Document Sample

ADAPTIVE WIENER FILTERING APPROACH FOR SPEECH ENHANCEMENT M. A. Abd El-Fattah*, M. I. Dessouky , S. M. Diab and F. E. Abd El-samie # Department of Electronics and Electrical communications, Faculty of Electronic Engineering Menoufia University, Menouf, Egypt E-mails: * maro_zizo2010@yahoo.com , # fathi_sayed@yahoo.com ABSTRACT This paper proposes the application of the Wiener filter in an adaptive manner in speech enhancement. The proposed adaptive Wiener filter depends on the adaptation of the filter transfer function from sample to sample based on the speech signal statistics(mean and variance). The adaptive Wiener filter is implemented in time domain rather than in frequency domain to accommodate for the varying nature of the speech signal. The proposed method is compared to the traditional Wiener filter and spectral subtraction methods and the results reveal its superiority. Keywords: Speech Enhancement, Spectral Subtraction, Adaptive Wiener Filter 1 INTRODUCTION Speech enhancement is one of the most is removed first. Decomposition of the vector space important topics in speech signal processing. of the noisy signal is performed by applying an Several techniques have been proposed for this eigenvalue or singular value decomposition or by purpose like the spectral subtraction approach, the applying the Karhunen-Loeve transform (KLT)[8]. signal subspace approach, adaptive noise canceling Mi. et. al. have proposed the signal / noise KLT and the iterative Wiener filter[1-5] . The based approach for colored noise removal[9]. The performances of these techniques depend on idea of this approach is that noisy speech frames quality and intelligibility of the processed speech are classified into speech-dominated frames and signal. The improvement of the speech signal-to- noise-dominated frames. In the speech-dominated noise ratio (SNR) is the target of most techniques. frames, the signal KLT matrix is used and in the noise-dominated frames, the noise KLT matrix is Spectral subtraction is the earliest method for used. enhancing speech degraded by additive noise[1]. In this paper, we present a new technique to This technique estimates the spectrum of the clean improve the signal-to-noise ratio in the enhanced (noise-free) signal by the subtraction of the speech signal by using an adaptive implementation estimated noise magnitude spectrum from the noisy of the Wiener filter. This implementation is signal magnitude spectrum while keeping the phase performed in time domain to accommodate for the spectrum of the noisy signal. The drawback of this varying nature of the signal. technique is the residual noise. The paper is organized as follows: in section Another technique is a signal subspace II, a review of the spectral subtraction technique is approach [3]. It is used for enhancing a speech presented. In section III, the traditional Wiener signal degraded by uncorrelated additive noise or filter in frequency domain is revisited. Section IV, colored noise [6,7]. The idea of this algorithm is proposes the adaptive Wiener filtering approach for based on the fact that the vector space of the noisy speech enhancement. In section V, a comparative signal can be decomposed into a signal plus noise study between the proposed adaptive Wiener filter, subspace and an orthogonal noise subspace. the Wiener filter in frequency domain and the Processing is performed on the vectors in the signal spectral subtraction approach is presented. plus noise subspace only, while the noise subspace Ubiquitous Computing and Communication Journal 1 2 SPECTRAL SUBTRACTION A noise-free signal estimate can then be obtained Spectral subtraction can be categorized as a with the inverse Fourier transform. This noise non-parametric approach, which simply needs an reduction method is a specific case of the general estimate of the noise spectrum. It is assume that technique given by Weiss, et al. and extended by there is an estimate of the noise spectrum that is Berouti , et al.