Cepstral Mean Subtraction Research Articles

Application of Teager Energy Operator on Linear and Mel Scales for Whispered Speech Recognition

This paper presents experimental results on whispered speech recognition based on Teager Energy Operator for linear and mel cepstral coefficients including the Cepstral Mean Subtraction normalization technique. The feature vectors taken into consideration are Linear Frequency Cepstral Coefficients, Teager Energy based Linear Frequency Cepstral Coefficients, Mel Frequency Cepstral Coefficients and Teager Energy based Mel Frequency Cepstral Coefficients. A speaker dependent scenario is used. For the recognition process, Dynamic Time Warping and Hidden Markov Models methods are applied. Results show a respectable improvement in whispered speech recognition as achieved by using the Teager Energy Operator with Cepstral Mean Subtraction.

Performance Evaluation of Normalization Techniques in Adverse Conditions

Abstract This paper explores the behavior of different normalization techniques viz. cepstral mean normalization, cepstral variance normalization, cepstral mean subtraction. cepstral mean and variance normalization, wiener filter, and spectral subtraction in noisy conditions. The performance parameters viz. EER (Equal Error Rate) and DCF (Detection Cost Function) has been calculated using NIST 2003 SRE and Aurora 2 with the help of various normalization techniques considered in this paper for different noisy backgrounds at 0, 5 and 10 dB signal-to-noise ratio. The experimental results obtained from these techniques reveal that cepstral mean normalization (CVN) normalization method is found to be better when compared to other normalization techniques used in this paper.

Whispered speech recognition based on gammatone filterbank cepstral coefficients

This paper presents the results on whispered speech recognition using gammatone filterbank cepstral coefficients for speaker dependent mode. The isolated words used for this experiment are taken from the Whi-Spe database. Whispered speech recognition is based on dynamic time warping and hidden Markov models methods. The experiments are focused on the following modes: normal speech, whispered speech and their combinations (normal/whispered and whispered/normal). The results demonstrated an important improvement in recognition after application of cepstral mean subtraction, especially in mixed train/test scenarios.

Evaluation of MSR Identity Toolbox under conditions reflecting those of a real forensic case (forensic_eval_01)

This paper presents an evaluation of the Microsoft Research Identity Toolbox version 1.0 developed at Microsoft Research, as a tool for forensic voice comparison under conditions reflecting those of a real forensic case. For this purpose we implement two systems: the first is based on Gaussian mixture model - universal background model (GMM-UBM) and the second on i-vectors with probabilistic linear discriminant analysis (i-vector PLDA). Three different feature-level mismatch compensation techniques were tested, before and after the application of voice activity detection (VAD). The three techniques were global cepstral mean subtraction (CMS), global cepstral mean and variance normalization (CMVN), and local feature warping (FW).

AN ANALYSIS OF TWO COMMON REFERENCE POINTS FOR EEGS.

Clinical electroencephalographic (EEG) data varies significantly depending on a number of operational conditions (e.g., the type and placement of electrodes, the type of electrical grounding used). This investigation explores the statistical differences present in two different referential montages: Linked Ear (LE) and Averaged Reference (AR). Each of these accounts for approximately 45% of the data in the TUH EEG Corpus. In this study, we explore the impact this variability has on machine learning performance. We compare the statistical properties of features generated using these two montages, and explore the impact of performance on our standard Hidden Markov Model (HMM) based classification system. We show that a system trained on LE data significantly outperforms one trained only on AR data (77.2% vs. 61.4%). We also demonstrate that performance of a system trained on both data sets is somewhat compromised (71.4% vs. 77.2%). A statistical analysis of the data suggests that mean, variance and channel normalization should be considered. However, cepstral mean subtraction failed to produce an improvement in performance, suggesting that the impact of these statistical differences is subtler.

New scheme based on GMM-PCA-SVM modelling for automatic speaker recognition

Most of the existing speaker recognition systems are based on the basic GMM, the state of the art GMM-UBM, the SVM or more recently the GMM-SVM modeling. In this paper, a new scheme for Automatic Speaker Recognition (ASR), namely GMM-PCA-SVM, is presented. Dimensionality reduction using Principal Component Analysis (PCA) technique, which was previously applied in the front-end process, is now incorporated in the core of the GMM-SVM modeling part, in order to reduce the size of the adapted means vectors issued from the Universal Background Model (UBM). A Comparative study, using Mel Frequency Cepstral Coefficients (MFCC) with Cepstral Mean Subtraction (CMS) extracted from the TIMIT database is performed for speaker recognition in clean and noisy environments. It is shown that the proposed scheme is a promising way for the ASR task. In fact, the recognition performances using GMM-PCA-SVM proposed method is significantly improved compared to the conventional SVM or GMM-SVM based systems.

