-known minimum mean-square error (MMSE) short-time spectral amplitude estimator technique in terms of subjective quality. In addition, the proposed method is shown to improve the robustness of a speech recognition system significantly.

/spl beta/-order MMSE spectral amplitude estimation for speech enhancement

This paper proposes /spl beta/-order minimum mean-square error (MMSE) speech enhancement approach for estimating the short time spectral amplitude (STSA) of a speech signal. We analyze the characteristics of the /spl beta/-order STSA MMSE estimator and the relation between the value of /spl beta/ and the spectral amplitude gain function of the MMSE method. We further investigate the effectiveness of a range of fixed-/spl beta/ values in estimating STSA based on the MMSE criterion, and discuss how the /spl beta/ value could be adapted using the frame signal-to-noise ratio (SNR). The performance of the proposed speech enhancement approach is then evaluated through spectrogram inspection, objective speech distortion measures and subjective listening tests using several types of noise sources from the NOISEX-92 database. Evaluation results show that our approach can achieve a more significant noise reduction and a better spectral estimation of weak speech spectral components from a noisy signal as compared to many existing speech enhancement algorithms.

Noise reduction with less processing delay for analysis and synthesis using a single microphone

To realize comfortable voice communication with reduced steady ambient noise such as air conditioning noise, there is a form of noise suppression processing based on short-time spectral amplitude estimation. This is a frequency domain process using one microphone. Since the amount of processing is rather small, this is an extremely useful method for realization of equipment and applications. However, the processing delay associated with frame processing in the frequency domain causes deterioration of the voice communications quality. To deal with this problem, the present research proposes a scheme to reduce the processing delay. In the proposed system, deterioration of the processing quality is prevented by retaining the frequency resolution and the processing delay is reduced by compressing the shift width of the processing frame by performing overlap-addition. Objective evaluation confirms that the processing delay can be reduced to one-eighth while retaining a noise suppression capability similar to that in the conventional method. A subjective evaluation experiment confirms that voice quality similar to that in the usual method is attained. © 2005 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 89(2): 24–32, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20168

A Noise Reduction Preprocessor for Mobile Voice Communication

We describe a speech enhancement algorithm which leads to significant quality and intelligibility improvements when used as a preprocessor to a low bit rate speech coder. This algorithm was developed in conjunction with the mixed excitation linear prediction (MELP) coder which, by itself, is highly susceptible to environmental noise. The paper presents novel as well as known speech and noise estimation techniques and combines them into a highly effective speech enhancement system. The algorithm is based on short-time spectral amplitude estimation, soft-decision gain modification, tracking of the a priori probability of speech absence, and minimum statistics noise power estimation. Special emphasis is placed on enhancing the performance of the preprocessor in nonstationary noise environments.

Speech Enhancement Based onWavelet Thresholding the Multitaper Spectrum

It is well known that the "musical noise" encountered in most frequency domain speech enhancement algorithms is partially due to the large variance estimates of the spectra. To address this issue, we propose in this paper the use of low-variance spectral estimators based on wavelet thresholding the multitaper spectra for speech enhancement. A short-time spectral amplitude estimator is derived which incorporates the wavelet-thresholded multitaper spectra. Listening tests showed that the use of multitaper spectrum estimation combined with wavelet thresholding suppressed the musical noise and yielded better quality than the subspace and MMSE algorithms.

Multichannel Direction-Independent Speech Enhancement Using Spectral Amplitude Estimation

This paper introduces two short-time spectral amplitude estimators for speech enhancement with multiple microphones. Based on joint Gaussian models of speech and noise Fourier coefficients, the clean speech amplitudes are estimated with respect to the MMSE or the MAP criterion. The estimators outperform single microphone minimum mean square amplitude estimators when the speech components are highly correlated and the noise components are sufficiently uncorrelated. Whereas the first MMSE estimator also requires knowledge of the direction of arrival, the second MAP estimator performs a direction-independent noise reduction. The estimators are generalizations of the well-known single channel MMSE estimator derived by Ephraim and Malah (1984) and the MAP estimator derived by Wolfe and Godsill (2001), respectively.

Speech enhancement based on audible noise spectrum and short-time spectral amplitude estimator

Subtractive-type algorithms are effective in reducing background noise, but always produce musical noise. An algorithm based on an audible noise spectrum is proposed and tested using various noise types and levels. The performance evaluation shows that the residual noise is significantly removed and it renders better quality than those without masking.

