Statistical Approach for Voice Personality Transformation

A voice transformation method which changes the source speaker's utterances so as to sound similar to those of a target speaker is described. Speaker individuality transformation is achieved by altering the LPC cepstrum, average pitch period and average speaking rate. The main objective of the work involves building a nonlinear relationship between the parameters for the acoustical features of two speakers, based on a probabilistic model. The conversion rules involve the probabilistic classification and a cross correlation probability between the acoustic features of the two speakers. The parameters of the conversion rules are estimated by estimating the maximum likelihood of the training data. To obtain transformed speech signals which are perceptually closer to the target speaker's voice, prosody modification is also involved. Prosody modification is achieved by scaling excitation spectrum and time scale modification with appropriate modification factors. An evaluation by objective tests and informal listening tests clearly indicated the effectiveness of the proposed transformation method. We also confirmed that the proposed method leads to smoothly evolving spectral contours over time, which, from a perceptual standpoint, produced results that were superior to conventional vector quantization (VQ)-based methods     

Advances in phone-based modeling for automatic accent classification

It is suggested that algorithms capable of estimating and characterizing accent knowledge would provide valuable information in the development of more effective speech systems such as speech recognition, speaker identification, audio stream tagging in spoken document retrieval, channel monitoring, or voice conversion. Accent knowledge could be used for selection of alternative pronunciations in a lexicon, engage adaptation for acoustic modeling, or provide information for biasing a language model in large vocabulary speech recognition. In this paper, we propose a text-independent automatic accent classification system using phone-based models. Algorithm formulation begins with a series of experiments focused on capturing the spectral evolution information as potential accent sensitive cues. Alternative subspace representations using principal component analysis and linear discriminant analysis with projected trajectories are considered. Finally, an experimental study is performed to compare the spectral trajectory model framework to a traditional hidden Markov model recognition framework using an accent sensitive word corpus. System evaluation is performed using a corpus representing five English speaker groups with native American English, and English spoken with Mandarin Chinese, French, Thai, and Turkish accents for both male and female speakers.     

A novel method for voice conversion based on non-parallel corpus

This article puts forward a new algorithm for voice conversion which not only removes the necessity of parallel corpus in the training phase but also resolves the issue of insufficiency of the target speaker’s corpus. The proposed approach is based on one of the new voice conversion models utilizing classical LPC analysis-synthesis model combined with GMM. Through this algorithm, the conversion functions among vowels and demi-syllables are derived. We assumed that these functions are rather the same for different speakers if their genders, accents, and languages are alike. Therefore, we will be able to produce the demi-syllables with just having access to few sentences from the target speaker and forming the GMM for one of his/her vowels. The results from the appraisal of the proposed method for voice conversion clarifies that this method has the ability to efficiently realize the speech features of the target speaker. It can also provide results comparable to the ones obtained through the parallel-corpus-based approaches.     

On robustness of speech based biometric systems against voice conversion attack

Voice conversion (VC) approach, which morphs the voice of a source speaker to be perceived as spoken by a specified target speaker, can be intentionally used to deceive the speaker identification (SID) and speaker verification (SV) systems that use speech biometric. Voice conversion spoofing attacks to imitate a particular speaker pose potential threat to these kinds of systems. In this paper, we first present an experimental study to evaluate the robustness of such systems against voice conversion disguise. We use Gaussian mixture model (GMM) based SID systems, GMM with universal background model (GMM-UBM) based SV systems and GMM supervector with support vector machine (GMM-SVM) based SV systems for this. Voice conversion is conducted by using three different techniques: GMM based VC technique, weighted frequency warping (WFW) based conversion method and its variation, where energy correction is disabled (WFW⁻). Evaluation is done by using intra-gender and cross-gender voice conversions between fifty male and fifty female speakers taken from TIMIT database. The result is indicated by degradation in the percentage of correct identification (POC) score in SID systems and degradation in equal error rate (EER) in all SV systems. Experimental results show that the GMM-SVM SV systems are more resilient against voice conversion spoofing attacks than GMM-UBM SV systems and all SID and SV systems are most vulnerable towards GMM based conversion than WFW and WFW⁻ based conversion. From the results, it can also be said that, in general terms, all SID and SV systems are slightly more robust to voices converted through cross-gender conversion than intra-gender conversion. This work extended the study to find out the relationship between VC objective score and SV system performance in CMU ARCTIC database, which is a parallel corpus. The results of this experiment show an approach on quantifying objective score of voice conversion that can be related to the ability to spoof an SV system.     

