Pub Date : 2013-06-01DOI: 10.1109/TASL.2013.2248717
S. Mousazadeh, I. Cohen
Voice activity detection has attracted significant research efforts in the last two decades. Despite much progress in designing voice activity detectors, voice activity detection (VAD) in presence of transient noise is a challenging problem. In this paper, we develop a novel VAD algorithm based on spectral clustering methods. We propose a VAD technique which is a supervised learning algorithm. This algorithm divides the input signal into two separate clusters (i.e., speech presence and speech absence frames). We use labeled data in order to adjust the parameters of the kernel used in spectral clustering methods for computing the similarity matrix. The parameters obtained in the training stage together with the eigenvectors of the normalized Laplacian of the similarity matrix and Gaussian mixture model (GMM) are utilized to compute the likelihood ratio needed for voice activity detection. Simulation results demonstrate the advantage of the proposed method compared to conventional statistical model-based VAD algorithms in presence of transient noise.
{"title":"Voice Activity Detection in Presence of Transient Noise Using Spectral Clustering","authors":"S. Mousazadeh, I. Cohen","doi":"10.1109/TASL.2013.2248717","DOIUrl":"https://doi.org/10.1109/TASL.2013.2248717","url":null,"abstract":"Voice activity detection has attracted significant research efforts in the last two decades. Despite much progress in designing voice activity detectors, voice activity detection (VAD) in presence of transient noise is a challenging problem. In this paper, we develop a novel VAD algorithm based on spectral clustering methods. We propose a VAD technique which is a supervised learning algorithm. This algorithm divides the input signal into two separate clusters (i.e., speech presence and speech absence frames). We use labeled data in order to adjust the parameters of the kernel used in spectral clustering methods for computing the similarity matrix. The parameters obtained in the training stage together with the eigenvectors of the normalized Laplacian of the similarity matrix and Gaussian mixture model (GMM) are utilized to compute the likelihood ratio needed for voice activity detection. Simulation results demonstrate the advantage of the proposed method compared to conventional statistical model-based VAD algorithms in presence of transient noise.","PeriodicalId":55014,"journal":{"name":"IEEE Transactions on Audio Speech and Language Processing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1109/TASL.2013.2248717","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"62888418","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-06-01DOI: 10.1109/TASL.2013.2245650
Lars-Johan Brännmark, A. Bahne, A. Ahlén
A new multichannel approach to robust broadband loudspeaker-room equalization is presented. Traditionally, the equalization (or room correction) problem has been treated primarily by single-channel methods, where loudspeaker input signals are prefiltered individually by separate scalar filters. Single-channel methods are generally able to improve the average spectral flatness of the acoustic transfer functions in a listening region, but they cannot reduce the variability of the transfer functions within the region. Most modern audio reproduction systems, however, contain two or more loudspeakers, and in this paper we aim at improving the equalization performance by using all available loudspeakers jointly. To this end we propose a polynomial based MIMO formulation of the equalization problem. The new approach, which is a generalization of an earlier single-channel approach by the authors, is found to reduce the average reproduction error and the transfer function variability over a region in space. Moreover, pre-ringing artifacts are avoided, and the reproduction error below 1000 Hz is significantly reduced with an amount that scales with the number of loudspeakers used.
