Modified Group Delay Function Research Articles

Multichannel CNN-BLSTM Architecture for Speech Emotion Recognition System by Fusion of Magnitude and Phase Spectral Features Using DCCA for Consumer Applications

Conventional Speech Emotion Recognition (SER) approaches put more emphasis on extracting magnitude spectrum-based features, such as Mel Frequency Cepstral Coefficients (MFCCs), and Mel spectrogram. However, phase information is ignored due to signal processing difficulties such as the phase wrapping issue. This work develops a multichannel Convolution Neural Network-Bidirectional Long Short Term Memory (CNN-BLSTM) architectures with an attention mechanism for speaker-independent SER by considering phase and magnitude spectrum-based features. The phase-based features are extracted using the Modified Group Delay Function (MODGD). The obtained phase features are combined with MFCC features. The CNN-BLSTM network extract learned representation from magnitude and phase features. The learned representation from MFCCs and MODGD are combined and given as an input to the Support Vector Machine (SVM) for classification. The Deep Canonical Correlation Analysis (DCCA) is used to maximize the correlation between magnitude and phase information to improve the conventional SER system’s performance. Here the IEMOCAP database is used for performance analysis. The experimental results show improvement over MFCC features and existing approaches for unimodal SER. In this work, we also developed real-time web server application for the proposed architecture.

Efficient Feature Extraction Algorithms to Develop an Arabic Speech Recognition System

This paper studies three feature extraction methods, Mel-Frequency Cepstral Coefficients (MFCC), Power-Normalized Cepstral Coefficients (PNCC), and Modified Group Delay Function (ModGDF) for the development of an Automated Speech Recognition System (ASR) in Arabic. The Support Vector Machine (SVM) algorithm processed the obtained features. These feature extraction algorithms extract speech or voice characteristics and process the group delay functionality calculated straight from the voice signal. These algorithms were deployed to extract audio forms from Arabic speakers. PNCC provided the best recognition results in Arabic speech in comparison with the other methods. Simulation results showed that PNCC and ModGDF were more accurate than MFCC in Arabic speech recognition.

Quantitative Assessment of Speech in Cerebellar Ataxia Using Magnitude and Phase Based Cepstrum.

The clinical assessment of speech abnormalities in Cerebellar Ataxia (CA) is time-consuming and inconsistent. We have developed an automated objective system to quantify CA severity and thereby facilitate remote monitoring and optimisation of therapeutic interventions. A quantitative acoustic assessment could prove to be a viable biomarker for this purpose. Our study explores the use of phase-based cepstral features extracted from the modified group delay function as a complement to the features obtained from the magnitude cepstrum. We selected a combination of 15 acoustic measurements using RELIEF feature selection algorithm during the feature optimisation process. These features were used to segregate ataxic speakers from normal speakers (controls) and objectively assess them based on their severity. The effectiveness of our study has been experimentally evaluated through a clinical study involving 42 patients diagnosed with CA and 23 age-matched controls. A radial basis function kernel based support vector machine (SVM) classifier achieved a classification accuracy of 84.6% in CA-Control discrimination [area under the ROC curve (AUC) of 0.97] and 74% in the modified 3-level CA severity estimation (AUC of 0.90) deduced from the clinical ratings. The strong classification ability of selected features and the SVM model supports this scheme's suitability for monitoring CA related speech motor abnormalities.

Developing Charcot-Marie-Tooth Disease Recognition System Using Bacterial Foraging Optimization Algorithm Based Spiking Neural Network.

In the developing technology Charcot-Marie-Tooth (CMT) disease is one of the teeth diseases which are occurred due to the genetic reason. The CMT disease affects the muscle tissue which reduces the progressive growth of the muscle. So, the CMT disease needs to be recognized carefully for eliminating the risk factors in the early stage. At the time of this process, the system handles the difficulties while performing feature extraction and classification part. So, the teeth images are processed by applying the normalization method which eliminates the salt and pepper noise from data. From that, modified group delay function along with Cepstral coefficient features are extracted with effective manner. After that Bacterial Foraging Optimization Algorithm based features are selected. Then the selected features are examined by applying the Bacterial Foraging Optimization Algorithm based spiking neural network which successfully recognizes the CMT disease. At that point the productivity of the framework is assessed with the assistance of exploratory outcomes.