[2,12]. typically estimated during periods of speaker The spectral subtraction approach can be silence. Let x(n) be a noisy speech signal : viewed as a filtering operation where high SNR regions of the measured spectrum are attenuated x ( n) = s ( n) + v ( n) (1) less than low SNR regions. This formulation can be given in terms of the SNR defined as: where s(n) is the clean (the noise-free) signal, and 2 v(n) is the white gaussian noise. Assume that the X (ω ) noise and the clean signals are uncorrelated. By SNR = (5) ˆ Pv (ω ) applying the spectral subtraction approach that estimates the short term magnitude spectrum of the Thus, equation (3) can be rewritten as: noise-free signal S (ω ) by subtraction of the 2 2 ˆ ˆ ˆ S (ω ) = X (ω ) − Pv (ω ) estimated noise magnitude spectrum V (ω ) from −1 (6) the noisy signal magnitude spectrum X (ω ) . It is 2⎡ 1 ⎤ ≈ X (ω ) 1 + sufficient to use the noisy signal phase spectrum as ⎢ ⎣ SNR ⎥ ⎦ an estimate of the clean speech phase spectrum,[10]: An important property of noise suppression using spectral subtraction is that the attenuation ˆ ˆ S(ω) = ( X (ω) − N(ω) ) exp(j∠X (ω)) (2) characteristics change with the length of the analysis window. A common problem for using The estimated time-domain speech signal is spectral subtraction is the musicality that results obtained as the inverse Fourier transform of from the rapid coming and going of waves over ˆ successive frames [13]. S (ω ) . Another way to recover a clean signal s(n) 3 WIENER FILTER IN FREQUNCY from the noisy signal x(n) using the spectral DOMAIN subtraction approach is performed by assuming that there is an the estimate of the power spectrum The Wiener filter is a popular technique that of the noise Pv (ω ) , that is obtained by averaging has been used in many signal enhancement over multiple frames of a known noise segment. methods. The basic principle of the Wiener filter is An estimate of the clean signal short-time squared to obtain a clean signal from that corrupted by magnitude spectrum can be obtained as follow [8]: additive noise. It is required estimate an optimal filter for the noisy input speech by minimizing the ⎧X(ω) 2 −Pv(ω), if X(ω) 2 −Pv(ω) ≥0 Mean Square Error (MSE) between the desired ˆ ˆ 2 ⎪ ˆ signal s(n) and the estimated signal s ( n) . The ˆ S(ω) = ⎨ (3) frequency domain solution to this optimization ⎪ 0, otherwise problem is given by[13]: ⎩ Psω) ( H(ω) = (7) It is possible combine this magnitude spectrum Psω) + Pvω) ( ( estimate with the measured phase and then get the Short Time Fourier Transform (STFT) estimate as where Ps (ω ) and Pv (ω ) are the power spectral follows: densities of the clean and the noise signals, respectively. This formula can be derived ˆ ˆ j∠X (ω ) considering the signal s and the noise signal v as S (ω ) = S (ω ) e (4) Ubiquitous Computing and Communication Journal 2 uncorrelated and stationary signals. The signal-to- Pv (ω ) =σv2 noise ratio is defined by[13]: (10) Ps (ω ) Consider a small segment of the speech SNR = (8) signal in which the signal x(n) is assumed to be ˆ Pv (ω ) stationary, The signal x(n) can be modeled by: This definition can be incorporated to the Wiener x(n) = mx + σxw(n) (11) filter equation as follows: where mx and σx are the local mean and standard deviation of x(n). w(n) is a unit variance noise. −1 ⎡ 1 ⎤ Within this small segment of speech, the H ( ω ) = ⎢1 + (9) SNR ⎥ Wiener filter transfer function can be approximated ⎣ ⎦ by: The drawback of the Wiener filter is the fixed Ps (ω ) σs 2 frequency response at all frequencies and the H (ω ) = = Ps (ω ) + Pv (ω ) σs + σv 2 2 requirement to estimate the power spectral density of the clean signal and noise prior to filtering. (12) From Eq.(12), because H (ω ) is constant over the 4 THE PROPOSED ADAPTIVE WIENER small segment of speech, the impulse response of FILTER the Wiener filter can be obtained by: This section presents and adaptive implementation of the Wiener filter which benefits σs 2 from the varying local statistics of the speech h( n) = δ ( n) (13) signal. A block diagram of the proposed approach σs + σv 2 2 is illustrated in Fig. (1). In this approach, the estimated speech signal mean mx and variance ˆ From Eq.(13), the enhanced speech s ( n) within this local segment can be expressed as: σx 2 are exploited. σs 2 A priori knowledge s (n) = mx + ( x(n) - mx ) ∗ ˆ δ ( n) σs + σv 2 2 Space- σs 2 Degraded speech variant Enhanced x(n) speech = mx + ( x(n) − mx ) σs + σv 2 2 h(n) ˆ signal s ( n) (14) If it is assumed that mx and σs are updated at each sample, we can say: Measure of σs (n) ( x(n) − mx(n)) 2 Local speech s (n) = mx (n) + ˆ σs (n) + σv 2 2 A priori statistics knowledge (15) In Eq.(15), the local mean mx(n) and Figure 1: Typical adaptive speech enhancement system for additive noise reduction ( x(n) − mx (n)) are modified separately from segment to segment and then the results are combined. If σs is much larger than σv the 2 2 It is assumed that the additive noise v(n) is ˆ output signal s ( n) is assumed to be primarily due of zero mean and has a white nature with variance to x(n) and the input signal x(n) is not attenuated. If of σv .Thus, the power spectrum Pv (ω ) can be 2 σs is smaller than σv , the filtering effect is 2 2 approximated by: performed. Ubiquitous Computing and Communication Journal 3 Notice that mx is identical to ms when In the first experiment , all the above- mentioned algorithms are carried out on the Handle mv is zero. So, we can estimate mx (n) in Eq.