{"title":"Compensation of Loudspeaker–Room Responses in a Robust MIMO Control Framework","authors":"Lars-Johan Brännmark, A. Bahne, A. Ahlén","doi":"10.1109/TASL.2013.2245650","DOIUrl":"https://doi.org/10.1109/TASL.2013.2245650","url":null,"abstract":"A new multichannel approach to robust broadband loudspeaker-room equalization is presented. Traditionally, the equalization (or room correction) problem has been treated primarily by single-channel methods, where loudspeaker input signals are prefiltered individually by separate scalar filters. Single-channel methods are generally able to improve the average spectral flatness of the acoustic transfer functions in a listening region, but they cannot reduce the variability of the transfer functions within the region. Most modern audio reproduction systems, however, contain two or more loudspeakers, and in this paper we aim at improving the equalization performance by using all available loudspeakers jointly. To this end we propose a polynomial based MIMO formulation of the equalization problem. The new approach, which is a generalization of an earlier single-channel approach by the authors, is found to reduce the average reproduction error and the transfer function variability over a region in space. Moreover, pre-ringing artifacts are avoided, and the reproduction error below 1000 Hz is significantly reduced with an amount that scales with the number of loudspeakers used.","PeriodicalId":55014,"journal":{"name":"IEEE Transactions on Audio Speech and Language Processing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1109/TASL.2013.2245650","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"62887740","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-06-01DOI: 10.1109/TASL.2013.2245653
John Kane, C. Gobl
This paper proposes a new parameter, the Maxima Dispersion Quotient (MDQ), for differentiating breathy to tense voice. Maxima derived following wavelet decomposition are often used for detecting edges in image processing, where locations of these maxima organize in the vicinity of the edge location. Similarly for tense voice, which typically displays sharp glottal closing characteristics, maxima following wavelet analysis are organized in the vicinity of the glottal closure instant (GCI). Contrastingly, as the phonation type tends away from tense voice towards a breathier phonation it is observed that the maxima become increasingly dispersed. The MDQ parameter is designed to quantify the extent of this dispersion and is shown to compare favorably to existing voice quality parameters, particularly for the analysis of continuous speech. Also, classification experiments reveal a significant improvement in the detection of the voice qualities when MDQ is included as an input to the classifier. Finally, MDQ is shown to be robust to additive noise down to a Signal-to-Noise Ratio of 10 dB.
{"title":"Wavelet Maxima Dispersion for Breathy to Tense Voice Discrimination","authors":"John Kane, C. Gobl","doi":"10.1109/TASL.2013.2245653","DOIUrl":"https://doi.org/10.1109/TASL.2013.2245653","url":null,"abstract":"This paper proposes a new parameter, the Maxima Dispersion Quotient (MDQ), for differentiating breathy to tense voice. Maxima derived following wavelet decomposition are often used for detecting edges in image processing, where locations of these maxima organize in the vicinity of the edge location. Similarly for tense voice, which typically displays sharp glottal closing characteristics, maxima following wavelet analysis are organized in the vicinity of the glottal closure instant (GCI). Contrastingly, as the phonation type tends away from tense voice towards a breathier phonation it is observed that the maxima become increasingly dispersed. The MDQ parameter is designed to quantify the extent of this dispersion and is shown to compare favorably to existing voice quality parameters, particularly for the analysis of continuous speech. Also, classification experiments reveal a significant improvement in the detection of the voice qualities when MDQ is included as an input to the classifier. Finally, MDQ is shown to be robust to additive noise down to a Signal-to-Noise Ratio of 10 dB.","PeriodicalId":55014,"journal":{"name":"IEEE Transactions on Audio Speech and Language Processing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1109/TASL.2013.2245653","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"62888092","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-06-01DOI: 10.1109/TASL.2013.2248722
Tom Ko, B. Mak
Most automatic speech recognizers employ tied-state triphone hidden Markov models (HMM), in which the corresponding triphone states of the same base phone are tied. State tying is commonly performed with the use of a phonetic regression class tree which renders robust context-dependent modeling possible by carefully balancing the amount of training data with the degree of tying. However, tying inevitably introduces quantization error: triphones tied to the same state are not distinguishable in that state. Recently we proposed a new triphone modeling approach called eigentriphone modeling in which all triphone models are, in general, distinct. The idea is to create an eigenbasis for each base phone (or phone state) and all its triphones (or triphone states) are represented as distinct points in the space spanned by the basis. We have shown that triphone HMMs trained using model-based or state-based eigentriphones perform at least as well as conventional tied-state HMMs. In this paper, we further generalize the definition of eigentriphones over clusters of acoustic units. Our experiments on TIMIT phone recognition and the Wall Street Journal 5K-vocabulary continuous speech recognition show that eigentriphones estimated from state clusters defined by the nodes in the same phonetic regression class tree used in state tying result in further performance gain.