Performance comparison of multitaper techniques for speaker verification with expressive speech

In this paper, we provide a comparative study of spectral front-end features used as representations for speech signals by processing multitaper magnitude and phase spectra, for speaker verification with expressive speech. In particular, the multitaper modified group delay function (MT-MOGDF) and multitaper magnitude (MT-MAG) spectra of the speech signals are employed to obtain low variance estimates of speech spectra. We observe that the cues that aid in representation of expressive speech are evident in the MT-MOGDF spectrum than the MT-MAG spectrum in terms of mean Formant value and Formant bandwidth. Our extensive experimental study on a speaker verification system with a Gaussian mixture model based universal background model classifier on expressive speech using the IITKGP-SESC and EMODB databases show that MT-MOGDF performs better than MT-MAG technique, in terms of equal error rate and minimum decision cost function. This improvement due to MT-MOGDF is owed to a better representation and a low-variance estimate of the speech spectrum. Our results highlight the utility of MT-MOGDF as a potential alternative for MT-MAG representation for speaker verification problems in general.

Two-pitch tracking in co-channel speech using modified group delay functions

Modified group delay functions are beginning to gain significance in the literature for formant estimation, speaker recognition and speech recognition. In particular, group delay functions have the property that they possess higher resolution compared to that of the magnitude spectrum. In this paper, modified group delay functions are used for the estimation and tracking of two pitches in concurrent speech. The power spectrum of the speech signal is first flattened to annihilate the system characteristics, while retaining the source characteristics. Group delay analysis of the flattened spectrum is performed and the predominant pitch is computed. Next, a comb filter is designed to remove the predominant pitch and its harmonics from the group delay spectrum. The residual spectrum is again subjected to group delay analysis and the next candidate pitch is again estimated using modified group delay processing. The first and second pass pitch trajectories are corrected using post processing. The performance of the proposed algorithm was evaluated on two datasets using two metrics; pitch accuracy and standard deviation of fine pitch error. Our results show that phase based processing holds promise in the context of multipitch estimation.

Melody extraction from music using modified group delay functions

Modified group delay based algorithms for estimation of melodic pitch sequences from heterphonic/polyphonic music are discussed in this paper. Two different variants of the modified group delay function are proposed, namely, (a) system based--MODGD (Direct) and (b) source based--MODGD (Source). In (a) the standard modified group delay function (MODGDF) is used to estimate prominent melodic pitch ($$f_0$$f0), which appears like a low frequency formant in the MODGDF spectrum. In (b), the power spectrum of the signal is first flattened to emphasise the source. The flattened power spectrum behaves like a sinusoid in noise, the frequency of the sinusoid being related to the pitch frequency. The modified group delay function of this signal produces peaks at $$T_0$$T0, $$2T_0, \ldots ,$$2T0,ź, where $$T_0=\frac{1}{f_0}$$T0=1f0. Continuity constraints in a dynamic programming framework are imposed across frames to reduce octave errors. Sudden changes in pitch are accommodated by changing the frame size dynamically using a multi-resolution framework. The performance of the proposed systems was evaluated on four datasets: ADC-2004, LabROSA, MIREX-2008 and Carnatic music dataset. The performance of the proposed approaches demonstrate the potential of the group delay based methods for melody extraction.

Group delay functions and its applications in speech technology

Traditionally, the information in speech signals is represented in terms of features derived from short-time Fourier analysis. In this analysis the features extracted from the magnitude of the Fourier transform (FT) are considered, ignoring the phase component. Although the significance of the FT phase was highlighted in several studies over the recent three decades, the features of the FT phase were not exploited fully due to difficulty in computing the phase and also in processing the phase function. The information in the short-time FT phase function can be extracted by processing the derivative of the FT phase, i.e., the group delay function. In this paper, the properties of the group delay functions are reviewed, highlighting the importance of the FT phase for representing information in the speech signal. Methods to process the group delay function are discussed to capture the characteristics of the vocal-tract system in the form of formants or through a modified group delay function. Applications of group delay functions for speech processing are discussed in some detail. They include segmentation of speech into syllable boundaries, exploiting the additive and high resolution properties of the group delay functions. The effectiveness of segmentation of speech, and the features derived from the modified group delay are demonstrated in applications such as language identification, speech recognition and speaker recognition. The paper thus demonstrates the need to exploit the potential of the group delay functions for development of speech systems.

