| Popis: |
This paper presents a novel speech enhancement algorithm based on β-order GARCH (Generalized Auto-regressive Conditional Heteroscedasticity) model. The speech signal is modeled as β-order GARCH process, and the a priori SNR is estimated effectively. The noisy signal is divided into several critical bands, and then the value of order β is updated adaptively according to the signal-to-noise ratios in each critical band. Besides, a novel estimation method for the parameters of β-order GARCH model is proposed in this paper. The performance of the proposed algorithm is evaluated under the standard of ITU-T G.160. The experimental results show that, in comparison with the reference method, the proposed algorithm can get a greater noise reduction and larger SNR improvement, better enhanced speech quality is also ensured in different noise environments. |