自动构建基于电视剧字幕和剧本的日常会话基础标注库

真实对话数据量不足已经成为限制数据驱动的对话生成系统性能提升的主要因素,尤其是汉语语料。为了获得丰富的日常会话语料,可以利用字幕时间戳信息把英语电视剧的英文字幕及其对应的中文字幕进行同步,从而生成大量的汉英双语同步字幕。然后通过信息检索的方法把双语同步字幕的英文句子跟英语剧本的演员台词进行自动对齐,从而将剧本中的场景和说话者信息映射到双语字幕中,最后得到含有场景及说话者标注的汉英双语日常会话库。该文利用这种方法,自动构建了包含978 109对双语话语消息的接近人类日常会话的多轮会话数据库CEDAC。经过抽样分析,场景边界的标注准确率达到97.0%,而说话者的标注准确率也达到91.57%。该标注库为后续进行影视剧字幕说话者自动标注和多轮会话自动生成研究打下了很好的基础。     

Maximum likelihood stochastic transformation adaptation for medium and small data sets

Speaker adaptation is recognized as an essential part of today’s large-vocabulary automatic speech recognition systems. A family of techniques that has been extensively applied for limited adaptation data is transformation-based adaptation. In transformation-based adaptation we partition our parameter space in a set of classes, estimate a transform (usually linear) for each class and apply the same transform to all the components of the class. It is known, however, that additional gains can be made if we do not constrain the components of each class to use the same transform. In this paper two speaker adaptation algorithms are described. First, instead of estimating one linear transform for each class (as maximum likelihood linear regression (MLLR) does, for example) we estimate multiple linear transforms per class of models and a transform weights vector which is specific to each component (Gaussians in our case). This in effect means that each component receives its own transform without having to estimate each one of them independently. This scheme, termed maximum likelihood stochastic transformation (MLST) achieves a good trade-off between robustness and acoustic resolution. MLST is evaluated on the Wall Street Journal(WSJ) corpus for non-native speakers and it is shown that in the case of 40 adaptation sentences the algorithm outperforms MLLR by more than 13%. In the second half of this paper, we introduce a variant of the MLST designed to operate under sparsity of data. Since the majority of the adaptation parameters are the transformations, we estimate them on the training speakers and adapt to a new speaker by estimating the transform weights only. First we cluster the speakers in a number of sets and estimate the transformations on each cluster. The new speaker will use transformations from all clusters to perform adaptation. This method, termed basis transformation, can be seen as a speaker similarity scheme. Experimental results on the WSJ show that when basis transformation is cascaded with MLLR marginal gains can be obtained from MLLR only, for adaptation of native speakers.     

Multilingual recognition of non-native speech using acoustic model transformation and pronunciation modeling

This article presents an approach for the automatic recognition of non-native speech. Some non-native speakers tend to pronounce phonemes as they would in their native language. Model adaptation can improve the recognition rate for non-native speakers, but has difficulties dealing with pronunciation errors like phoneme insertions or substitutions. For these pronunciation mismatches, pronunciation modeling can make the recognition system more robust. Our approach is based on acoustic model transformation and pronunciation modeling for multiple non-native accents. For acoustic model transformation, two approaches are evaluated: MAP and model re-estimation. For pronunciation modeling, confusion rules (alternate pronunciations) are automatically extracted from a small non-native speech corpus. This paper presents a novel approach to introduce confusion rules in the recognition system which are automatically learned through pronunciation modelling. The modified HMM of a foreign spoken language phoneme includes its canonical pronunciation along with all the alternate non-native pronunciations, so that spoken language phonemes pronounced correctly by a non-native speaker could be recognized. We evaluate our approaches on the European project HIWIRE non-native corpus which contains English sentences pronounced by French, Italian, Greek and Spanish speakers. Two cases are studied: the native language of the test speaker is either known or unknown. Our approach gives better recognition results than the classical acoustic adaptation of HMM when the foreign origin of the speaker is known. We obtain 22% WER reduction compared to the reference system.     

