Speaker Adaptation Method Research Articles

Comparing Speaker Adaptation Methods for Visual Speech Recognition for Continuous Spanish

Visual speech recognition (VSR) is a challenging task that aims to interpret speech based solely on lip movements. However, although remarkable results have recently been reached in the field, this task remains an open research problem due to different challenges, such as visual ambiguities, the intra-personal variability among speakers, and the complex modeling of silence. Nonetheless, these challenges can be alleviated when the task is approached from a speaker-dependent perspective. Our work focuses on the adaptation of end-to-end VSR systems to a specific speaker. Hence, we propose two different adaptation methods based on the conventional fine-tuning technique, the so-called Adapters. We conduct a comparative study in terms of performance while considering different deployment aspects such as training time and storage cost. Results on the Spanish LIP-RTVE database show that both methods are able to obtain recognition rates comparable to the state of the art, even when only a limited amount of training data is available. Although it incurs a deterioration in performance, the Adapters-based method presents a more scalable and efficient solution, significantly reducing the training time and storage cost by up to 80%.

Chinese Speech Recognition System Based on Neural Network Acoustic Network Model

In order to understand the Chinese speech recognition system based on network models, the author proposes a research on a Chinese speech recognition system based on neural network acoustic network models. The author first analyzed a speech recognition system based on neural networks. The system first performs sampling and filtering, pre weighting, signal framing, and endpoint detection operations on the speech signal, extracts the LPC, LPCC, and MFCC values of the preprocessed data, improves the MFCC values, and trains and constructs a neural network model, secondly, in view of the problem that the accuracy of System identification recognition declines when there is a large difference between the target Chinese speech tested and the speaker speech of training data in the speech recognition system, a speaker language adaptation SA (Speaker Adaptation) method based on deep neural network (DNN) is proposed. It is speaker adaptation in the feature space. By adding speaker identity vector information to the DNN Acoustic model, speaker difference information in the feature is removed to reduce the impact of speaker difference and retain semantic information. Finally, the experimental results on TEDLIUM open source dataset show that when the features of this method are fbank and fMLLR respectively, the system word error rate WER (Word Error Rate) is improved by 7.8% and 6.8% compared with the baseline neural network DNN Acoustic model. The reasonable selection of speech features can not only improve the training efficiency of the recognition model, but also effectively improve the recognition accuracy of the model, thereby improving recognition performance.

Meta-TTS: Meta-Learning for Few-Shot Speaker Adaptive Text-to-Speech

Personalizing a speech synthesis system is a highly desired application, where the system can generate speech with the user’s voice with rare enrolled recordings. There are two main approaches to build such a system in recent works: speaker adaptation and speaker encoding. On the one hand, speaker adaptation methods fine-tune a trained multi-speaker text-to-speech (TTS) model with few enrolled samples. However, they require at least thousands of fine-tuning steps for high-quality adaptation, making it hard to apply on devices. On the other hand, speaker encoding methods encode enrollment utterances into a speaker embedding. The trained TTS model can synthesize the user’s speech conditioned on the corresponding speaker embedding. Nevertheless, the speaker encoder suffers from the generalization gap between the seen and unseen speakers. In this paper, we propose applying a meta-learning algorithm to the speaker adaptation method. More specifically, we use Model Agnostic Meta-Learning (MAML) as the training algorithm of a multi-speaker TTS model, which aims to find a great meta-initialization to adapt the model to any few-shot speaker adaptation tasks quickly. Therefore, we can also adapt the meta-trained TTS model to unseen speakers efficiently. Our experiments compare the proposed method (Meta-TTS) with two baselines: a speaker adaptation method baseline and a speaker encoding method baseline. The evaluation results show that Meta-TTS can synthesize high speaker-similarity speech from few enrollment samples with fewer adaptation steps than the speaker adaptation baseline and outperforms the speaker encoding baseline under the same training scheme. When the speaker encoder of the baseline is pre-trained with extra 8371 speakers of data, Meta-TTS can still outperform the baseline on LibriTTS dataset and achieve comparable results on VCTK dataset.