감정 음성 인식을 위한 강인한 음성 파라메터

This paper studied the speech parameters less affected by the human emotion for the development of the robust emotional speech recognition system. For this purpose, the effect of emotion on the speech recognition system and robust speech parameters of speech recognition system were studied using speech database containing various emotions. In this study, mel-cepstral coefficient, delta-cepstral coefficient, RASTA mel-cepstral coefficient, root-cepstral coefficient, PLP coefficient and frequency warped mel-cepstral coefficient in the vocal tract length normalization method were used as feature parameters. And CMS (Cepstral Mean Subtraction) and SBR(Signal Bias Removal) method were used as a signal bias removal technique. Experimental results showed that the HMM based speaker independent word recognizer using frequency warped RASTA mel-cepstral coefficient in the vocal tract length normalized method, its derivatives and CMS as a signal bias removal showed the best performance.

Combining log-spectral mean subtraction at different frequency resolutions for handset-channel compensation in single utterance speaker verification

Cepstral mean subtraction (CMS) is a well-known feature domain channel compensation technique employed to eliminate the effects of convolutive channel distortion. However, as the authors use in log-spectral mean subtraction (LSMS), the compensation might be applied in spectral domain before the filter-bank analysis with a higher-frequency resolution. LSMS can also be combined with CMS to further improve the recognition performance. In this study, the authors compare the performances of LSMS and CMS methods using a multi-channel, text-dependent single utterance speaker recognition database. In the experiments, the authors observe that LSMS outperforms CMS especially in the high false acceptance region. Moreover, the accuracy is further improved when the methods are combined together. With the combination, the authors achieve 15.5% relative reduction in equal error rate for no score normalisation and 9.4% for test normalisation cases when compared with the baseline CMS experiment.

Spectro-temporal modulation subspace-spanning filter bank features for robust automatic speech recognition

In an attempt to increase the robustness of automatic speech recognition (ASR) systems, a feature extraction scheme is proposed that takes spectro-temporal modulation frequencies (MF) into account. This physiologically inspired approach uses a two-dimensional filter bank based on Gabor filters, which limits the redundant information between feature components, and also results in physically interpretable features. Robustness against extrinsic variation (different types of additive noise) and intrinsic variability (arising from changes in speaking rate, effort, and style) is quantified in a series of recognition experiments. The results are compared to reference ASR systems using Mel-frequency cepstral coefficients (MFCCs), MFCCs with cepstral mean subtraction (CMS) and RASTA-PLP features, respectively. Gabor features are shown to be more robust against extrinsic variation than the baseline systems without CMS, with relative improvements of 28% and 16% for two training conditions (using only clean training samples or a mixture of noisy and clean utterances, respectively). When used in a state-of-the-art system, improvements of 14% are observed when spectro-temporal features are concatenated with MFCCs, indicating the complementarity of those feature types. An analysis of the importance of specific MF shows that temporal MF up to 25 Hz and spectral MF up to 0.25 cycles/channel are beneficial for ASR.

감정에 강인한 음성 인식을 위한 음성 파라메터

This paper studied the speech parameters less affected by the human emotion for the development of the robust speech recognition system. For this purpose, the effect of emotion on the speech recognition system and robust speech parameters of speech recognition system were studied using speech database containing various emotions. In this study, mel-cepstral coefficient, delta-cepstral coefficient, RASTA mel-cepstral coefficient and frequency warped mel-cepstral coefficient were used as feature parameters. And CMS (Cepstral Mean Subtraction) method were used as a signal bias removal technique. Experimental results showed that the HMM based speaker independent word recognizer using vocal tract length normalized mel-cepstral coefficient, its derivatives and CMS as a signal bias removal showed the best performance of 0.78% word error rate. This corresponds to about a 50% word error reduction as compare to the performance of baseline system using mel-cepstral coefficient, its derivatives and CMS.

The effect of reverberation on the performance of cepstral mean subtraction in speaker verification

Speaker verification (SVR) performance is degraded under reverberation conditions. Cepstral mean subtraction (CMS) is often applied to the feature vectors in order to compensate for convolutive effects of transmission channels, which are considered to have a short-duration impulse response. The effect of reverberation on the performance of CMS applied to the feature vectors in SVR is investigated. Although CMS was found effective in reducing the effect of reverberation for short reverberation time (RT), in cases of long RT, it is shown that CMS may degrade SVR performance rather than improve it. Hence, CMS should not to be used in these cases. In addition, the effect of the room volume was tested and found less critical than the effect of long RT.