{"title":"Eigentriphones for Context-Dependent Acoustic Modeling","authors":"Tom Ko, B. Mak","doi":"10.1109/TASL.2013.2248722","DOIUrl":"https://doi.org/10.1109/TASL.2013.2248722","url":null,"abstract":"Most automatic speech recognizers employ tied-state triphone hidden Markov models (HMM), in which the corresponding triphone states of the same base phone are tied. State tying is commonly performed with the use of a phonetic regression class tree which renders robust context-dependent modeling possible by carefully balancing the amount of training data with the degree of tying. However, tying inevitably introduces quantization error: triphones tied to the same state are not distinguishable in that state. Recently we proposed a new triphone modeling approach called eigentriphone modeling in which all triphone models are, in general, distinct. The idea is to create an eigenbasis for each base phone (or phone state) and all its triphones (or triphone states) are represented as distinct points in the space spanned by the basis. We have shown that triphone HMMs trained using model-based or state-based eigentriphones perform at least as well as conventional tied-state HMMs. In this paper, we further generalize the definition of eigentriphones over clusters of acoustic units. Our experiments on TIMIT phone recognition and the Wall Street Journal 5K-vocabulary continuous speech recognition show that eigentriphones estimated from state clusters defined by the nodes in the same phonetic regression class tree used in state tying result in further performance gain.","PeriodicalId":55014,"journal":{"name":"IEEE Transactions on Audio Speech and Language Processing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1109/TASL.2013.2248722","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"62888586","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-05-01DOI: 10.1109/TASL.2012.2234110
Guangzhao Bao, Z. Ye, Xu Xu, Yingyue Zhou
This paper discusses underdetermined blind source separation (BSS) using a compressed sensing (CS) approach, which contains two stages. In the first stage we exploit a modified K-means method to estimate the unknown mixing matrix. The second stage is to separate the sources from the mixed signals using the estimated mixing matrix from the first stage. In the second stage a two-layer sparsity model is used. The two-layer sparsity model assumes that the low frequency components of speech signals are sparse on K-SVD dictionary and the high frequency components are sparse on discrete cosine transformation (DCT) dictionary. This model, taking advantage of two dictionaries, can produce effective separation performance even if the sources are not sparse in time-frequency (TF) domain.
{"title":"A Compressed Sensing Approach to Blind Separation of Speech Mixture Based on a Two-Layer Sparsity Model","authors":"Guangzhao Bao, Z. Ye, Xu Xu, Yingyue Zhou","doi":"10.1109/TASL.2012.2234110","DOIUrl":"https://doi.org/10.1109/TASL.2012.2234110","url":null,"abstract":"This paper discusses underdetermined blind source separation (BSS) using a compressed sensing (CS) approach, which contains two stages. In the first stage we exploit a modified K-means method to estimate the unknown mixing matrix. The second stage is to separate the sources from the mixed signals using the estimated mixing matrix from the first stage. In the second stage a two-layer sparsity model is used. The two-layer sparsity model assumes that the low frequency components of speech signals are sparse on K-SVD dictionary and the high frequency components are sparse on discrete cosine transformation (DCT) dictionary. This model, taking advantage of two dictionaries, can produce effective separation performance even if the sources are not sparse in time-frequency (TF) domain.","PeriodicalId":55014,"journal":{"name":"IEEE Transactions on Audio Speech and Language Processing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1109/TASL.2012.2234110","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"62884843","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-05-01DOI: 10.1109/TASL.2013.2239990
H. Sawada, H. Kameoka, S. Araki, N. Ueda
This paper presents new formulations and algorithms for multichannel extensions of non-negative matrix factorization (NMF). The formulations employ Hermitian positive semidefinite matrices to represent a multichannel version of non-negative elements. Multichannel Euclidean distance and multichannel Itakura-Saito (IS) divergence are defined based on appropriate statistical models utilizing multivariate complex Gaussian distributions. To minimize this distance/divergence, efficient optimization algorithms in the form of multiplicative updates are derived by using properly designed auxiliary functions. Two methods are proposed for clustering NMF bases according to the estimated spatial property. Convolutive blind source separation (BSS) is performed by the multichannel extensions of NMF with the clustering mechanism. Experimental results show that 1) the derived multiplicative update rules exhibited good convergence behavior, and 2) BSS tasks for several music sources with two microphones and three instrumental parts were evaluated successfully.