CampNet: Context-Aware Mask Prediction for End-to-End Text-Based Speech Editing

The text-based speech editor allows the editing of speech through intuitive cutting, copying, and pasting operations to speed up the process of editing speech. However, the major drawback of current systems is that edited speech often sounds unnatural due to cut-copy-paste operation. In addition, it is not obvious how to synthesize records according to a new word not appearing in the transcript, which often needs the help of text-to-speech (TTS) and voice conversion (VC) technology at the same time. This paper first proposes a novel end-to-end text-based speech editing method called context-aware mask prediction network (CampNet). The model can simulate the text-based speech editing process by randomly masking part of speech and then predicting the masked region by sensing the speech context. It can solve unnatural prosody in the edited region and synthesize the speech corresponding to the unseen words in the transcript. Secondly, for the possible operation of text-based speech editing, we design three text-based operations based on CampNet: deletion, insertion, and replacement. These operations can cover various situations of speech editing. Thirdly, to synthesize the speech corresponding to long text in insertion and replacement operations, a word-level autoregressive generation method is proposed, which can synthesize the speech of arbitrary length text. Fourthly, we propose a speaker adaptation method using only one sentence for CampNet and explore the ability of few-shot learning based on CampNet, which provides a new idea for speech forgery tasks. The subjective and objective experiments<xref ref-type="fn" rid="fn1">¹</xref><fn id="fn1"><label>¹</label> Examples of generated speech can be found at <uri>https://hairuo55.github.io/CampNet</uri>. </fn> on VCTK and LibriTTS datasets show that the speech editing results based on CampNet are better than TTS technology, manual editing, and VoCo method (the combination of TTS and VC). We also conduct detailed ablation experiments to explore the effect of the CampNet structure on its performance. Finally, the experiment shows that speaker adaptation with only one sentence can further improve the naturalness of speech editing for one-shot learning.

A formant modification method for improved ASR of children’s speech

Differences in acoustic characteristics between children’s and adults’ speech degrade performance of automatic speech recognition systems when systems trained using adults’ speech are used to recognize children’s speech. This performance degradation is due to the acoustic mismatch between training and testing. One of the main sources of the acoustic mismatch is the difference in vocal tract resonances (formant frequencies) between adult and child speakers. The present study aims to reduce the mismatch in formant frequencies by modifying formants of children’s speech to better correspond to formants of adults’ speech. This is carried out by warping the linear prediction (LP) spectrum computed from children’s speech. The warped LP spectra computed in a frame-based manner from children’s speech are used with the corresponding LP residuals to synthesize speech whose formant structure is closer to that of adults’ speech. When used in testing of an ASR system trained using adults’ speech, the warping reduces the spectral mismatch in speech between training and testing and improves the system performance in recognition of children’s speech. Experiments were conducted using narrowband (8 kHz) and wideband (16 kHz) speech of adult and child speakers from the WSJCAM0 and PF_STAR databases, respectively, and by recognizing children’s speech using acoustic models trained with adults’ speech. The proposed method gave relative improvements of 24% and 11% for the DNN and TDNN acoustic models, respectively, for narrowband speech. For wideband speech, the technique gave relative improvements of 27% and 13% for the DNN and TDNN acoustic models, respectively. The performance of the proposed method was also compared to two speaker adaptation methods: vocal tract length normalization (VTLN) and speaking rate adaptation (SRA). This comparison showed the best recognition performance for the proposed method. We also combined the proposed method with VTLN and SRA, and found that the combined method gave a further reduction in WER. Moreover, our experiments carried out for noisy speech using various types of additive noise and signal-to-noise ratios showed that the proposed method performs well also for degraded speech.