A parametric formulation of the generalized spectral subtraction method

In this paper, two short-time spectral amplitude estimators of the speech signal are derived based on a parametric formulation of the original generalized spectral subtraction method. The objective is to improve the noise suppression performance of the original method while maintaining its computational simplicity. The proposed parametric formulation describes the original method and several of its modifications. Based on the formulation, the speech spectral amplitude estimator is derived and optimized by minimizing the mean-square error (MSE) of the speech spectrum. With a constraint imposed on the parameters inherent in the formulation, a second estimator is also derived and optimized. The two estimators are different from those derived in most modified spectral subtraction methods, which are predominantly nonstatistical. When tested under stationary white Gaussian noise and semistationary Jeep noise, they showed improved noise suppression results.

Speech enhancement based on audible noise suppression

A novel speech enhancement technique is presented based on the definition of the psychoacoustically derived quantity of audible noise spectrum and its subsequent suppression using optimal nonlinear filtering of the short-time spectral amplitude (STSA) envelope. The filter operates with sparse spectral estimates obtained from the STSA, and, when these parameters are accurately known, significant intelligibility gains, up to 40%, result in the processed speech signal. These parameters can be also estimated from noisy data, resulting into smaller but significant intelligibility gains.

Enhancement of noisy speech signals: Application to mobile radio communications

This paper deals with the enhancement of noisy speech signals recorded in a car for mobile radio applications. Our concern is the signal estimation when 1 or 2 observations are available; each one is composed of a speech signal, s i , and an additive noise n i ( i = 1, 2). In the case of one microphone, we consider new techniques that capitalize on the aspect of speech perception by focusing on enhancing only the short-time spectral amplitude. A “modified spectral subtraction” method is proposed; it uses a frequency-dependent over-estimate of the noise and can be combined with segmentation of the observation. When two microphones are available, in the case of uncorrelated (or slightly correlated) noises, we introduce a new technique based on the coherence function which is used to filter the observations or to determine a speech/noise classification algorithm. Finally, listening tests have been conducted to compare the simplest methods. In the case of stationary noise, the modified spectral subtraction is very promising and in the case of non-stationary and decorrelated noises, the method based on the coherence is more attractive.

Soustraction des bruits de fond basée sur l'estimation du spectre à court terme. Réalisation en temps réel

In order to restore degraded audio recordings in real time for French National Library, we developed a parallel structure system with PC-based multiprocessor Motorola DSP96002. The study is for some methods of speech enhancement using short-time spectral amplitude estimation. The rapidity and flexibility of our system programme for WINDOWS give us the possibility to realise many applications in sound analysis and synthesis with great performance

A short-time Fourier transform

This paper describes a method for computing the short-time Fourier transform of a discrete-time signal by convolving it with one of a family of infinite-duration windows h( nT)=( nT) k e − αnT , k integer. The values of k and α may be chosen to specify the order and bandwidth, respectively, of each equivalent analysing filter. This flexibility in controlling the response of individual analysing filter channels greatly alleviates the problem of spectral leakage which is inherent in the more traditional DFT method. The merit of the technique for computing the short-time spectral amplitude of human speech is illustrated.

Parametric coding of speech spectra

We suggest that a class of speech coders can be designed from criteria that are perception-specific. Such a class represents a middle ground between ’’waveform’’ coders and speech-specific ’’source’’ coders. We argue that, within perceptually acceptable limits, a bandpass portion of a speech signal can be represented by its short-time spectral amplitude ‖ Sn ‖ and its phase derivative ?n. Both parameters are evaluated at the center frequency ωn of the passband whose spectral width is Wn rad/s. We show that the short-time spectral amplitude is identical to the Hilbert envelope An of the passband waveform. As a by-product, we show that the bandwidth occupied by both An2 and An2 ?n is the same as that for their parent bandpass signal, but their spectra are translated to the range 0 to Wn. No comparable statements can be adduced for An and ?n, but on practical grounds, we conjecture that useful low-pass limits of less than Wn/2 can be applied to both. Finally, we outline techniques to utilize the passband parameters An, ?n, An2 and An2 ?n in digital voice coding.