Environment adaptation for robust speaker verification by cascading maximum likelihood linear regression and reinforced learning

In speaker verification over public telephone networks, utterances can be obtained from different types of handsets. Different handsets may introduce different degrees of distortion to the speech signals. This paper attempts to combine a handset selector with (1) handset-specific transformations, (2) reinforced learning, and (3) stochastic feature transformation to reduce the effect caused by the acoustic distortion. Specifically, during training, the clean speaker models and background models are firstly transformed by MLLR-based handset-specific transformations using a small amount of distorted speech data. Then reinforced learning is applied to adapt the transformed models to handset-dependent speaker models and handset-dependent background models using stochastically transformed speaker patterns. During a verification session, a GMM-based handset classifier is used to identify the most likely handset used by the claimant; then the corresponding handset-dependent speaker and background model pairs are used for verification. Experimental results based on 150 speakers of the HTIMIT corpus show that environment adaptation based on the combination of MLLR, reinforced learning and feature transformation outperforms CMS, Hnorm, Tnorm, and speaker model synthesis.     

Improving robustness of MLLR adaptation with speaker-clustered regression class trees

We introduce a strategy for modeling speaker variability in speaker adaptation based on maximum likelihood linear regression (MLLR). The approach uses a speaker-clustering procedure that models speaker variability by partitioning a large corpus of speakers in the eigenspace of their MLLR transformations and learning cluster-specific regression class tree structures. We present experiments showing that choosing the appropriate regression class tree structure for speakers leads to a significant reduction in overall word error rates in automatic speech recognition systems. To realize these gains in unsupervised adaptation, we describe an algorithm that produces a linear combination of MLLR transformations from cluster-specific trees using weights estimated by maximizing the likelihood of a speaker’s adaptation data. This algorithm produces small improvements in overall recognition performance across a range of tasks for both English and Mandarin. More significantly, distributional analysis shows that it reduces the number of speakers with performance loss due to adaptation across a range of adaptation data sizes and word error rates.     

Quality-enhanced voice morphing using maximum likelihood transformations

Voice morphing is a technique for modifying a source speaker's speech to sound as if it was spoken by some designated target speaker. The core process in a voice morphing system is the transformation of the spectral envelope of the source speaker to match that of the target speaker and linear transformations estimated from time-aligned parallel training data are commonly used to achieve this. However, the naive application of envelope transformation combined with the necessary pitch and duration modifications will result in noticeable artifacts. This paper studies the linear transformation approach to voice morphing and investigates these two specific issues. First, a general maximum likelihood framework is proposed for transform estimation which avoids the need for parallel training data inherent in conventional least mean square approaches. Second, the main causes of artifacts are identified as being due to glottal coupling, unnatural phase dispersion and the high spectral variance of unvoiced sounds, and compensation techniques are developed to mitigate these. The resulting voice morphing system is evaluated using both subjective and objective measures. These tests show that the proposed approaches are capable of effectively transforming speaker identity whilst maintaining high quality. Furthermore, they do not require carefully prepared parallel training data.     

Expression of affect in spontaneous speech: Acoustic correlates and automatic detection of irritation and resignation

The majority of previous studies on vocal expression have been conducted on posed expressions. In contrast, we utilized a large corpus of authentic affective speech recorded from real-life voice controlled telephone services. Listeners rated a selection of 200 utterances from this corpus with regard to level of perceived irritation, resignation, neutrality, and emotion intensity. The selected utterances came from 64 different speakers who each provided both neutral and affective stimuli. All utterances were further automatically analyzed regarding a comprehensive set of acoustic measures related to F0, intensity, formants, voice source, and temporal characteristics of speech. Results first showed that several significant acoustic differences were found between utterances classified as neutral and utterances classified as irritated or resigned using a within-persons design. Second, listeners’ ratings on each scale were associated with several acoustic measures. In general the acoustic correlates of irritation, resignation, and emotion intensity were similar to previous findings obtained with posed expressions, though the effect sizes were smaller for the authentic expressions. Third, automatic classification (using LDA classifiers both with and without speaker adaptation) of irritation, resignation, and neutral performed at a level comparable to human performance, though human listeners and machines did not necessarily classify individual utterances similarly. Fourth, clearly perceived exemplars of irritation and resignation were rare in our corpus. These findings were discussed in relation to future research.     