(15) signal with different SNRs and the output PSNR from x(n) by: results are shown in Fig. (2). The same experiment n+M is repeated for the Laughter and Gong signals and 1 ms (n) = mx (n) = ˆ ˆ ∑ x(k ) the results are shown in Figs.(3) and (4), respectively. (2 M +1) k =n−M From these figures, it is clear that the proposed (16) adaptive Wiener filter approach has the best performance for different SNRs. The adaptive where ( 2 M + 1) is the number of samples in the Wiener filter approach gives about 3-5 dB improvement at different values of SNR. The non- short segment used in the estimation. linearity between input SNR and output PSNR is due to the adaptive nature of the filter. To measure the local signal statistics in the system of Figure 1, the algorithm developed uses the signal variance σs . The specific method 2 used to designing the space-variant h(n) is given by (17.b). Since σx = σs + σv may be estimated 2 2 2 80 from x(n) by: 70 ⎧σx (n) − σv , if σx (n) > σv ˆ2 ˆ2 ˆ2 ˆ2 σs (n) = ⎨ ˆ 2 60 ⎩0, otherwise O u tp u t P S N R (d B ) 50 (17.a) Where 40 30 n+ M 1 σx (n) = ˆ 2 (2 M + 1) ∑ ( x(k ) − m (n)) k =n−M ˆ x 2 20 Spectral Subtraction (17.b) 10 Wiener Filter By this proposed method, we guarantee that Adaptive Wiener Filter the filter transfer function is adapted from sample to sample based on the speech signal statistics. 0 -10 -5 0 5 10 15 20 25 30 35 Input SNR (dB) 5 EXPERIMENTAL RESULTS For evaluation purposes, we use different Figure 2: PSNR results for white noise speech signals like the handel, laughter and gong case at-10 dB to +35 dB SNR levels for Handle signal signals. White Gaussian noise is added to each speech signal with different SNRs. The different speech enhancement algorithms such as the spectral subtraction method, the Weiner filter in frequency domain and the proposed adaptive Wiener filter are carried out on the noisy speech signals. The peak signal to noise ratio (PSNR) results for each enhancement algorithm are compared. Ubiquitous Computing and Communication Journal 4 reveal that the best performance is that of the 60 proposed adaptive Wiener filter. 1 1 50 A m p lit u d e A m p lit u d e 0 0 40 -1 -1 O u tp u t P S N R (d B ) 0 2000 4000 6000 8000 0 2000 4000 6000 8000 (a) (b) 30 1 1 A m p lit u d e A m p lit u d e 0 0 20 -1 -1 0 2000 4000 6000 8000 0 2000 4000 6000 8000 Spectral Subtraction (c) (d) 10 Wiener Filter 1 A m p lit u d e Adaptive Wiener Filter 0 0 -10 -5 0 5 10 15 20 25 30 35 -1 0 2000 4000 6000 8000 Input SNR (dB) (e) Time(msec) Figure 3: PSNR results for white noise case at -10 dB to +35 dB SNR levels for Laughter signal Figure 5: Time domain results of the Handel sig. at SNR = +5dB (a) original sig. (b) noisy sig. (c) spectral subtraction. (d) Wiener filtering. (e) adaptive Wiener filtering. 80 70 Amplitude (dB) Amplitude (dB) 0 0 60 -20 -20 -40 O u tp u t P S N R (d B ) 50 -40 00 00 00 00 0 10 20 30 40 00 00 00 00 0 10 20 30 40 (a) (b) 40 Amplitude (dB) Amplitude (dB) 0 0 30 -20 -20 20 Spectral Subtraction -40 -40 0 10 20 30 40 00 00 00 00 00 00 00 00 0 10 20 30 40 Wiener Filter (c) (d) 10 Adaptive Wiener Filter Amplitude (dB) 0 0 -10 -5 0 5 10 15 20 25 30 35 -20 Input SNR (dB) -40 00 00 00 00 0 10 20 30 40 Figure 4: PSNR results for white noise case at -10 dB ) re .(H (e F q z) to +35 dB SNR levels for Gong signal The results of the different enhancement Figure 6:The spectrum of the Handel sig. in Fig.(5) (a) original sig. (b) noisy sig. (c) spectral subtraction. (d) algorithms for the handle signal with SNRs of 5, Wiener filtering. (e) adaptive Wiener filtering. 10,15 and 20 dB in the both time and frequency domain are given in Figs. (5) to (12). These results Ubiquitous Computing and Communication Journal 5 1 1 A m p lit u d e A m p lit u d e 1 1 0 0 A m p lit u d e A m p lit u d e 0 0 -1 -1 0 2000 4000 6000 8000 0 2000 4000 6000 8000 -1 -1 (a) (b) 0 2000 4000 6000 8000 0 2000 4000 6000 8000 (a) (b) 1 1 A m p lit u d e A m p lit u d e 1 1 0 0 A m p lit u d e A m p lit u d e 0 0 -1 -1 0 2000 4000 6000 8000 0 2000 4000 6000 8000 -1 -1 (c) (d) 0 2000 4000 6000 8000 0 2000 4000 6000 8000 (c) (d) 1 1 A m p lit u d e A m p lit u d e 0 0 -1 -1 0 2000 4000 6000 8000 0 2000 4000 6000 8000 (e) Time (msec) (e) Time(msec) Figure 7: Time domain results of the Handel sig. at SNR = 10 dB (a) original sig. (b) noisy sig. (c) spectral Figure 9: Time domain results of the Handel sig. at subtraction. (d) Wiener filtering. (e) adaptive Wiener SNR = 15 dB (a) original sig. (b) noisy sig. (c) spectral filtering. subtraction. (d) Wiener filtering. (e) adaptive Wiener filtering. Amplitude (dB) Amplitude (dB) Amplitude (dB) Amplitude (dB) Amplitude (dB) Amplitude (dB) 0 0 0 Amplitude (dB) Amplitude (dB) 0 -20 -20 -20 -20 -40 -40 -40 -40 0 0 0 0 00 00 0 1 0 20 30 4 0 0 0 0 0 00 00 0 10 20 30 40 00 00 00 00 0 10 20 30 40 00 00 00 00 0 10 20 30 40 (a) (b) (a) (b) 0 0 0 0 -20 -20 -20 -20 -40 -40 0 0 0 0 00 00 0 10 2 0 30 4 0 -40 -40 00 0 0 0 0 00 0 10 2 0 30 4 0 00 00 00 00 0 10 20 30 40 00 00 00 00 0 10 20 30 40 (c) (d) (c) (d) Amplitude (dB) 0 Amplitude (dB) 0 -20 -20 -40 -40 00 0 0 00 0 0 0 10 2 0 3 0 40 0 00 00 00 00 10 20 30 40 ) re. z (e F q (H) ) re. z (e F q (H) Figure 8: The spectrum of the Handel sig. in Fig.