The integration of principal component analysis and cepstral mean subtraction in parallel model combination for robust speech recognition

This paper addresses the problem of automatic speech recognition in real applications in which the speech signal is altered by various noises. Feature compensation and model compensation robustness methods are studied. Parallel model combination (PMC) and its recent advances are reviewed and a novel algorithm called PC-PMC is proposed. This algorithm utilizes cepstral mean subtraction (CMS) normalization ability and principal component analysis (PCA) compression and de-correlation capability in the combination with PMC model transformation method. PC-PMC algorithm takes the advantages of additive noise compensation ability of PMC and convolutional noise removal capability of CMS and PCA. In realizing PC-PMC mainly two problems should be solved. The first problem is that PMC method requires invertible modules in the front-end of the system while CMS normalization is not an invertible process. Moreover, when the recognition system is exposed to noisy speech, the adaptation of the PCA transform is required; therefore a framework is to be designed for adaptation of the PCA transform in the presence of noise. The method presented in this paper provides solutions to these problems. Our evaluations are done on the four different real noisy tasks using Nevisa HMM-based, Persian continuous speech recognition system. Experimental results demonstrate significant reduction in the system word error rate using PC-PMC. In addition, the effects of covariance matrix compensation, dynamic features adaptation and the effect of gain parameter in the PMC method are also studied and experimented in the real acoustic conditions. Besides, we have investigated the using of maximum likelihood linear regression (MLLR) and maximum a posteriori (MAP) adaptation techniques in the combination with the introduced PC-PMC method. Finally, a comprehensive discussion on the capabilities of the studied robustness methods is presented.

Soft‐decision a priori knowledge interpolation for robust telephone speaker identification

Handsets which are not seen in the training phase (a.k.a unseen handsets) are main sources of performance degradation for speaker identification (SID) applications in telecommunication environments. To alleviate the problem, a soft‐decision a priori knowledge interpolation (SD‐AKI) method of handset characteristic estimation for handset mismatch‐compensated SID is proposed in this paper. The idea of the SD‐AKI method is to first collect a set of characteristics of seen handsets in the training phase, and to then estimate the characteristic of the unknown testing handset by interpolating the set of seen handset characteristics in the test phase. The estimated handset characteristic is then used to compensate for handset mismatch for robust SID. The SD‐AKI method can be realized in both feature and model spaces. Experimental results on the handset TIMIT (HTIMIT) database showed that both the proposed feature‐ and model‐space SD‐AKI schemes were more robust than the blind cepstral mean subtraction (CMS), feature warping (FW) methods and their hard‐decision counterpart (HD‐AKI) for both cases of all‐handset and unseen‐handset SID tests. It is therefore a promising robust SID method.

Higher Order Cepstral Moment Normalization for Improved Robust Speech Recognition

Cepstral normalization has widely been used as a powerful approach to produce robust features for speech recognition. Good examples of this approach include cepstral mean subtraction, and cepstral mean and variance normalization, in which either the first or both the first and the second moments of the Mel-frequency cepstral coefficients (MFCCs) are normalized. In this paper, we propose the family of higher order cepstral moment normalization, in which the MFCC parameters are normalized with respect to a few moments of orders higher than 1 or 2. The basic idea is that the higher order moments are more dominated by samples with larger values, which are very likely the primary sources of the asymmetry and abnormal flatness or tail size of the parameter distributions. Normalization with respect to these moments therefore puts more emphasis on these signal components and constrains the distributions to be more symmetric with more reasonable flatness and tail size. The fundamental principles behind this approach are also analyzed and discussed based on the statistical properties of the distributions of the MFCC parameters. Experimental results based on the AURORA 2, AURORA 3, AURORA 4, and Resource Management (RM) testing environments show that with the proposed approach, recognition accuracy can be significantly and consistently improved for all types of noise and all SNR conditions.

강인한 음성 인식 시스템을 사용한 감정 인식

본 논문은 음성을 사용한 인간의 감정 인식 시스템의 성능을 향상시키기 위하여 감정 변화에 강인한 음성 인식 시스템과 결합된 감정 인식 시스템에 관하여 연구하였다. 이를 위하여 우선 다양한 감정이 포함된 음성 데이터베이스를 사용하여 감정 변화가 음성 인식 시스템의 성능에 미치는 영향에 관한 연구와 감정 변화의 영향을 적게 받는 음성 인식 시스템을 구현하였다. 감정 인식은 음성 인식의 결과에 따라 입력 문장에 대한 각각의 감정 모델을 비교하여 입력 음성에 대한 최종감정 인식을 수행한다. 실험 결과에서 강인한 음성 인식 시스템은 음성 파라메터로 RASTA 멜 켑스트럼과 델타 켑스트럼을 사용하고 신호편의 제거 방법으로 CMS를 사용한 HMM 기반의 화자독립 단어 인식기를 사용하였다. 이러한 음성 인식기와 결합된 감정 인식을 수행한 결과 감정 인식기만을 사용한 경우보다 좋은 성능을 나타내었다. This paper studied the emotion recognition system combined with robust speech recognition system in order to improve the performance of emotion recognition system. For this purpose, the effect of emotional variation on the speech recognition system and robust feature parameters of speech recognition system were studied using speech database containing various emotions. Final emotion recognition is processed using the input utterance and its emotional model according to the result of speech recognition. In the experiment, robust speech recognition system is HMM based speaker independent word recognizer using RASTA mel-cepstral coefficient and its derivatives and cepstral mean subtraction(CMS) as a signal bias removal. Experimental results showed that emotion recognizer combined with speech recognition system showed better performance than emotion recognizer alone.