Survey of Speaker Adaptation Methods in Speech Recognition

Survey of Speaker Adaptation Methods in Speech Recognition

Rapid Speaker Adaptation Based on Combination of KPCA and Latent Variable Model

Speaker adaptation is implemented in order to shift the speaker-independent model closer to the new speaker speech characteristics to improve the speech recognition performance. The kernel eigenspace-based speaker adaptation methods provide satisfactory performance using only a small amount of adaptation data. In such adaptation methods, kernel principal component analysis (KPCA) is applied to the training speaker space in order to create kernel eigenspace. Then, the adapted acoustic model to the new user is calculated in that space. One limitation of KPCA is its inability to define a precise pre-image of the model adapted in the kernel eigenspace, back to the speaker space. Therefore, a huge amount of computations is required to perform adaptation. The previously developed solutions for calculation of an approximate pre-image of the adapted model do not necessarily lead to the optimal conditions. Therefore, in this paper, we propose an efficient solution for this problem to construct more reliable pre-image of the adapted model in the speaker space. For this purpose, we benefit from the latent variable model to define a probabilistic model for description of the applied mapping between the kernel eigenspace and the speaker space. The experiments were conducted on two speech databases: FARSDAT, a Persian, and TIMIT, an English speech database. Implementing a typical HMM-based automatic speech recognition system, it was verified that the proposed method, utilizing about three seconds of adaptation data, achieves up to 4.4% and 7.6% relative phoneme recognition accuracy rate over the speaker-independent model on FARSDAT and TIMIT, respectively. Moreover, the proposed approach demonstrated superior performance compared to the other kernel eigenspace-based adaptation methods.

Speaker Adaptation Using i-Vector Based Clustering

We propose a novel speaker adaptation method using acoustic model clustering. The similarity of different speakers is defined by the cosine distance between their i-vectors (intermediate vectors), and various efficient clustering algorithms are applied to obtain a number of speaker subsets with different characteristics. The speaker-independent model is then retrained with the training data of the individual speaker subsets grouped by the clustering results, and an unknown speech is recognized by the retrained model of the closest cluster. The proposed method is applied to a large-scale speech recognition system implemented by a hybrid hidden Markov model and deep neural network framework. An experiment was conducted to evaluate the word error rates using Resource Management database. When the proposed speaker adaptation method using i-vector based clustering was applied, the performance, as compared to that of the conventional speaker-independent speech recognition model, was improved relatively by as much as 12.2% for the conventional fully neural network, and by as much as 10.5% for the bidirectional long short-term memory.

Online Speaker Adaptation Using Memory-Aware Networks for Speech Recognition

In our previous work, we introduced our attention-based speaker adaptation method, which has been proved to be an efficient online speaker adaptation method for real-time speech recognition. In this paper, we present a more complete framework of this method named memory-aware networks, which consists of the main network, the memory module, the attention module and the connection module. A gate mechanism and a multiple-connections strategy are presented to connect the memory with the main network in order to take full advantage of the memory. An auxiliary speaker classification task is provided to improve the accuracy of the attention module. The fixed-size ordinally forgetting encoding method is used together with average pooling to gather both short-term and long-term information. Furthermore, instead of only using traditional speaker embeddings such as i-vectors or d-vectors as the memory, we design a new form of memory called residual vectors, which can represent different pronunciation habits. Experiments on both the Switchboard and AISHELL-2 tasks show that our method can perform online speaker adaptation very well with no additional adaptation data and with only a relative 3% increase in decoding computation complexity. Under the cross-entropy criterion, our method achieves a relative word error rate reduction of 9.4% and 8.3% compared to that of the speaker-independent model on the Switchboard task and the AISHELL-2 task, respectively, and approximately 7.0% compared to that of the traditional d-vector-based speaker adaptation method.

