Corpus-Based Speech Enhancement With Uncertainty Modeling and Cepstral Smoothing

• Published in: IEEE transactions on audio, speech, and language processing Vol. 21; no. 5; pp. 983 - 997
• Main Authors: Nickel, R. M., Astudillo, R. F., Kolossa, D., Martin, R.
• Format: Journal Article
• Language: English
• Published: Piscataway, NJ IEEE 01-05-2013; Institute of Electrical and Electronics Engineers
• Subjects: Applied sciences; Background noise; Cepstral analysis; Coding, codes; Exact sciences and technology; Information, signal and communications theory; Inventory-style speech enhancement; Miscellaneous; modified imputation; Nickel; Noise; Recognition; Sampling, quantization; Signal and communications theory; Signal processing; Smoothing; Sound filters; Speech; Speech enhancement; Speech processing; Speech recognition; Telecommunications and information theory; Uncertainty; uncertainty-of-observation techniques; Performance evaluation; Mixture theory; Background noise; Modeling; Sound recognition; Learning; Audio signal; User interface; Speech enhancement; Corpus based approach; uncertainty-of-observation techniques; Phoneme; Inventory-style speech enhancement; Vector quantization; Objective analysis; Man machine dialogue; Decoding; modified imputation; Cepstral analysis; Gaussian process; Vocal signal; Quality control; Speech recognition; Speech processing; Signal to noise ratio
• ISSN: 1558-7916; 1558-7924
• DOI: 10.1109/TASL.2013.2243434

We present a new approach for corpus-based speech enhancement that significantly improves over a method published by Xiao and Nickel in 2010. Corpus-based enhancement systems do not merely filter an incoming noisy signal, but resynthesize its speech content via an inventory of pre-recorded clean signals. The goal of the procedure is to perceptually improve the sound of speech signals in background noise. The proposed new method modifies Xiao's method in four significant ways. Firstly, it employs a Gaussian mixture model (GMM) instead of a vector quantizer in the phoneme recognition front-end. Secondly, the state decoding of the recognition stage is supported with an uncertainty modeling technique. With the GMM and the uncertainty modeling it is possible to eliminate the need for noise dependent system training. Thirdly, the post-processing of the original method via sinusoidal modeling is replaced with a powerful cepstral smoothing operation. And lastly, due to the improvements of these modifications, it is possible to extend the operational bandwidth of the procedure from 4 kHz to 8 kHz. The performance of the proposed method was evaluated across different noise types and different signal-to-noise ratios. The new method was able to significantly outperform traditional methods, including the one by Xiao and Nickel, in terms of PESQ scores and other objective quality measures. Results of subjective CMOS tests over a smaller set of test samples support our claims.