The Experts below are selected from a list of 21603 Experts worldwide ranked by ideXlab platform
Matthias Wölfel - One of the best experts on this subject based on the ideXlab platform.
-
Signal Adaptive Spectral Envelope Estimation for Robust Speech Recognition
Speech Communication, 2009Co-Authors: Matthias WölfelAbstract:This paper describes a novel Spectral Envelope estimation technique which adapts to the characteristics of the observed signal. This is possible via the introduction of a second bilinear transformation into warped (MVDR) Spectral Envelope estimation. As opposed to the first bilinear transformation, however, which is applied in the time domain, the second bilinear transformation must be applied in the frequency domain. This extension enables the resolution of the Spectral Envelope estimate to be steered to lower or higher frequencies, while keeping the overall resolution of the estimate and the frequency axis fixed. When embedded in the feature extraction process of an automatic speech recognition system, it provides for the emphasis of the characteristics of speech features that are relevant for robust classification, while simultaneously suppressing characteristics that are irrelevant for classification. The change in resolution may be steered, for each observation window, by the normalized first autocorrelation coefficient.
-
Signal adaptive Spectral Envelope estimation for robust speech recognition
Speech Communication, 2009Co-Authors: Matthias WölfelAbstract:This paper describes a novel Spectral Envelope estimation technique which adapts to the characteristics of the observed signal. This is possible via the introduction of a second bilinear transformation into warped minimum variance distortionless response (MVDR) Spectral Envelope estimation. As opposed to the first bilinear transformation, however, which is applied in the time domain, the second bilinear transformation must be applied in the frequency domain. This extension enables the resolution of the Spectral Envelope estimate to be steered to lower or higher frequencies, while keeping the overall resolution of the estimate and the frequency axis fixed. When embedded in the feature extraction process of an automatic speech recognition system, it provides for the emphasis of the characteristics of speech features that are relevant for robust classification, while simultaneously suppressing characteristics that are irrelevant for classification. The change in resolution may be steered, for each observation window, by the normalized first autocorrelation coefficient. To evaluate the proposed adaptive Spectral Envelope technique, dubbed warped-twice MVDR, we use two objective functions: class separability and word error rate. Our test set consists of development and evaluation data as provided by NIST for the Rich Transcription 2005 Spring Meeting Recognition Evaluation. For both measures, we observed consistent improvements for several speaker-to-microphone distances. In average, over all distances, the proposed front-end reduces the word error rate by 4% relative compared to the widely used mel-frequency cepstral coefficients as well as perceptual linear prediction.
Xavier Rodet - One of the best experts on this subject based on the ideXlab platform.
-
Evolutionary Spectral Envelope Morphing by Spectral Shape Descriptors
2009Co-Authors: Marcelo Caetano, Xavier RodetAbstract:There has been a great collective effort in the search for perceptually meaningful sound transformation techniques. The transformation of sounds matching target sound descriptors is a promising candidate because the descriptors are thought to capture timbral dimensions corresponding to relevant perceptual features. However, matching the descriptors alone is not enough because there are a large number of perceptually different sounds with the same values of descriptors. In this work, we use evolutionary computation to search for the Spectral Envelope variation that best matches the target Spectral shape descriptors. We were able to achieve a more independent control of the descriptors while preserving the overall perceptual features.
-
Extending efficient Spectral Envelope modeling to Mel-frequency based representation
2008 IEEE International Conference on Acoustics Speech and Signal Processing, 2008Co-Authors: Fernando Villavicencio, Axel Röbel, Xavier RodetAbstract:In this work we consider the problem of Spectral Envelope estimation using spectra with perceptually warped frequency axis. The goal of this work is the reduction of the order of the Spectral Envelope model which will facilitate the use of these Envelopes for training of voice conversion systems. We adapt the true-Envelope estimator to Mel-frequency representations and adapt a recently proposed cepstral model order selection criterion taking into account the distortion of the frequency axis. We evaluate the modified order selection procedure using a perceptual framework for the evaluation of Envelope estimation errors. The experimental evaluation carried out with real speech confirms our modifications. The results demonstrate that the Mel frequency based true Envelope estimator achieves superior Envelope estimation with significantly reduced model order.
-
ICASSP - Extending efficient Spectral Envelope modeling to Mel-frequency based representation
2008 IEEE International Conference on Acoustics Speech and Signal Processing, 2008Co-Authors: Fernando Villavicencio, Axel Röbel, Xavier RodetAbstract:In this work we consider the problem of Spectral Envelope estimation using spectra with perceptually warped frequency axis. The goal of this work is the reduction of the order of the Spectral Envelope model which will facilitate the use of these Envelopes for training of voice conversion systems. We adapt the true-Envelope estimator to Mel-frequency representations and adapt a recently proposed cepstral model order selection criterion taking into account the distortion of the frequency axis. We evaluate the modified order selection procedure using a perceptual framework for the evaluation of Envelope estimation errors. The experimental evaluation carried out with real speech confirms our modifications. The results demonstrate that the Mel frequency based true Envelope estimator achieves superior Envelope estimation with significantly reduced model order.
-
All-Pole Spectral Envelope Modeling with Order Selection for Harmonic Signals
2007Co-Authors: Fernando Villavicencio, Axel Röbel, Xavier RodetAbstract:We present a study into all-pole Spectral Envelope estimation for the case of harmonic signals. We address the problem of the selection of the model order and propose to make use of the fact that the Spectral Envelope is sampled by means of the harmonic structure to derive a reasonable choice for an appropriate model order. The experimental investigation uses synthetic ARMA featured signals with varying fundamental frequency and differing model structure to evaluate the performance of the selected all-pole models. The experimental results confirm the relation between optimal model order and the fundamental frequency.