Significance of the Modified Group Delay Feature in Speech Recognition

Spectral representation of speech is complete when both the Fourier transform magnitude and phase spectra are specified. In conventional speech recognition systems, features are generally derived from the short-time magnitude spectrum. Although the importance of Fourier transform phase in speech perception has been realized, few attempts have been made to extract features from it. This is primarily because the resonances of the speech signal which manifest as transitions in the phase spectrum are completely masked by the wrapping of the phase spectrum. Hence, an alternative to processing the Fourier transform phase, for extracting speech features, is to process the group delay function which can be directly computed from the speech signal. The group delay function has been used in earlier efforts, to extract pitch and formant information from the speech signal. In all these efforts, no attempt was made to extract features from the speech signal and use them for speech recognition applications. This is primarily because the group delay function fails to capture the short-time spectral structure of speech owing to zeros that are close to the unit circle in the z-plane and also due to pitch periodicity effects. In this paper, the group delay function is modified to overcome these effects. Cepstral features are extracted from the modified group delay function and are called the modified group delay feature (MODGDF). The MODGDF is used for three speech recognition tasks namely, speaker, language, and continuous-speech recognition. Based on the results of feature and performance evaluation, the significance of the MODGDF as a new feature for speech recognition is discussed

Significance of Joint Features Derived from the Modified Group Delay Function in Speech Processing

This paper investigates the significance of combining cepstral features derived from the modified group delay function and from the short-time spectral magnitude like the MFCC. The conventional group delay function fails to capture the resonant structure and the dynamic range of the speech spectrum primarily due to pitch periodicity effects. The group delay function is modified to suppress these spikes and to restore the dynamic range of the speech spectrum. Cepstral features are derived from the modified group delay function, which are called the modified group delay feature (MODGDF). The complementarity and robustness of the MODGDF when compared to the MFCC are also analyzed using spectral reconstruction techniques. Combination of several spectral magnitude-based features and the MODGDF using feature fusion and likelihood combination is described. These features are then used for three speech processing tasks, namely, syllable, speaker, and language recognition. Results indicate that combining MODGDF with MFCC at the feature level gives significant improvements for speech recognition tasks in noise. Combining the MODGDF and the spectral magnitude-based features gives a significant increase in recognition performance of 11% at best, while combining any two features derived from the spectral magnitude does not give any significant improvement.

Estimation of evolutionary spectrum based on short time Fourier transform and modified group delay

This paper proposes a new estimator for evolutionary spectrum (ES) based on short time Fourier transform (STFT) and modified group delay function (GDFM). The STFT due to its built-in averaging suppresses the crossterms and the GDFM preserves the frequency resolution of the rectangular window as it reduces the Gibbs ripple without using any window function. The new estimator is applicable to random signals as the GDFM removes the effect of the zeros due to input noise driving the time-varying system and provides the system information effectively. The GDFM also provides signal-to-noise ratio enhancement as it removes the zeros due to the associated noise. The performance of the method is illustrated for linear chirp signals, frequency shift keying and for time-varying random process which indicate that its frequency resolution is better than evolutionary periodogram (EP) and STFT and nearer to that of Wigner Ville distribution. Further, its noise immunity is better than those of EP and STFT.

Improved instantaneous power spectrum (IPS) performance: A group delay approach

The instantaneous power spectrum (IPS) has many advantages over the Wigner–Ville distribution. However, it suffers from ringing effect and poor frequency resolution. In this paper, a new method, improved IPS (IIPS) based on modified group delay function (GDFM) has been suggested for reducing the ringing effect. In this approach, since the ringing effect has been removed without using any common window function, it provides improved frequency resolution. Further, in IIPS, the GDFM enables signal-to-noise ratio (SNR) enhancement, in the time–frequency representation (TFR), as it not only reduces the effect of zeros which cause ringing but also those due to noise associated with the signal. The performance of the proposed method is illustrated by simulation for FSK and linear chirp signals and it has been observed that it improves: the frequency resolution significantly and the SNR.