A voice command detection system for aerospace applications

Nowadays, according to ever-increasing volumes of audio content, audio processing is a vital need. In the aerospace field, voice commands could be used instead of data commands in order to speed up the command transmission, help crewmembers to complete their tasks by allowing hands-free control of supplemental equipment and as a redundant system for increasing the reliability of command transmission. In this paper, a voice command detection (VCD) framework is proposed for aerospace applications, which decodes the voice commands to comprehensible and executable commands, in an acceptable speed with a low false alarm rate. The framework is mainly based on a keyword spotting method, which extracts some pre-defined target keywords from the input voice commands. The mentioned keywords are input arguments to the proposed rule-based language model (LM). The rule-based LM decodes the voice commands based on the input keywords and their locations. Two keyword spotters are trained and used in the VCD system. The phone-based keyword spotter is trained on TIMIT database. Then, speaker adaptation methods are exploited to modify the parameters of the trained models using non-native speaker utterances. The word-based keyword spotter is trained on a database prepared and specialized for aerospace applications. The experimental results show that the word-based VCD system decodes the voice commands with true detection rate equal to 88% and false alarm rate equal to 12%, in average. Additionally, using speaker adaptation methods in the phone-based VCD system improves the true detection and false alarm rates about 21% and 21%, respectively.

Personalized Spontaneous Speech Synthesis Using a Small-Sized Unsegmented Semispontaneous Speech

A systematic approach is proposed to synthesizing personalized spontaneous speech using a small-sized unsegmented speech corpus of the target speaker. First, an automatic segmentation algorithm is employed to segment and label a collected semispontaneous speech corpus of the target speaker. Then, a pretrained average voice model is adapted to the voice model of the target speaker by using the segmented data. A postfilter based on modulation spectrum is adopted to further improve the speaker similarity of the synthesized speech as well as alleviate the over-smoothing problem of the synthesized speech. For generating spontaneous speech, a smoothing method applied at the prosodic word level is proposed to improve speech fluency. For objective evaluation on spontaneous speech segmentation, the segmentation accuracy of the proposed method is superior to that of Viterbi-based forced alignment. The results of subjective listening test also show that the proposed method can improve the spontaneity and speaker similarity of the synthesized speech compared to the maximum likelihood linear regression based speaker adaptation method.

Regularized Speaker Adaptation of KL-HMM for Dysarthric Speech Recognition.

This paper addresses the problem of recognizing the speech uttered by patients with dysarthria, which is a motor speech disorder impeding the physical production of speech. Patients with dysarthria have articulatory limitation, and therefore, they often have trouble in pronouncing certain sounds, resulting in undesirable phonetic variation. Modern automatic speech recognition systems designed for regular speakers are ineffective for dysarthric sufferers due to the phonetic variation. To capture the phonetic variation, Kullback-Leibler divergence-based hidden Markov model (KL-HMM) is adopted, where the emission probability of state is parameterized by a categorical distribution using phoneme posterior probabilities obtained from a deep neural network-based acoustic model. To further reflect speaker-specific phonetic variation patterns, a speaker adaptation method based on a combination of L2 regularization and confusion-reducing regularization, which can enhance discriminability between categorical distributions of the KL-HMM states while preserving speaker-specific information is proposed. Evaluation of the proposed speaker adaptation method on a database of several hundred words for 30 speakers consisting of 12 mildly dysarthric, 8 moderately dysarthric, and 10 non-dysarthric control speakers showed that the proposed approach significantly outperformed the conventional deep neural network-based speaker adapted system on dysarthric as well as non-dysarthric speech.