{"title":"Multichannel Extensions of Non-Negative Matrix Factorization With Complex-Valued Data","authors":"H. Sawada, H. Kameoka, S. Araki, N. Ueda","doi":"10.1109/TASL.2013.2239990","DOIUrl":"https://doi.org/10.1109/TASL.2013.2239990","url":null,"abstract":"This paper presents new formulations and algorithms for multichannel extensions of non-negative matrix factorization (NMF). The formulations employ Hermitian positive semidefinite matrices to represent a multichannel version of non-negative elements. Multichannel Euclidean distance and multichannel Itakura-Saito (IS) divergence are defined based on appropriate statistical models utilizing multivariate complex Gaussian distributions. To minimize this distance/divergence, efficient optimization algorithms in the form of multiplicative updates are derived by using properly designed auxiliary functions. Two methods are proposed for clustering NMF bases according to the estimated spatial property. Convolutive blind source separation (BSS) is performed by the multichannel extensions of NMF with the clustering mechanism. Experimental results show that 1) the derived multiplicative update rules exhibited good convergence behavior, and 2) BSS tasks for several music sources with two microphones and three instrumental parts were evaluated successfully.","PeriodicalId":55014,"journal":{"name":"IEEE Transactions on Audio Speech and Language Processing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1109/TASL.2013.2239990","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"62886114","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-05-01DOI: 10.1109/TASL.2013.2244087
Coskun Mermer, M. Saraçlar, R. Sarikaya
We present a Bayesian approach to word alignment inference in IBM Models 1 and 2. In the original approach, word translation probabilities (i.e., model parameters) are estimated using the expectation-maximization (EM) algorithm. In the proposed approach, they are random variables with a prior and are integrated out during inference. We use Gibbs sampling to infer the word alignment posteriors. The inferred word alignments are compared against EM and variational Bayes (VB) inference in terms of their end-to-end translation performance on several language pairs and types of corpora up to 15 million sentence pairs. We show that Bayesian inference outperforms both EM and VB in the majority of test cases. Further analysis reveals that the proposed method effectively addresses the high-fertility rare word problem in EM and unaligned rare word problem in VB, achieves higher agreement and vocabulary coverage rates than both, and leads to smaller phrase tables.