Voice conversion using General Regression Neural Network

The objective of voice conversion system is to formulate the mapping function which can transform the source speaker characteristics to that of the target speaker. In this paper, we propose the General Regression Neural Network (GRNN) based model for voice conversion. It is a single pass learning network that makes the training procedure fast and comparatively less time consuming. The proposed system uses the shape of the vocal tract, the shape of the glottal pulse (excitation signal) and long term prosodic features to carry out the voice conversion task. In this paper, the shape of the vocal tract and the shape of source excitation of a particular speaker are represented using Line Spectral Frequencies (LSFs) and Linear Prediction (LP) residual respectively. GRNN is used to obtain the mapping function between the source and target speakers. The direct transformation of the time domain residual using Artificial Neural Network (ANN) causes phase change and generates artifacts in consecutive frames. In order to alleviate it, wavelet packet decomposed coefficients are used to characterize the excitation of the speech signal. The long term prosodic parameters namely, pitch contour (intonation) and the energy profile of the test signal are also modified in relation to that of the target (desired) speaker using the baseline method. The relative performances of the proposed model are compared to voice conversion system based on the state of the art RBF and GMM models using objective and subjective evaluation measures. The evaluation measures show that the proposed GRNN based voice conversion system performs slightly better than the state of the art models.     

Nonlinear normalization of input patterns to speaker variability in speech recognition neural networks

The issue of input variability resulting from speaker changes is one of the most crucial factors influencing the effectiveness of speech recognition systems. A solution to this problem is adaptation or normalization of the input, in a way that all the parameters of the input representation are adapted to that of a single speaker, and a kind of normalization is applied to the input pattern against the speaker changes, before recognition. This paper proposes three such methods in which some effects of the speaker changes influencing speech recognition process is compensated. In all three methods, a feed-forward neural network is first trained for mapping the input into codes representing the phonetic classes and speakers. Then, among the 71 speakers used in training, the one who is showing the highest percentage of phone recognition accuracy is selected as the reference speaker so that the representation parameters of the other speakers are converted to the corresponding speech uttered by him. In the first method, the error back-propagation algorithm is used for finding the optimal point of every decision region relating to each phone of each speaker in the input space for all the phones and all the speakers. The distances between these points and the corresponding points related to the reference speaker are employed for offsetting the speaker change effects and the adaptation of the input signal to the reference speaker. In the second method, using the error back-propagation algorithm and maintaining the reference speaker data as the desirable speaker output, we correct all the speech signal frames, i.e., the train and the test datasets, so that they coincide with the corresponding speech of the reference speaker. In the third method, another feed-forward neural network is applied inversely for mapping the phonetic classes and speaker information to the input representation. The phonetic output retrieved from the direct network along with the reference speaker data are given to the inverse network. Using this information, the inverse network yields an estimation of the input representation adapted to the reference speaker. In all three methods, the final speech recognition model is trained using the adapted training data, and is tested by the adapted testing data. Implementing these methods and combining the final network results with un-adapted network based on the highest confidence level, an increase of 2.1, 2.6 and 3% in phone recognition accuracy on the clean speech is obtained from the three methods, respectively.     

A modified HME architecture for text-dependent speakeridentification

A modified hierarchical mixtures of experts (HME) architecture is presented for text-dependent speaker identification. A new gating network is introduced to the original HME architecture for the use of instantaneous and transitional spectral information in text-dependent speaker identification. The statistical model underlying the proposed architecture is presented and learning is treated as a maximum likelihood problem; in particular, an expectation-maximization (EM) algorithm is also proposed for adjusting the parameters of the proposed architecture. An evaluation has been carried out using a database of isolated digit utterances by 10 male speakers. Experimental results demonstrate that the proposed architecture outperforms the original HME architecture in text-dependent speaker identification.     

VoCMex: a voice corpus in Mexican Spanish for research in speaker recognition

Voice corpus is an essential element for automatic speaker recognition systems. In order for a corpus to be useful in recognition tasks, it must contain recordings from several speakers pronouncing phonetically balanced utterances; recorded through several sessions using different recording media. This work shows the methodology, development and evaluation of a Mexican Spanish Corpus referred as to VoCMex, which is aimed to support research on speaker recognition. It contains telephone and microphone recordings of 20 male and 13 female speakers, obtained through three sessions. In order to validate the usefulness of the corpus, a speaker identification system was developed and the recognition results were similar compared against those obtained using a known voice corpus.     

基于话者无关模型的说话人转换方法

提出一种基于话者无关模型的说话人转换方法.考虑到音素信息共同存在于所有说话人的语音中,假设存在一个可以用高斯混合模型来描述的话者无关空间,且可用分段线性变换来描述该空间到各说话人相关空间之间的映射关系.在一个多说话人的数据库上,用话者自适应训练算法来训练模型,并在转换阶段使用源目标说话人空间到话者无关空间的变换关系来构造源与目标之间的特征变换关系,快速、灵活的构造说话人转换系统.通过主观测听实验来验证该算法相对于传统的基于话者相关模型方法的优点.     

