Speech Synthesizers have traditionally been built on carefully read speech that is recorded in studio environment. Such voices are suboptimal for use in noisy conditions, which is inevitable in a majority of deployed speech systems. In this work, we attempt to modify the output of the speech synthesizers to make it more appropriate for noisy environments. Comparison of spectral and prosodic fea...

We introduce a framework for speech enhancement based on convolutive non-negative matrix factorization that leverages available speech data to enhance arbitrary noisy utterances with no a priori knowledge of the speakers or noise types present. Previous approaches have shown the utility of a sparse reconstruction of the speech-only components of an observed noisy utterance. We demonstrate that ...

In-Car speech recognition will be pervasive over the coming years. The goal of speech enhancement is to increase the quality and intelligibility of speech in a noisy environment. The focus of the present research is to evaluate the effect of speech enhancement on the intelligibility of spoken language in a moving vehicle. Here, an ECoS network is used as a model to evaluate the intelligibility....

Recognition performance decreases when automatic recognition systems are used over the telephone network, especially wireless network and noisy environments. Previous studies have shown that non efficient speech/non-speech detection is a very important source of this degradation. Hence, speech detector robustness to noise is highly required, in order to improve recognition performance for the v...

We describe SRI’s recognition system as used in the 2001 DARPA Speech in Noisy Environments (SPINE) evaluation. The SPINE task involves recognition of speech in simulated military environments. The task had some unique challenges, including segmentation of foreground speech from noisy background, the need for robust acoustic models to handle noisy speech, and development of language models from...

Objectives: The objective of this project is to develop a preprocessor for speech recognition systems operating in noisy environments. The preprocessor, consisting of a nonlinear inhomogeneous transmission line, will be realized in software, although realization in hardware in FYgl should be possible. More specifically we will: 1) Develop a nonlinear transmission line preprocessor that accurate...

In adverse conditions, the speech recognition performance decreases in part due to imperfect speech/non-speech detection. In this paper, a new combination of voicing parameter and energy for speech/non-speech detection is described. This combination avoids especially the noise detections in real life very noisy environments and provides better performance for continuous speech recognition. This...

In this work an exemplar-based technique for speech enhancement of noisy speech is proposed. The technique works by finding a sparse representation of the noisy speech in a dictionary containing both speech and noise exemplars, and uses the activated dictionary atoms to create a time-varying filter to enhance the noisy speech. The speech enhancement algorithm is evaluated using measured signal ...

Research has shown that degrading acoustic background noise innuences speech quality across phoneme classes in a non-uniform manner. This results in variable quality performance of many speech enhancement algorithms in noisy environments. A phoneme classiication procedure is proposed which directs single-channel constrained speech enhancement. The procedure performs broad phoneme class partitio...

This paper addresses the study of the speech intelligibility enhancement. The speech model, noise sources, perceptual aspects of speech, and performance evaluation are reviewed. The intelligibility enhancement system based on spectral subtraction technique is investigated. Spectral density estimation device based on the algorithm of smoothed periodograms is analysed. Determination of the silenc...