We previously have applied deep autoencoder (DAE) for noise reduction and speech enhancement. However, the DAE was trained using only clean speech. In this study, by using noisyclean training pairs, we further introduce a denoising process in learning the DAE. In training the DAE, we still adopt greedy layer-wised pretraining plus fine tuning strategy. In pretraining, each layer is trained as a...

In this paper, we present a new speech enhancement method based on robust principal component analysis. In the proposed method, noisy signal is transformed into time-frequency domain where background noise is assumed as a low-rank component and human speech is regarded as a sparse compone. An inexact augmented Lagrange multipliers algorithm is conducted for solving the noise and speech separati...

In this paper we address the problem of vector quantization of speech in a noisy environment. We show that the performance of a vector quantization system can be improved by adapting the distortion measure to the changing environmental conditions. The proposed method emphasizes the distortion in spectral regions where the speech signal dominates. The method functions well even when conventional...

In this work, a new feature extracting method in noisy environments is proposed. The approach is based on subband decomposition of speech signals followed by adaptive filtering in the noisiest subbbands of speech. The speech decomposition is obtained using low complexity octave filter bank, while adaptive filtering is performed using the normalized least mean square algorithm. The performance o...

MFCCs perform well when used for clean speech recognition. However, for noisy speech the recognition rates go down. Augmenting the MFCC feature vector by dynamic features improves both discrimination and robustness of the MFCC-based recognizer. In this paper, we present an alternative para meterization based on the frequency filtering (FF) technique. By using FF, a significant improvement with ...

The quality of text-to-speech (TTS) voices built from noisy speech is compromised. Enhancing the speech data before training has been shown to improve quality but voices built with clean speech are still preferred. In this paper we investigate two different approaches for speech enhancement to train TTS systems. In both approaches we train a recursive neural network (RNN) to map acoustic featur...

In this paper we address the problem of enhancing speech which has been degraded by additive noise. As proposed by Ephraim et al., autoregressive hidden Markov models (AR-HMM) for the clean speech and an autoregressive Gaussian for the noise are used. The filter applied to a given frame of noisy speech is estimated using the noise model and the autoregressive Gaussian having the highest a poste...

The performance of speech recognition systems degrades significantly when they are operated in noisy conditions. For example, the automatic speech recognition (ASR) frontend of a speech-to-speech (S2S) translation prototype that is currently developed at IBM [11] shows noticeable increase in its word error rate (WER) when it is operated in real field noise. Thus, adding noise robustness to spee...

Hynek Hermansky, Eric A. Wan, and Carlos Avendano Oregon Graduate Institute of Science & Technology Department of Electrical Engineering and Applied Physics P.O. Box 91000, Portland, OR 97291 ABSTRACT Finite Impulse Response (FIR) Wiener-like lters are applied to time trajectories of cubic-root compressed short-term power spectrum of noisy speech recorded over cellular telephone communications....

We demonstrate a system to integrate adaptive beam-forming and auditory features in order to improve speech recognition accuracy in noisy environments. Adaptive beam-forming based on a microphone array can utilize spatial information to improve the sound recording signal-to-noise ratio (SNR) on a focused speaker for robust speech recognition. Auditory features based on modeling the signal proce...