Rapid Speaker Adaptation Using Clustered Maximum-Likelihood Linear Basis With Sparse Training Data

Speaker space-based adaptation methods for automatic speech recognition have been shown to provide significant performance improvements for tasks where only a few seconds of adaptation speech is available. However, these techniques are not widely used in practical applications because they require large amounts of speaker-dependent training data and large amounts of computer memory. The authors propose a robust, low-complexity technique within this general class that has been shown to reduce word error rate, reduce the large storage requirements associated with speaker space approaches, and eliminate the need for large numbers of utterances per speaker in training. The technique is based on representing speakers as a linear combination of clustered linear basis vectors and a procedure is presented for maximum-likelihood estimation of these vectors from training data. Significant word error rate reduction was obtained using these methods relative to speaker independent performance for the Resource Management and Wall Street Journal task domains.     

Voice comparison between smokers and non-smokers using HMM speech recognition system

Automatic speech recognition is a technology that allows a computer to transcribe in real time spoken words into readable text. In this work an HMM automatic speech recognition system was created to detect smoker speaker. This research project is carried out using Amazigh language for comparison of the voice of normal persons to smokers one. To achieve this goal, two experiments were performed, the first one to test the performance of the system for non-smokers for different parameters. The second experiment concern smokers speakers. The corpus used in this system is collected from two groups of speaker, non-smokers and smokers native Morocan tarifit speakers aged between 25 and 55 years. Our experimental results show that we can use our system to make diagnostic for smoking people and confirm that a speaker is smoker when the observed recognition rate is below 50%.     

Robust processing techniques for voice conversion

Differences in speaker characteristics, recording conditions, and signal processing algorithms affect output quality in voice conversion systems. This study focuses on formulating robust techniques for a codebook mapping based voice conversion algorithm. Three different methods are used to improve voice conversion performance: confidence measures, pre-emphasis, and spectral equalization. Analysis is performed for each method and the implementation details are discussed. The first method employs confidence measures in the training stage to eliminate problematic pairs of source and target speech units that might result from possible misalignments, speaking style differences or pronunciation variations. Four confidence measures are developed based on the spectral distance, fundamental frequency (f₀) distance, energy distance, and duration distance between the source and target speech units. The second method focuses on the importance of pre-emphasis in line-spectral frequency (LSF) based vocal tract modeling and transformation. The last method, spectral equalization, is aimed at reducing the differences in the source and target long-term spectra when the source and target recording conditions are significantly different. The voice conversion algorithm that employs the proposed techniques is compared with the baseline voice conversion algorithm with objective tests as well as three subjective listening tests. First, similarity to the target voice is evaluated in a subjective listening test and it is shown that the proposed algorithm improves similarity to the target voice by 23.0%. An ABX test is performed and the proposed algorithm is preferred over the baseline algorithm by 76.4%. In the third test, the two algorithms are compared in terms of the subjective quality of the voice conversion output. The proposed algorithm improves the subjective output quality by 46.8% in terms of mean opinion score (MOS).     

Self-learning speaker identification for enhanced speech recognition

A novel approach for joint speaker identification and speech recognition is presented in this article. Unsupervised speaker tracking and automatic adaptation of the human-computer interface is achieved by the interaction of speaker identification, speech recognition and speaker adaptation for a limited number of recurring users. Together with a technique for efficient information retrieval a compact modeling of speech and speaker characteristics is presented. Applying speaker specific profiles allows speech recognition to take individual speech characteristics into consideration to achieve higher recognition rates. Speaker profiles are initialized and continuously adapted by a balanced strategy of short-term and long-term speaker adaptation combined with robust speaker identification. Different users can be tracked by the resulting self-learning speech controlled system. Only a very short enrollment of each speaker is required. Subsequent utterances are used for unsupervised adaptation resulting in continuously improved speech recognition rates. Additionally, the detection of unknown speakers is examined under the objective to avoid the requirement to train new speaker profiles explicitly. The speech controlled system presented here is suitable for in-car applications, e.g. speech controlled navigation, hands-free telephony or infotainment systems, on embedded devices. Results are presented for a subset of the SPEECON database. The results validate the benefit of the speaker adaptation scheme and the unified modeling in terms of speaker identification and speech recognition rates.