采用韵律特征的说话人确认系统

在文本无关的说话人识别中,韵律特征由于其对信道环境噪声不敏感等特性而被应用于话者识别任务中.本文对韵律参数采用基于高斯混合模型超向量的支持向量机建模方法,并将类内协方差特征映射方法应用于模型超向量上,单系统的性能比传统方法的混合高斯-通用背景模型(Gaussian mixture model-universal background model,GMM-UBM)基线系统有了40.19%的提升.该方法与本文的基于声学倒谱参数的确认系统融合后,能使整体系统的识别性能有9.25%的提升.在NIST(National institute of standards and technology mixture)2006说话人测试数据库上,融合后的系统能够取得4.9%的等错误率.     

Robust Speaker Verification with Principal Pitch Components

We are presenting a new method that improves the accuracy of text dependent speaker verification systems. The new method exploits a set of novel speech features derived from a principal component analysis of pitch synchronous voiced speech segments. We use the term principal pitch components (PPCs) or optimal pitch bases (OPBs) to denote the new feature set. Utterance distances computed from these new PPC features are only loosely correlated with utterance distances computed from cepstral features. A distance measure that combines both cepstral and PPC features provides a discriminative power that cannot be achieved with cepstral features alone. By augmenting the feature space of a cepstral baseline system with PPC features we achieve a significant reduction of the equal error probability of incorrect customer rejection versus incorrect impostor acceptance. The proposed method delivers robust performance in various noise conditions.     

Pseudo pitch synchronous analysis of speech with applications to speaker recognition

The fine spectral structure related to pitch information is conveyed in Mel cepstral features, with variations in pitch causing variations in the features. For speaker recognition systems, this phenomenon, known as "pitch mismatch" between training and testing, can increase error rates. Likewise, pitch-related variability may potentially increase error rates in speech recognition systems for languages such as English in which pitch does not carry phonetic information. In addition, for both speech recognition and speaker recognition systems, the parsing of the raw speech signal into frames is traditionally performed using a constant frame size and a constant frame offset, without aligning the frames to the natural pitch cycles. As a result the power spectral estimation that is done as part of the Mel cepstral computation may include artifacts. Pitch synchronous methods have addressed this problem in the past, at the expense of adding some complexity by using a variable frame size and/or offset. This paper introduces Pseudo Pitch Synchronous (PPS) signal processing procedures that attempt to align each individual frame to its natural cycle and avoid truncation of pitch cycles while still using constant frame size and frame offset, in an effort to address the above problems. Text independent speaker recognition experiments performed on NIST speaker recognition tasks demonstrate a performance improvement when the scores produced by systems using PPS are fused with traditional speaker recognition scores. In addition, a better distribution of errors across trials may be obtained for similar error rates, and some insight regarding of role of the fundamental frequency in speaker recognition is revealed. Speech recognition experiments run on the Aurora-2 noisy digits task also show improved robustness and better accuracy for extremely low signal-to-noise ratio (SNR) data.     

Emotion recognition from speech using source, system, and prosodic features

In this work, source, system, and prosodic features of speech are explored for characterizing and classifying the underlying emotions. Different speech features contribute in different ways to express the emotions, due to their complementary nature. Linear prediction residual samples chosen around glottal closure regions, and glottal pulse parameters are used to represent excitation source information. Linear prediction cepstral coefficients extracted through simple block processing and pitch synchronous analysis represent the vocal tract information. Global and local prosodic features extracted from gross statistics and temporal dynamics of the sequence of duration, pitch, and energy values represent the prosodic information. Emotion recognition models are developed using above mentioned features separately, and in combination. Simulated Telugu emotion database (IITKGP-SESC) is used to evaluate the proposed features. The emotion recognition results obtained using IITKGP-SESC are compared with the results of internationally known Berlin emotion speech database (Emo-DB). Autoassociative neural networks, Gaussian mixture models, and support vector machines are used to develop emotion recognition systems with source, system, and prosodic features, respectively. Weighted combination of evidence has been used while combining the performance of systems developed using different features. From the results, it is observed that, each of the proposed speech features has contributed toward emotion recognition. The combination of features improved the emotion recognition performance, indicating the complementary nature of the features.     

