Soft Mask Methods for Single-Channel Speaker Separation

Reddy, Aarthi M.; Raj, Bhiksha

doi:10.1109/tasl.2007.901310

Cited by 102 publications

(102 citation statements)

References 16 publications

Supporting

Mentioning

102

Contrasting

Order By: Relevance

“…It is worth mentioning here that the estimation algorithm presented in this section is similar in some aspects to other algorithms proposed in the literature for feature compensation [18,19,33], model decomposition [43,47] and single-channel speaker separation [40,41]. Nevertheless, contrary to previous work, the problem we address here is that of speech feature enhancement for noise-robust ASR under the assumption that the corrupting source (noise) is distributed according to a GMM.…”

Section: Spectral Reconstruction Using the Masking Modelmentioning

confidence: 99%

Spectral Reconstruction and Noise Model Estimation Based on a Masking Model for Noise Robust Speech Recognition

González

Gómez²,

Peinado³

et al. 2017

Circuits Syst Signal Process

View full text Add to dashboard Cite

An effective way to increase noise robustness in automatic speech recognition (ASR) systems is feature enhancement based on an analytical distortion model that describes the effects of noise on the speech features. One of such distortion models that has been reported to achieve a good tradeoff between accuracy and simplicity is the masking model. Under this model, speech distortion caused by environmental noise is seen as a spectral mask and, as a result, noisy speech features can be either reliable (speech is not masked by noise) or unreliable (speech is masked). In this paper we present a detailed overview of this model and its applications to noise-robust ASR. Firstly, using the masking model, we derive a spectral reconstruction technique aimed at enhancing the noisy speech features. Two problems must be solved in order to perform spectral reconstruction using the masking model: i) mask estimation, i.e. determining the reliability of the noisy features, and ii) feature imputation, i.e. estimating speech for the unreliable features. Unlike missing-data imputation techniques where the two problems are considered as independent, our technique jointly addresses them by exploiting a priori knowledge of the speech and noise sources in the form of a statistical model. Secondly, we propose an algorithm for estimating the noise model required by the feature enhancement technique. The proposed algorithm fits a Gaussian mixture model (GMM) to the noise by iteratively maximising the likelihood of the noisy speech signal so that noise can be estimated even during speech-dominating frames. A comprehensive set of experiments carried out on the Aurora-2 and Aurora-4 databases shows that the proposed method achieves significant improvements over the baseline system and other similar missing-data imputation techniques.

show abstract

Section: Spectral Reconstruction Using the Masking Modelmentioning

confidence: 99%

Spectral Reconstruction and Noise Model Estimation Based on a Masking Model for Noise Robust Speech Recognition

González

Gómez²,

Peinado³

et al. 2017

Circuits Syst Signal Process

View full text Add to dashboard Cite

show abstract

“…a time-varying Wiener filter, or a more extreme binary time-frequency mask) to recover an estimate of the original target source. Related approaches have been investigated by several other researchers, including [3], who derive soft masks from the posterior probabilities of each cell belonging to a particular source, [4], who learn separate but coupled models for multiple frequency subbands, and [5] who infer distributions over the target speech magnitudes.…”

Section: Introductionmentioning

confidence: 99%

Model-Based Monaural Source Separation Using a Vector-Quantized Phase-Vocoder Representation

Ellis

Weiss

2006 IEEE International Conference on Acoustics Speed and Signal Processing Proceedings

View full text Add to dashboard Cite

A vector quantizer (VQ) trained on short-time frames of a particular source can form an accurate non-parametric model of that source. This principle has been used in several previous source separation and enhancement schemes as a basis for filtering the original mixture. In this paper, we propose the "projection" of a corrupted target signal onto the constrained space represented by the model as a viable model for source separation. We investigate some parameters of VQ encoding, including a more perceptuallymotivated distance measure, and an encoding of phase derivatives that supports reconstruction directly from quantizer output alone. For the problem of separating speech from noise, we highlight some problems with this approach, including the need for sequential constraints (which we introduce with a simple hidden Markov model), and choices for choosing the best quantization for overlapping sources.

show abstract

“…This is in general an ill-posed problem and can not be solved without further knowledge about the sources or their interrelationship. Possible SCSS methods are mainly divided into two categories: source driven [4][5][6][7] and model-based methods [12][13][14][15][16][17][18]. As a major example for the first group, computational auditory scene analysis (CASA) has widely been studied [4].…”

Section: Introductionmentioning

confidence: 99%

“…This speaker dependent model is used as source prior knowledge and is applied for separation without considering the interfering component. The most prominent models are vector quantization (VQ) [12], [16], Gaussian mixture models (GMM) [14], [15] and Hidden Marcov models (HMM) [17]. In most recent proposed model-based SCSS techniques, it is assumed that the test speech files are recorded at a condition similar to that of the training phase recording.…”

Section: Introductionmentioning

confidence: 99%

A Bayesian Approach for Single Channel Speech Separation

Kammi¹,

Karami²

2012

IJMLC

View full text Add to dashboard Cite

Abstract-This paper addresses the problem of single channel speech separation to extract and enhance the desired speech signals from mixed speech signals. We propose a new speech separation algorithm by utilizing Bayesian approach for the case in which, underlying sources are mixed at different levels of energies. This situation is not considered in many single channel speech separation methods. To validate the effectiveness of our proposed method, it is compared with a state-of-the art method which is a gain adapted Maximum Likelihood estimator. Through the experiments, we show that our proposed method outperforms the compared method.Index Terms-Single channel speech separation, bayesian approach, gain estimation.

show abstract

Soft Mask Methods for Single-Channel Speaker Separation

Cited by 102 publications

References 16 publications

Spectral Reconstruction and Noise Model Estimation Based on a Masking Model for Noise Robust Speech Recognition

Spectral Reconstruction and Noise Model Estimation Based on a Masking Model for Noise Robust Speech Recognition

Model-Based Monaural Source Separation Using a Vector-Quantized Phase-Vocoder Representation

A Bayesian Approach for Single Channel Speech Separation

Contact Info

Product

Resources

About