This paper focuses on optimal estimators of the magnitude spectrum for speech enhancement. We present an analytical solution for estimating in the MMSE sense the magnitude spectrum when the clean speech DFT coefficients are modeled by a Laplacian distribution and the noise DFT coefficients are modeled by a Gaussian distribution. Furthermore, we derive the MMSE estimator under speech presence un...

In this paper we investigate an alternate, RI-modulation (R = real, I = imaginary) AMS framework for speech enhancement, in which the real and imaginary parts of the modulation signal are processed in secondary AMS procedures. This framework offers theoretical advantages over the previously proposed modulation AMS frameworks in that noise is additive in the modulation signal and noisy acoustic ...

Speech enhancement employing Deep Neural Networks (DNNs) is gaining strength as a data-driven alternative to classical Minimum Mean Square Error (MMSE) enhancement approaches. In the past, Observation Uncertainty approaches to integrate MMSE speech enhancement with Automatic Speech Recognition (ASR) have yielded good results as a lightweight alternative for robust ASR. In this paper we thus exp...

The theory of social presence is perhaps the most popular construct used to describe and understand how people socially interact in online learning environments. However, despite its intuitive appeal, researchers and practitioners alike often define and conceptualize this popular construct differently. In fact, it is often hard to distinguish between whether someone is talking about social inte...

To improve the performance of multi-channel speech enhancement algorithms, we previously proposed a hybrid Wiener postfilter for microphone arrays under the assumption of a diffuse noise field [4]. In this paper, considering the speech presence uncertainty, we further improve the hybrid post-filter presented before by integrating a novel robust estimator for the a priori speech absence probabil...

This paper proposes a framework for spectral enhancement of reverberant speech based on inversion of the modulation transfer function. All-pole modeling of modulation spectra of clean and degraded speech are utilized to derive the linear prediction inverse modulation transfer function (LP-IMTF) solution as a low-order IIR filter in the modulation envelope domain. By considering spectral estimat...

The term uncertainty decoding has been phrased for a class of robustness enhancing algorithms in automatic speech recognition that replace point estimates and plug-in rules by posterior densities and optimal decision rules. While uncertainty can be incorporated in the model domain, in the feature domain, or even in both, we concentrate here on feature domain approaches as they tend to be comput...

Designing a machine that is capable for understanding human speech and responds properly to speech utterance or spoken language has intrigued speech research community for centuries. Among others, one of the fundamental problems to building speech recognition system is acoustic noise. The performance of speech recognition system significantly degrades in the presence of ambient noise. Backgroun...

The paper presents a new short-time spectrum estimation algorithm for speech enhancement. A novel multivariate Laplace speech model is utilized to characterize the dependencies between adjacent DFT coefficients of speech, based on which a minimum mean-square error (MMSE) estimator of speech spectral components is derived. Moreover, the speech presence uncertainty is incorporated to modify the M...

In this paper, we present an optimally-modi#ed log-spectral amplitude (OM-LSA) speech estimator and a minima controlled recursive averaging (MCRA) noise estimation approach for robust speech enhancement. The spectral gain function, which minimizes the mean-square error of the log-spectra, is obtained as a weighted geometric mean of the hypothetical gains associated with the speech presence unce...