Improvement of phone recognition accuracy using speech mode classification

In this work, we have developed a speech mode classification model for improving the performance of phone recognition system (PRS). In this paper, we have explored vocal tract system, excitation source and prosodic features for development of speech mode classification (SMC) model. These features are extracted from voiced regions of a speech signal. In this study, conversation, extempore, and read speech are considered as three different modes of speech. The vocal tract component of speech is extracted using Mel-frequency cepstral coefficients (MFCCs). The excitation source features are captured through Mel power differences of spectrum in sub-bands (MPDSS) and residual Mel-frequency cepstral coefficients (RMFCCs) of the speech signal. The prosody information is extracted from pitch and intensity. Speech mode classification models are developed using above described features independently, and in fusion. The experiments carried out on Bengali speech corpus to analyze the accuracy of the speech mode classification model using the artificial neural network (ANN), naive Bayes, support vector machines (SVMs) and k-nearest neighbor (KNN). We proposed four classification models which are combined using maximum voting approach for optimal performance. From the results, it is observed that speech mode classification model developed using the fusion of vocal tract system, excitation source and prosodic features of speech, yields the best performance of 98%. Finally, the proposed speech mode classifier is integrated to the PRS, and the accuracy of phone recognition system is observed to be improved by 11.08%.     

Analysis and modeling of acoustic information for automatic dialect classification

A primary challenge in the field of automatic speech recognition is to understand and create acoustic models to represent individual differences in their spoken language. Individual’s age, gender; their speaking styles influenced by their dialect may be few of the reasons for these differences. This work investigates the dialectal differences by measuring the analysis of variance of acoustic features such as, formant frequencies, pitch, pitch slope, duration and intensity for vowel sounds. This paper attempts to discuss methods to capture dialect specific knowledge through vocal tract and prosody information extracted from speech that can be utilized for automatic identification of dialects. Kernel based support vector machine is utilized for measuring the dialect discriminating ability of acoustic features. For the spectral feature shifted delta cepstral coefficients along with Mel frequency cepstral coefficients gives a recognition performance of 66.97 %. Combination of prosodic features performs better with a classification score of 74 %. The model is further evaluated for the combination of spectral and prosodic feature set and achieves a classification accuracy of 88.77 %. The proposed model is compared with the human perception of dialects. The overall work is based on four dialects of Hindi; one of the world’s major languages.     

Speaker recognition under stressed condition

This paper presents the feature analysis and design of compensators for speaker recognition under stressed speech conditions. Any condition that causes a speaker to vary his or her speech production from normal or neutral condition is called stressed speech condition. Stressed speech is induced by emotion, high workload, sleep deprivation, frustration and environmental noise. In stressed condition, the characteristics of speech signal are different from that of normal or neutral condition. Due to changes in speech signal characteristics, performance of the speaker recognition system may degrade under stressed speech conditions. Firstly, six speech features (mel-frequency cepstral coefficients (MFCC), linear prediction (LP) coefficients, linear prediction cepstral coefficients (LPCC), reflection coefficients (RC), arc-sin reflection coefficients (ARC) and log-area ratios (LAR)), which are widely used for speaker recognition, are analyzed for evaluation of their characteristics under stressed condition. Secondly, Vector Quantization (VQ) classifier and Gaussian Mixture Model (GMM) are used to evaluate speaker recognition results with different speech features. This analysis help select the best feature set for speaker recognition under stressed condition. Finally, four VQ based novel compensation techniques are proposed and evaluated for improvement of speaker recognition under stressed condition. The compensation techniques are speaker and stressed information based compensation (SSIC), compensation by removal of stressed vectors (CRSV), cepstral mean normalization (CMN) and combination of MFCC and sinusoidal amplitude (CMSA) features. Speech data from SUSAS database corresponding to four different stressed conditions, Angry, Lombard, Question and Neutral, are used for analysis of speaker recognition under stressed condition.     