{"title":"Improving Statistical Machine Translation Using Bayesian Word Alignment and Gibbs Sampling","authors":"Coskun Mermer, M. Saraçlar, R. Sarikaya","doi":"10.1109/TASL.2013.2244087","DOIUrl":"https://doi.org/10.1109/TASL.2013.2244087","url":null,"abstract":"We present a Bayesian approach to word alignment inference in IBM Models 1 and 2. In the original approach, word translation probabilities (i.e., model parameters) are estimated using the expectation-maximization (EM) algorithm. In the proposed approach, they are random variables with a prior and are integrated out during inference. We use Gibbs sampling to infer the word alignment posteriors. The inferred word alignments are compared against EM and variational Bayes (VB) inference in terms of their end-to-end translation performance on several language pairs and types of corpora up to 15 million sentence pairs. We show that Bayesian inference outperforms both EM and VB in the majority of test cases. Further analysis reveals that the proposed method effectively addresses the high-fertility rare word problem in EM and unaligned rare word problem in VB, achieves higher agreement and vocabulary coverage rates than both, and leads to smaller phrase tables.","PeriodicalId":55014,"journal":{"name":"IEEE Transactions on Audio Speech and Language Processing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1109/TASL.2013.2244087","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"62887557","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-05-01DOI: 10.1109/TASL.2013.2243435
N. Mohammadiha, A. Leijon
Deriving a good model for multitalker babble noise can facilitate different speech processing algorithms, e.g., noise reduction, to reduce the so-called cocktail party difficulty. In the available systems, the fact that the babble waveform is generated as a sum of N different speech waveforms is not exploited explicitly. In this paper, first we develop a gamma hidden Markov model for power spectra of the speech signal, and then formulate it as a sparse nonnegative matrix factorization (NMF). Second, the sparse NMF is extended by relaxing the sparsity constraint, and a novel model for babble noise (gamma nonnegative HMM) is proposed in which the babble basis matrix is the same as the speech basis matrix, and only the activation factors (weights) of the basis vectors are different for the two signals over time. Finally, a noise reduction algorithm is proposed using the derived speech and babble models. All of the stationary model parameters are estimated using the expectation-maximization (EM) algorithm, whereas the time-varying parameters, i.e., the gain parameters of speech and babble signals, are estimated using a recursive EM algorithm. The objective and subjective listening evaluations show that the proposed babble model and the final noise reduction algorithm significantly outperform the conventional methods.
{"title":"Nonnegative HMM for Babble Noise Derived From Speech HMM: Application to Speech Enhancement","authors":"N. Mohammadiha, A. Leijon","doi":"10.1109/TASL.2013.2243435","DOIUrl":"https://doi.org/10.1109/TASL.2013.2243435","url":null,"abstract":"Deriving a good model for multitalker babble noise can facilitate different speech processing algorithms, e.g., noise reduction, to reduce the so-called cocktail party difficulty. In the available systems, the fact that the babble waveform is generated as a sum of N different speech waveforms is not exploited explicitly. In this paper, first we develop a gamma hidden Markov model for power spectra of the speech signal, and then formulate it as a sparse nonnegative matrix factorization (NMF). Second, the sparse NMF is extended by relaxing the sparsity constraint, and a novel model for babble noise (gamma nonnegative HMM) is proposed in which the babble basis matrix is the same as the speech basis matrix, and only the activation factors (weights) of the basis vectors are different for the two signals over time. Finally, a noise reduction algorithm is proposed using the derived speech and babble models. All of the stationary model parameters are estimated using the expectation-maximization (EM) algorithm, whereas the time-varying parameters, i.e., the gain parameters of speech and babble signals, are estimated using a recursive EM algorithm. The objective and subjective listening evaluations show that the proposed babble model and the final noise reduction algorithm significantly outperform the conventional methods.","PeriodicalId":55014,"journal":{"name":"IEEE Transactions on Audio Speech and Language Processing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1109/TASL.2013.2243435","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"62885800","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-05-01DOI: 10.1109/TASL.2013.2239290
J. Jensen, M. G. Christensen, S. H. Jensen
In this paper, we consider the problem of joint direction-of-arrival (DOA) and fundamental frequency estimation. Joint estimation enables robust estimation of these parameters in multi-source scenarios where separate estimators may fail. First, we derive the exact and asymptotic Cramér-Rao bounds for the joint estimation problem. Then, we propose a nonlinear least squares (NLS) and an approximate NLS (aNLS) estimator for joint DOA and fundamental frequency estimation. The proposed estimators are maximum likelihood estimators when: 1) the noise is white Gaussian, 2) the environment is anechoic, and 3) the source of interest is in the far-field. Otherwise, the methods still approximately yield maximum likelihood estimates. Simulations on synthetic data show that the proposed methods have similar or better performance than state-of-the-art methods for DOA and fundamental frequency estimation. Moreover, simulations on real-life data indicate that the NLS and aNLS methods are applicable even when reverberation is present and the noise is not white Gaussian.
