L.I. Averina1, D.S. Kuznetsov2, D.Yu. Charkin3

1-3 JSC «Concern «Sozvezdie» (Voronezh, Russia)

Single-channel noise reduction systems have an important role in many applications, either as stand-alone systems, or as a post-processing step for multi-microphone methods. There are a large number of theoretical works that offer various approaches for solving the noise reduction problem of broadband signals. However, they usually consider a limited type of interfering signals with given statistical properties; they also do not analyze the aspects of practical applicability and feasibility of algorithms.

Purpose of this work is to carry out a comparative analysis of single-channel noise reduction algorithms for broadband real signals based on the processing of amplitude and phase spectrum; to propose efficient, reliable and practical algorithmic approaches that increase the noise resistance of a single-channel system in the presence of broadband interference; to evaluate their effectiveness in suppressing various types of interference and software complexity using a low-level implementation on microcontrollers.

Spectral noise reduction algorithms of broadband real signals using their statistical data on the receiving side are considered. Using simulation modeling, a comparative analysis of the effectiveness of their functioning under various interference conditions was carried out. At the same time, as a special case, the effectiveness of the considered approaches in relation to acoustic signals and noise is analyzed. Based on low-level practical implementation of the methods, the speed of their work was estimated.

The considered theoretical algorithms are practically implemented in voice communication systems to enhance the target speech. Advantage of the method based on the processing of the amplitude spectrum of the signal has been established, both in terms of the effectiveness of combating various types of broadband interference, and in terms of the computational complexity.

Averina L.I., Kuznetsov D.S., Charkin D.Yu. Comparative analysis of single-channel spectral noise reduction algorithms for broadband signals. Radiotekhnika. 2023. V. 87. № 5. P. 157−165. DOI: https://doi.org/10.18127/j00338486-202305-16 (In Russian)

1. Dendrinos M., Bakamides S., Carayannis G. Speech enhancement from noise: A regenerative approach. ELSEVIER Speech Commun. Feb. 1991. V. 10. № 2. P. 45–57.
2. Gazor S., Zhang W. Speech enhancement employing Laplacian-Gaussian mixture. IEEE Trans. Speech Audio Process. Sep. 2005. V. 13, № 5, P. 896–904.
3. Gülzow T., Engelsberg A., Heute U. Comparison of a discrete wavelet transformation and a nonuniform polyphase filterbank applied to spectral-subtraction speech enhancement. ELSEVIER Signal Process. Jan. 1998. V. 64. № 1. P. 5–19.
4. Ephraim Y. Statistical-model-based speech enhancement systems. Proc. IEEE. Oct. 1992. V. 80. № 10. P. 1526–1555.
5. Hendriks R.C., Gerkmann T., Jensen J. DFT-domain based single-microphone noise reduction for speech enhancement - a survey of the state of the art. Morgan & Claypool Publishers. 2013.
6. Erkelens J.S., Hendriks R.C, Heusdens R., Jensen J. Minimum mean-square error estimation of discrete Fourier coefficients with generalized Gamma priors. IEEE Trans. Audio, Speech, Language Process. Aug. 2007. V. 15, № 6, P. 1741–1752.
7. Gerkmann T. and Hendriks R.C. Unbiased MMSE-based noise power estimation with low complexity and low tracking delay. IEEETrans. Audio, Speech, Language Process. May 2012. V. 20. № 4. P. 1383–1393.
8. Ephraim Y., Malah D. Speech enhancement using a minimum mean-square error shorttime spectral amplitude estimator. IEEE Trans. Acoust., Speech, Signal Process. Dec. 1984. V. 32. № 6. P. 1109–1121.
9. Wójcicki K., Milacic M., Stark A., Lyons J., Paliwal K. Exploiting conjugate symmetry of the short-time Fourier spectrum for speech enhancement. IEEE Signal Process. Lett. 2008. V. 15, P. 461–464.
10. Stark A.P., Wojcicki K.K., Lyons J.G., Paliwal K.K. Noise driven short time phase spectrum compensation procedure for speech enhancement. Proc. INTERSPEECH 2008. Brisbane, Australia. Sep. 2008. P. 549-552.
11. Rix A., Beerends J., Hollier M., Hekstra A. Perceptual Evaluation of Speech Quality (PESQ) - A new method for speech quality assessment of telephone networks and codecs. In Proc. IEEE Int. Conf. Acoustics, Speech, and Signal Processing (ICASSP’01). Salt Lake City. UT. 2001. V. 2. P. 749–752.