Local spectral variability features for speaker verification

Speaker verification techniques neglect the short-time variation in the feature space even though it contains speaker related attributes. We propose a simple method to capture and characterize this spectral variation through the eigenstructure of the sample covariance matrix. This covariance is computed using sliding window over spectral features. The newly formulated feature vectors representing local spectral variations are used with classical and state-of-the-art speaker recognition systems. Results on multiple speaker recognition evaluation corpora reveal that eigenvectors weighted with their normalized singular values are useful in representing local covariance information. We have also shown that local variability features can be extracted using mel frequency cepstral coefficients (MFCCs) as well as using three recently developed features: frequency domain linear prediction (FDLP), mean Hilbert envelope coefficients (MHECs) and power-normalized cepstral coefficients (PNCCs). Since information conveyed in the proposed feature is complementary to the standard short-term features, we apply different fusion techniques. We observe considerable relative improvements in speaker verification accuracy in combined mode on text-independent (NIST SRE) and text-dependent (RSR2015) speech corpora. We have obtained up to 12.28% relative improvement in speaker recognition accuracy on text-independent corpora. Conversely in experiments on text-dependent corpora, we have achieved up to 40% relative reduction in EER. To sum up, combining local covariance information with the traditional cepstral features holds promise as an additional speaker cue in both text-independent and text-dependent recognition.     

Automatic speaker age and gender recognition using acoustic and prosodic level information fusion

The paper presents a novel automatic speaker age and gender identification approach which combines seven different methods at both acoustic and prosodic levels to improve the baseline performance. The three baseline subsystems are (1) Gaussian mixture model (GMM) based on mel-frequency cepstral coefficient (MFCC) features, (2) Support vector machine (SVM) based on GMM mean supervectors and (3) SVM based on 450-dimensional utterance level features including acoustic, prosodic and voice quality information. In addition, we propose four subsystems: (1) SVM based on UBM weight posterior probability supervectors using the Bhattacharyya probability product kernel, (2) Sparse representation based on UBM weight posterior probability supervectors, (3) SVM based on GMM maximum likelihood linear regression (MLLR) matrix supervectors and (4) SVM based on the polynomial expansion coefficients of the syllable level prosodic feature contours in voiced speech segments. Contours of pitch, time domain energy, frequency domain harmonic structure energy and formant for each syllable (segmented using energy information in the voiced speech segment) are considered for analysis in subsystem (4). The proposed four subsystems have been demonstrated to be effective and able to achieve competitive results in classifying different age and gender groups. To further improve the overall classification performance, weighted summation based fusion of these seven subsystems at the score level is demonstrated. Experiment results are reported on the development and test set of the 2010 Interspeech Paralinguistic Challenge aGender database. Compared to the SVM baseline system (3), which is the baseline system suggested by the challenge committee, the proposed fusion system achieves 5.6% absolute improvement in unweighted accuracy for the age task and 4.2% for the gender task on the development set. On the final test set, we obtain 3.1% and 3.8% absolute improvement, respectively.     

Prosody dependent speech recognition on radio news corpus of American English

Does prosody help word recognition? This paper proposes a novel probabilistic framework in which word and phoneme are dependent on prosody in a way that reduces word error rates (WER) relative to a prosody-independent recognizer with comparable parameter count. In the proposed prosody-dependent speech recognizer, word and phoneme models are conditioned on two important prosodic variables: the intonational phrase boundary and the pitch accent. An information-theoretic analysis is provided to show that prosody dependent acoustic and language modeling can increase the mutual information between the true word hypothesis and the acoustic observation by exciting the interaction between prosody dependent acoustic model and prosody dependent language model. Empirically, results indicate that the influence of these prosodic variables on allophonic models are mainly restricted to a small subset of distributions: the duration PDFs (modeled using an explicit duration hidden Markov model or EDHMM) and the acoustic-prosodic observation PDFs (normalized pitch frequency). Influence of prosody on cepstral features is limited to a subset of phonemes: for example, vowels may be influenced by both accent and phrase position, but phrase-initial and phrase-final consonants are independent of accent. Leveraging these results, effective prosody dependent allophonic models are built with minimal increase in parameter count. These prosody dependent speech recognizers are able to reduce word error rates by up to 11% relative to prosody independent recognizers with comparable parameter count, in experiments based on the prosodically-transcribed Boston Radio News corpus.     