Analysis of Feature Extraction and Channel Compensation in a GMM Speaker Recognition System

<para xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> In this paper, several feature extraction and channel compensation techniques found in state-of-the-art speaker verification systems are analyzed and discussed. For the NIST SRE 2006 submission, cepstral mean subtraction, feature warping, RelAtive SpecTrAl (RASTA) filtering, heteroscedastic linear discriminant analysis (HLDA), feature mapping, and eigenchannel adaptation were incrementally added to minimize the system's error rate. This paper deals with eigenchannel adaptation in more detail and includes its theoretical background and implementation issues. The key part of the paper is, however, the post-evaluation analysis, undermining a common myth that “the more boxes in the scheme, the better the system.” All results are presented on NIST Speaker Recognition Evaluation (SRE) 2005 and 2006 data. </para>

HMM-based noise-robust feature compensation

In this paper, we describe a hidden Markov model (HMM)-based feature-compensation method. The proposed method compensates for noise-corrupted speech features in the mel-frequency cepstral coefficient (MFCC) domain using the output probability density functions (pdfs) of the HMM. In compensating the features, the output pdfs are adaptively weighted according to forward path probabilities. Because of this, the proposed method can minimize degradation of feature-compensation accuracy due to a temporarily changing noise environment. We evaluated the proposed method based on the AURORA2 database. All the experiments were conducted under clean conditions. The experimental results indicate that the proposed method, combined with cepstral mean subtraction, can achieve a word accuracy of 87.64%. We also show that the proposed method is useful in a transient pulse noise environment.

Feature normalization for robust individual identification of the ortolan bunting (emberiza hortulana L)

This paper presents the use of cepstral moment normalization for the purpose of improving the accuracy of individual identification in the ortolan bunting (emberiza hortulana L). The underlying classification algorithm is based on hidden markov models (HMMs) using Greenwood function cepstral coefficients (GFCCs), and the proposed feature normalization is implemented via a combination of cepstral mean subtraction and cepstral variance normalization, which are equivalent to a statistical z-score normalization across each example utterance. Results indicate a significant reduction in error on the individual identification task, improving accuracy from 89% to 96% in initial experiments, using a single call type over a small population of 6 individuals.

Blind Stochastic Feature Transformation for Channel Robust Speaker Verification

To improve the reliability of telephone-based speaker verification systems, channel compensation is indispensable. However, it is also important to ensure that the channel compensation algorithms in these systems surpress channel variations and enhance interspeaker distinction. This paper addresses this problem by a blind feature-based transformation approach in which the transformation parameters are determined online without any a priori knowledge of channel characteristics. Specifically, a composite statistical model formed by the fusion of a speaker model and a background model is used to represent the characteristics of enrollment speech. Based on the difference between the claimant's speech and the composite model, a stochastic matching type of approach is proposed to transform the claimant's speech to a region close to the enrollment speech. Therefore, the algorithm can estimate the transformation online without the necessity of detecting the handset types. Experimental results based on the 2001 NIST evaluation set show that the proposed transformation approach achieves significant improvement in both equal error rate and minimum detection cost as compared to cepstral mean subtraction and Znorm.

Automatic speech recognition with an adaptation model motivated by auditory processing

The mel-frequency cepstral coefficient (MFCC) or perceptual linear prediction (PLP) feature extraction typically used for automatic speech recognition (ASR) employ several principles which have known counterparts in the cochlea and auditory nerve: frequency decomposition, mel- or bark-warping of the frequency axis, and compression of amplitudes. It seems natural to ask if one can profitably employ a counterpart of the next physiological processing step, synaptic adaptation. We, therefore, incorporated a simplified model of short-term adaptation into MFCC feature extraction. We evaluated the resulting ASR performance on the AURORA 2 and AURORA 3 tasks, in comparison to ordinary MFCCs, MFCCs processed by RASTA, and MFCCs processed by cepstral mean subtraction (CMS), and both in comparison to and in combination with Wiener filtering. The results suggest that our approach offers a simple, causal robustness strategy which is competitive with RASTA, CMS, and Wiener filtering and performs well in combination with Wiener filtering. Compared to the structurally related RASTA, our adaptation model provides superior performance on AURORA 2 and, if Wiener filtering is used prior to both approaches, on AURORA 3 as well.