(7) Figure 10: The spectrum of the Handel sig. in Fig.(9) (a) original sig. (b) noisy sig. (c) spectral subtraction. (d) (a) original sig. (b) noisy sig. (c) spectral subtraction. (d) Wiener filtering. (e) adaptive Wiener filtering. Wiener filtering. (e) adaptive Wiener filtering. Ubiquitous Computing and Communication Journal 6 6 CONCLUSION 1 1 A m p lit u d e A m p lit u d e An adaptive Wiener filter approach for 0 0 speech enhancement is proposed in this papaper. This approach depends on the adaptation of the -1 -1 0 2000 4000 6000 8000 0 2000 4000 6000 8000 filter transfer function from sample to sample (a) (b) based on the speech signal statistics(mean and variance). This results indicates that the proposed 1 1 approach provides the best SNR improvement A m p lit u d e A m p lit u d e 0 0 among the spectral subtraction approach and the traditional Wiener filter approach in frequency -1 -1 domain. The results also indicate that the proposed 0 2000 4000 6000 8000 0 2000 4000 6000 8000 approach can treat musical noise better than the (c) (d) spectral subtraction approach and it can avoid the 1 drawbacks of Wiener filter in frequency domain . A m p lit u d e 0 REFERENCES -1 [1] S. F. Boll: Suppression of acoustic noise in 0 2000 4000 6000 8000 speech using spectral subtraction, IEEE Trans. (e) Time(msec) Acoust., Speech, Signal Processing, vol. ASSP-27,. pp. 113-120 (1979). Figure 11: Time domain results of the Handel sig. at [2] M. Berouti, R. Schwartz, and J. Makhoul: SNR = 20 dB (a) original sig. (b) noisy sig. (c) spectral Enhancement of speech corrupted by acoustic subtraction. (d) Wiener filtering. (e) adaptive Wiener noise, Proc. IEEE Int. Conf. Acoust., Speech filtering. Signal Processing, pp. 208-211 (1979). Amplitude (dB) Amplitude (dB) 0 [3] Y. Ephriam and H. L. Van Trees: A signal 0 subspace approach for speech enhancement, in -20 -20 Proc. International Conference on Acoustic, -40 Speech and Signal Processing, vol. II, Detroit, -40 0 10 20 3 0 40 00 00 00 00 0 0 00 00 00 0 1 0 20 30 40 MI, U.S.A., pp. 355-358, May (1993). (a) (b) [4] Simon Haykin: Adaptive Filter Theory, Prentice-Hall, ISBN 0-13-322760-X, (1996). Amplitude (dB) Amplitude (dB) 0 0 [5] J. S. Lim and A. V. Oppenheim.: All-pole -20 -20 Modelling of Degraded Speech, IEEE Trans. Acoust., Speech, Signal Processing, ASSP-26, -40 -40 June (1978). 0 10 20 3 0 40 00 00 0 0 00 0 0 00 00 00 0 1 0 20 30 40 [6] Y. Ephraim and H. L. Van Trees, A spectrally- (c) (d) based signal subspace approach for speech Amplitude (dB) 0 enhancement, in IEEE ICASSP, pp. 804-807 (1995). -20 [7] Y. Hu and P. Loizou: A subspace approach -40 for enhancing speech corrupted by colored noise, 0 00 00 00 00 10 20 30 40 in Proc. International Conference on ) re. z (e F q (H) Acoustics, Speech and Signal Processing, vol. I, Orlando, FL, U.S.A., pp. 573-576, May (2002). Figure 12: The spectrum of the Handel sig. in Fig.(11) [8] A. Rezayee and S. Gazor: An adaptive KLT (a) original sig. (b) noisy sig. (c) spectral subtraction. (d) approach for speech enhancement, IEEE Trans. Wiener filtering. (e) adaptive Wiener filtering. Speech Audio Processing, vol. 9, pp. 87-95 Feb. (2001). [9] U. Mittal and N. Phamdo: Signal/noise KLT based approach for enhancing speech degraded by colored noise, IEEE Trans. Speech Audio Processing, vol. 8, NO. 2, pp. 159-167,(2000). [10] John R. Deller, John G. Proakis, and John H. L. Hansen. Discrete- Time Processing of Speech Ubiquitous Computing and Communication Journal 7 Signals. Prentice-Hall, ISBN 0-02-328301-7 (1997). [11] S. F. Boll: Suppression of Acoustic Noise in Speech Using Spectral Sub- traction. IEEE Trans. Acoustics, Speech, and Signal Processing. vol. ASSP-29. no. 2, pp. 113-120, April (1979). [12] M. R. Weiss, E. Aschkenasy, and T. W. Parsons: Processing Speech Signal to Attenuate Interference, in Proc. IEEE Symp. Speech Recognition, pp. 292-293, April (1974). [13] J. S. Lim and A. V. Oppenheim: Enhancement and band width compression of Noisy speech, Proc. of the IEEE, vol. 67, No..12, pp. 1586-1604, Dec. (1979). Ubiquitous Computing and Communication Journal 8