Speaker identification based on normalized pitch frequency and Mel Frequency Cepstral Coefficients

This paper presents an efficient approach for automatic speaker identification based on cepstral features and the Normalized Pitch Frequency (NPF). Most relevant speaker identification methods adopt a cepstral strategy. Inclusion of the pitch frequency as a new feature in the speaker identification process is expected to enhance the speaker identification accuracy. In the proposed framework for speaker identification, a neural classifier with a single hidden layer is used. Different transform domains are investigated for reliable feature extraction from the speech signal. Moreover, a pre-processing noise reduction step, is used prior to the feature extraction process to enhance the performance of the speaker identification system. Simulation results prove that the NPF as a feature in speaker identification enhances the performance of the speaker identification system, especially with the Discrete Cosine Transform (DCT) and wavelet denoising pre-processing step.     

From English pitch accent detection to Mandarin stress detection, where is the difference?

Although English pitch accent detection has been studied extensively, there relatively a few works explore Mandarin stress detection. Moreover, the comparison and analysis between Mandarin stress detection and English pitch accent detection have not been touched for such counterpart tasks. In this paper, we discuss Mandarin stress detection and compare it with English pitch accent detection. The contributions of the paper are two aspects: one is that we use classifier combination method to detect Mandarin stress and English pitch accent by using acoustic, lexical and syntactic evidence. Our proposed method achieves better performance on both the Mandarin prosodic annotation corpus—ASCCD and the English prosodic annotation corpus—Boston University Radio News Corpus (BURNC) when compared with the baseline system. We also verify our proposed method on other prosodic annotation corpus and continuous speech corpus. The other is the feature analysis. Duration, pitch, energy and intensity features are compared for Mandarin stress detection and English pitch accent detection. Based on the analysis of prosodic annotation corpora, we also verify some linguistic conclusions.     

Artificially intelligent recognition of Arabic speaker using voice print-based local features

Local features for any pattern recognition system are based on the information extracted locally. In this paper, a local feature extraction technique was developed. This feature was extracted in the time–frequency plain by taking the moving average on the diagonal directions of the time–frequency plane. This feature captured the time–frequency events producing a unique pattern for each speaker that can be viewed as a voice print of the speaker. Hence, we referred to this technique as voice print-based local feature. The proposed feature was compared to other features including mel-frequency cepstral coefficient (MFCC) for speaker recognition using two different databases. One of the databases used in the comparison is a subset of an LDC database that consisted of two short sentences uttered by 182 speakers. The proposed feature attained 98.35% recognition rate compared to 96.7% for MFCC using the LDC subset.     

Neural network based feature transformation for emotion independent speaker identification

In this paper we are proposing neural network based feature transformation framework for developing emotion independent speaker identification system. Most of the present speaker recognition systems may not perform well during emotional environments. In real life, humans extensively express emotions during conversations for effectively conveying the messages. Therefore, in this work we propose the speaker recognition system, robust to variations in emotional moods of speakers. Neural network models are explored to transform the speaker specific spectral features from any specific emotion to neutral. In this work, we have considered eight emotions namely, Anger, Sad, Disgust, Fear, Happy, Neutral, Sarcastic and Surprise. The emotional databases developed in Hindi, Telugu and German are used in this work for analyzing the effect of proposed feature transformation on the performance of speaker identification system. In this work, spectral features are represented by mel-frequency cepstral coefficients, and speaker models are developed using Gaussian mixture models. Performance of the speaker identification system is analyzed with various feature mapping techniques. Results have demonstrated that the proposed neural network based feature transformation has improved the speaker identification performance by 20?%. Feature transformation at the syllable level has shown the better performance, compared to sentence level.     