{"title":"Nonlinear Least Squares Methods for Joint DOA and Pitch Estimation","authors":"J. Jensen, M. G. Christensen, S. H. Jensen","doi":"10.1109/TASL.2013.2239290","DOIUrl":"https://doi.org/10.1109/TASL.2013.2239290","url":null,"abstract":"In this paper, we consider the problem of joint direction-of-arrival (DOA) and fundamental frequency estimation. Joint estimation enables robust estimation of these parameters in multi-source scenarios where separate estimators may fail. First, we derive the exact and asymptotic Cramér-Rao bounds for the joint estimation problem. Then, we propose a nonlinear least squares (NLS) and an approximate NLS (aNLS) estimator for joint DOA and fundamental frequency estimation. The proposed estimators are maximum likelihood estimators when: 1) the noise is white Gaussian, 2) the environment is anechoic, and 3) the source of interest is in the far-field. Otherwise, the methods still approximately yield maximum likelihood estimates. Simulations on synthetic data show that the proposed methods have similar or better performance than state-of-the-art methods for DOA and fundamental frequency estimation. Moreover, simulations on real-life data indicate that the NLS and aNLS methods are applicable even when reverberation is present and the noise is not white Gaussian.","PeriodicalId":55014,"journal":{"name":"IEEE Transactions on Audio Speech and Language Processing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1109/TASL.2013.2239290","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"62885905","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-05-01DOI: 10.1109/TASL.2013.2244085
Ramón Fernández Astudillo, R. Orglmeister
In this paper we demonstrate how uncertainty propagation allows the computation of minimum mean square error (MMSE) estimates in the feature domain for various feature extraction methods using short-time Fourier transform (STFT) domain distortion models. In addition to this, a measure of estimate reliability is also attained which allows either feature re-estimation or the dynamic compensation of automatic speech recognition (ASR) models. The proposed method transforms the posterior distribution associated to a Wiener filter through the feature extraction using the STFT Uncertainty Propagation formulas. It is also shown that non-linear estimators in the STFT domain like the Ephraim-Malah filters can be seen as special cases of a propagation of the Wiener posterior. The method is illustrated by developing two MMSE-Mel-frequency Cepstral Coefficient (MFCC) estimators and combining them with observation uncertainty techniques. We discuss similarities with other MMSE-MFCC estimators and show how the proposed approach outperforms conventional MMSE estimators in the STFT domain on the AURORA4 robust ASR task.
{"title":"Computing MMSE Estimates and Residual Uncertainty Directly in the Feature Domain of ASR using STFT Domain Speech Distortion Models","authors":"Ramón Fernández Astudillo, R. Orglmeister","doi":"10.1109/TASL.2013.2244085","DOIUrl":"https://doi.org/10.1109/TASL.2013.2244085","url":null,"abstract":"In this paper we demonstrate how uncertainty propagation allows the computation of minimum mean square error (MMSE) estimates in the feature domain for various feature extraction methods using short-time Fourier transform (STFT) domain distortion models. In addition to this, a measure of estimate reliability is also attained which allows either feature re-estimation or the dynamic compensation of automatic speech recognition (ASR) models. The proposed method transforms the posterior distribution associated to a Wiener filter through the feature extraction using the STFT Uncertainty Propagation formulas. It is also shown that non-linear estimators in the STFT domain like the Ephraim-Malah filters can be seen as special cases of a propagation of the Wiener posterior. The method is illustrated by developing two MMSE-Mel-frequency Cepstral Coefficient (MFCC) estimators and combining them with observation uncertainty techniques. We discuss similarities with other MMSE-MFCC estimators and show how the proposed approach outperforms conventional MMSE estimators in the STFT domain on the AURORA4 robust ASR task.","PeriodicalId":55014,"journal":{"name":"IEEE Transactions on Audio Speech and Language Processing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1109/TASL.2013.2244085","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"62886263","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}