DOCUMENT INFO

Shared By:

Categories:

Tags:
Ubiquitous Computing and Communication Journal, www.ubicc.org, networks, e-Learning, UBICC Journal, UBICC, Journal, Research, Ubiquitous, Computing, Communication, Ubiquitous Computing and Communication Journal, 6Lowpan, Security, Privacy, Mobile Computing, Mobile, Notebook, Computer, Research, electronics, electronic, electrical, conferences, UBICC conferences, UBICC membership, biomed, conference, UBICC fellow, nanotechnology, UBICC xplore, fellow, UBICC pes, UBICC explore

Stats:

views: | 32 |

posted: | 1/3/2011 |

language: | English |

pages: | 8 |

Description:
UBICC, the Ubiquitous Computing and Communication Journal [ISSN 1992-8424], is an international scientific and educational organization dedicated to advancing the arts, sciences, and applications of information technology. With a world-wide membership, UBICC is a leading resource for computing professionals and students working in the various fields of Information Technology, and for interpreting the impact of information technology on society.
www.ubicc.org

OTHER DOCS BY tabindah

How are you planning on using Docstoc?
BUSINESS
PERSONAL

By registering with docstoc.com you agree to our
privacy policy and
terms of service, and to receive content and offer notifications.

Docstoc is the premier online destination to start and grow small businesses. It hosts the best quality and widest selection of professional documents (over 20 million) and resources including expert videos, articles and productivity tools to make every small business better.

Search or Browse for any specific document or resource you need for your business. Or explore our curated resources for Starting a Business, Growing a Business or for Professional Development.

Feel free to Contact Us with any questions you might have.