提高MFCC鲁棒性的方法

在噪声环境中,基于笔记本电脑录音的情况下,采用特征参数加窗的方法,以提高系统的噪声鲁棒性.在 Matlab 环境下,建立了基于高斯混合模型 (GMM) 的说话人辨认系统,并进行实验.通过对多种窗口的正识率比较,发现对美尔倒谱(MFCC)高阶参数的加窗提升,可以改善系统的鲁棒性.实验结果表明,采用加窗后的系统识别率得到了明显改善.     

说话人识别技术中Mel倒谱参数改进算法的C程序实现

基于语音信号的频谱特性,本文对说话人识别技术中Mel倒谱参数做了改进,并通过Microsoft Visual C 6.0验证了在低信噪比时使用改进后的Mel倒谱参数可以提高说话人识别系统的正确识别率.     

Emotion recognition from speech using global and local prosodic features

In this paper, global and local prosodic features extracted from sentence, word and syllables are proposed for speech emotion or affect recognition. In this work, duration, pitch, and energy values are used to represent the prosodic information, for recognizing the emotions from speech. Global prosodic features represent the gross statistics such as mean, minimum, maximum, standard deviation, and slope of the prosodic contours. Local prosodic features represent the temporal dynamics in the prosody. In this work, global and local prosodic features are analyzed separately and in combination at different levels for the recognition of emotions. In this study, we have also explored the words and syllables at different positions (initial, middle, and final) separately, to analyze their contribution towards the recognition of emotions. In this paper, all the studies are carried out using simulated Telugu emotion speech corpus (IITKGP-SESC). These results are compared with the results of internationally known Berlin emotion speech corpus (Emo-DB). Support vector machines are used to develop the emotion recognition models. The results indicate that, the recognition performance using local prosodic features is better compared to the performance of global prosodic features. Words in the final position of the sentences, syllables in the final position of the words exhibit more emotion discriminative information compared to the words and syllables present in the other positions.     

Vocal fatigue induced by prolonged oral reading: Analysis and detection

This article uses prolonged oral reading corpora for various experiments to analyze and detect vocal fatigue. Vocal fatigue particularly concerns voice professionals, including teachers, telemarketing operators, users of automatic speech recognition technology and actors. In analyzing and detecting vocal fatigue, we focused our investigations on three main experiments: a prosodic analysis that can be compared to the results found in related work, a two-class Support Vector Machines (SVM) classifier into Fatigue and Non-Fatigue states using a large set of audio features and a comparison function that estimates the difference in fatigue level between two speech segments using a combination of multiple phoneme-based comparison functions. The experiments on prosodic analysis showed that vocal fatigue was not associated with an increase in fundamental frequency and voice intensity. A two-class SVM classifier using the Paralinguistic Challenge 2010 audio feature set gave an unweighted accuracy of 94.1% for the training set (10-fold cross-validation) and 68.2% for the test set. These results show that the phenomenon of vocal fatigue can be modeled and detected. The comparison function was assessed by detecting increased fatigue levels between two speech segments. The fatigue level detection performance in Equal Error Rate (EER) was 31% using all phonetic segments and yielded EER of 21% after filtering phonetic segments and 19% after filtering phonetic segments and cepstral features. These results show that some phonemes are more sensitive than others to vocal fatigue. These experiments show that the fatigued voice has specific characteristics for prolonged oral reading and suggest the feasibility of vocal fatigue detection.     

基于韵律特征的SVM说话人确认

提出了一种基于韵律特征和SVM的文本无关说话人确认系统。采用小波分析方法,从语音信号的MFCC、F0和能量轨迹中提取出超音段韵律特征,通过实验研究三者的韵律特征在特征层的最佳互补融合,得到信号的韵律特征PMFCCFE,用韵律特征的GMM均值超矢量作为参数训练目标话者的SVM模型,以更有效地区分目标话者和冒认话者。在NIST06 8side-1side数据库的实验表明,以短时倒谱参数的GMM-UBM系统为基准,超音段韵律特征的GMM-SVM系统的EER相对下降了57.9%,MinDCF相对下降了41.4%。