-
On cepstral and all-pole based Spectral Envelope modeling with unknown model order
Pattern Recognition Letters, 2007Co-Authors: Axel Röbel, Fernando Villavicencio, Xavier RodetAbstract:In this work, we investigate Spectral Envelope estimation for harmonic signals. We address the issue of model order selection and propose to make use of the fact that the Spectral Envelope is sampled by means of the harmonic structure of the signal in order to derive upper bounds for the estimator order. An experimental study is performed using synthetic test signals with various fundamental frequencies and different model structures to evaluate the performance of the Envelope models. Experimental results confirm the relation between optimal model order and fundamental frequency.
Josep Morera - One of the best experts on this subject based on the ideXlab platform.
-
Spectral Envelope analysis in snoring signals from simple snorers and patients with obstructive sleep apnea
International Conference of the IEEE Engineering in Medicine and Biology Society, 2003Co-Authors: Jordi Solasoler, Raimon Jane, José Antonio Fiz, Josep MoreraAbstract:Several studies have shown a relationship between snoring and obstructive sleep apnea syndrome (OSAS). Beyond conventional sound intensity analysis, some Spectral differences have been found between snores from simple snores and post-apneic snores from OSAS patients. Snoring spectrum has two components, a fundamental frequency and a Spectral Envelope. Spectral density studies analyze them altogether. In this work we approach the estimation of the Spectral Envelope alone, which is closely related to the mechanism of snoring production. Linear prediction autoregressive (AR) modeling with a low order is used for Spectral Envelope estimation. Two methods are proposed for automatic order estimation. Formant frequencies of the Spectral Envelope are calculated. A total of 447 snores from 8 simple snorers, and 236 normal and 429 post-apneic snores from 8 OSAS patients are analyzed. Significant differences are found (p<0.005) in formant frequencies variability between simple snorers and OSAS patients, even when non postapneic snores are considered.
-
Spectral Envelope analysis in snoring signals from simple snorers and patients with Obstructive Sleep Apnea
Proceedings of the 25th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (IEEE Cat. No.03CH37439), 1Co-Authors: Jordi Sola-soler, Raimon Jane, José Antonio Fiz, Josep MoreraAbstract:Several studies have shown a relationship between snoring and obstructive sleep apnea syndrome (OSAS). Beyond conventional sound intensity analysis, some Spectral differences have been found between snores from simple snores and post-apneic snores from OSAS patients. Snoring spectrum has two components, a fundamental frequency and a Spectral Envelope. Spectral density studies analyze them altogether. In this work we approach the estimation of the Spectral Envelope alone, which is closely related to the mechanism of snoring production. Linear prediction autoregressive (AR) modeling with a low order is used for Spectral Envelope estimation. Two methods are proposed for automatic order estimation. Formant frequencies of the Spectral Envelope are calculated. A total of 447 snores from 8 simple snorers, and 236 normal and 429 post-apneic snores from 8 OSAS patients are analyzed. Significant differences are found (p
Hwang Soo Lee - One of the best experts on this subject based on the ideXlab platform.
-
Use of Spectral autocorrelation in Spectral Envelope linear prediction for speech recognition
IEEE Transactions on Speech and Audio Processing, 1999Co-Authors: Hong Kook Kim, Hwang Soo LeeAbstract:This paper proposes a linear predictive (LP) analysis method where sample autocorrelations are estimated from the Spectral Envelope of a speech signal on the basis of the Spectral autocorrelation. The Spectral autocorrelation is defined as discrete quantities of speech spectrum with Spectral resolution identical to the discrete Fourier transform (DFT) used to obtain the speech spectrum. From analytical and empirical derivation of its properties, we can estimate the fundamental frequency and the maximally correlated frequency for voiced and unvoiced speech, respectively, and then obtain the Spectral Envelope by sampling at a rate of the estimated frequency. A frequency normalization can be applied to the estimated Spectral Envelope because the number of samples of the Spectral Envelope usually differs from frame to frame. The Spectral Envelope is warped into the mel-frequency scale and the inverse DFT is applied to extract the estimate of sample autocorrelations. From the result of LP analysis on the sample autocorrelations, we finally obtain the Spectral Envelope cepstral coefficients (SECC). Hidden Markov model (HMM) recognition experiments show that SECC significantly improves the performance of a recognizer at low signal-to-noise ratios (SNRs) over several other representations.
-
Spectral Envelope Linear Predictive Analysis of Speech based on the Spectral Autocorrelation
1994Co-Authors: Hong Kook Kim, Hwang Soo LeeAbstract:In this paper, we propose a new linear predictive analysis method where the autocorrelation of speech signal is estimated from the Spectral Envelope of the speech signal on the basis of the Spectral autocorrelation. The Spectral autocorrelation is defined as the autocorrelation of discrete quantities of speech spectrum with Spectral resolution identical to the discrete Fourier transform (DFT) used to obtain the speech spectrum. The characteristic of the Spectral autocorrelation for voiced and unvoiced speech is derived analytically and imperially, respectively. The Spectral Envelope is obtained by smoothing the fine structure of a speech spectrum using the fundamental frequency estimated by the Spectral autocorrelation of speech spectrum. The conventional linear prediction analysis is applied to the sample autocorrelation sequences obtained from the inverse DFT of the Spectral Envelope. From the comparison to LP we can observe that SELP provides lower order representation of speech than LP.
Wölfelmatthias - One of the best experts on this subject based on the ideXlab platform.
-
Signal adaptive Spectral Envelope estimation for robust speech recognition
Speech Communication, 2009Co-Authors: WölfelmatthiasAbstract:This paper describes a novel Spectral Envelope estimation technique which adapts to the characteristics of the observed signal. This is possible via the introduction of a second bilinear transforma...
