In the missing data approach to robust Automatic Speech Recognition (ASR), time-frequency regions which carry reliable speech information are identified. Recognition is then based on these regions alone. In this paper, we address the problem of identifying reliable regions and propose two criteria to solve this based on negative energy ( $ s < 0 ) and SNR ( $ s s n 2 2 2 < + ). These criteria a...

Spectral subtraction (SS) approach has been widely used for speech enhancement and recognition accuracy improvement, but becomes less effective when the additive noise is not white. In this paper, we propose to integrate wavelet transform and the SS algorithm. The spectrum of the additive noise in each frequency band obtained in this way can then be better approximated as white if the number of...

In the field of speech signal processing, Spectral subtraction method (SSM) has been successfully implemented to suppress the noise that is added acoustically. SSM does reduce the noise at satisfactory level but musical noise is a major drawback of this method. To implement spectral subtraction method, transformation of speech signal from time domain to frequency domain is required. On the othe...

S-transform (ST) is an effective time-frequency representation method with the advantages of short-time Fourier transform and wavelet transform. This paper utilizes the advantages of the power spectral subtraction method and the speech enhancement method to presents a novel speech enhancement algorithm based on discrete ST. Firstly, the speech in time domain is transformed by ST to joint time-f...

We propose a noise robust speech recognition method based on combining novel features extracted from fundamental frequency (F0) information and spectral subtraction. F0 features have been shown to be effective in speech recognition in noisy environments. Recently, F0 features obtained by Hough transform were developed for concatenated digit recognition and significantly improved recognition per...

In this study, we develop a new method to realize speech intelligibility prediction of synthetic sounds processed by nonlinear speech enhancement algorithms. A speech envelope power spectrum model (sEPSM) was proposed to account for subjective results on a spectral subtraction, but it is untested by recent state-of-the-art speech enhancement algorithms. We introduce a dynamic compressive gammac...

In order to meet the requirements of robustness and discrimination of content preserving operations after conversion of speech communication format on the heterogeneous mobile terminal, and noise reduction and efficient authentication, a new efficient speech perceptual hashing authentication algorithm based on spectral subtraction and energy to entropy ratio was proposed. Firstly the proposed a...

In this paper a HTK-compatible robust speech parametrization tool CtuCopy is presented. This tool allows for the usage of several additive noise suppression preprocessing techniques, nonlinear spectrum transformation, RASTA-like filtration, and direct final feature computation. The tool is general, it is easily extendible, and it may be also used for speech enhancement purposes. In the second p...

In order to make the speech perception hashing authentication algorithm has strong robustness and discrimination to content preserving operations and speech communication under the common background noise, a new robust speech perceptual hashing authentication algorithm based on spectral subtraction and multi-feature tensor was proposed. The proposed algorithm uses spectral subtraction method to...

A new approach for single channel speech enhancement based on Kalman filtering and masking properties of the human auditory system is proposed in the paper. A standard time-varying Kalman filtering method is extended by combining the calculation of noise masking thresholds during the process of parameter updating. Simulation results of a traditional spectral subtraction method, an extended spec...