Toward growing modular deep neural networks for continuous speech recognition

The performance drop of typical automatic speech recognition systems in real applications is related to their not properly designed structure and training procedure. In this article, a growing modular deep neural network (MDNN) for speech recognition is introduced. According to its structure, this network is pre-trained in a special manner. The ability of the MDNN to grow enables it to implement spatiotemporal information of the frame sequences at the input and their labels at the output layer at the same time. The trained network with such a double spatiotemporal (DST) structure has learned valid phonetic sequences subspace. Therefore, it can filter out invalid output sequences in its own structure. In order to improve the proposed network performance in speaker variations, two speaker adaptation methods are also presented in this work. In these adaptation methods, the network trains how to move distorted input representations nonlinearly to their optimal positions or to adapt itself based on the input information. To evaluate the proposed MDNN structure and its modified versions, two Persian speech datasets are used: FARSDAT and Large FARSDAT. As there is no frame-level transcription for large vocabulary speech datasets, a semi-supervised learning algorithm is explored to train MDNN on Large FARSDAT. Experimental results on FARSDAT verify that implementing the DST structure besides speaker adaptation methods achieves up to 7.3 and 10.6 % absolute phone accuracy rate improvement over the MDNN and typical hidden Markov model, respectively. Likewise, semi-supervised training of the grown MDNN on Large FARSDAT improves its recognition performance up to 5 %.

Scaled norm-based Euclidean projection for sparse speaker adaptation

To reduce data storage for speaker adaptive (SA) models, in our previous work, we proposed a sparse speaker adaptation method which can efficiently reduce the number of adapted parameters by using Euclidean projection onto the L 1-ball (EPL1) while maintaining recognition performance comparable to maximum a posteriori (MAP) adaptation. In the EPL1-based sparse speaker adaptation framework, however, the adapted Gaussian mean vectors are mostly concentrated on dimensions having large variances because of assuming unit variance for all dimensions. To make EPL1 more flexible, in this paper, we propose scaled norm-based Euclidean projection (SNEP) which can consider dimension-specific variances. By using SNEP, we also propose a new sparse speaker adaptation method which can consider the variances of a speaker-independent model. Our experiments show that the adapted components of mean vectors are evenly distributed in all dimensions, and we can obtain sparsely adapted models with no loss of phone recognition performance from the proposed method compared with MAP adaptation.

Speaker Adaptation of Hybrid NN/HMM Model for Speech Recognition Based on Singular Value Decomposition

Recently several speaker adaptation methods have been proposed for deep neural network (DNN) in many large vocabulary continuous speech recognition (LVCSR) tasks. However, only a few methods rely on tuning the connection weights in trained DNNs directly to optimize system performance since it is very prone to over-fitting especially when some class labels are missing in the adaptation data. In this paper, we propose a new speaker adaptation method for the hybrid NN/HMM speech recognition model based on singular value decomposition (SVD). We apply SVD on the weight matrices in trained DNNs and then tune rectangular diagonal matrices with the adaptation data. This alleviates the over-fitting problem via updating the weight matrices slightly by only modifying the singular values. We evaluate the proposed adaptation method in two standard speech recognition tasks, namely TIMIT phone recognition and large vocabulary speech recognition in the Switchboard task. Experimental results have shown that it is effective to adapt large DNN models using only a small amount of adaptation data. For example, recognition results in the Switchboard task have shown that the proposed SVD-based adaptation method may achieve up to 3-6 % relative error reduction using only a few dozens of adaptation utterances per speaker.

Basis-Based Speaker Adaptation Using Partitioned HMM Mean Parameters of Training Speaker Models

This paper presents the basis-based speaker adaptation method that includes approaches using principal component analysis (PCA) and two-dimensional PCA (2DPCA). The proposed method partitions the hidden Markov model (HMM) mean vectors of training models into subvectors of smaller dimension. Consequently, the sample covariance matrix computed using the partitioned HMM mean vectors has various dimensions according to the dimension of the subvectors. From the eigen-decomposition of the sample covariance matrix, basis vectors are constructed. Thus, the dimension of basis vectors varies according to the dimension of the sample covariance matrix, and the proposed method includes PCA and 2DPCA-based approaches. We present the adaptation equation in both the maximum likelihood (ML) and maximum a posteriori (MAP) frameworks. We perform continuous speech recognition experiments using the Wall Street Journal (WSJ) corpus. The results show that the model with basis vectors whose dimensions are between those of PCA and 2DPCA-based approaches shows good overall performance. The proposed approach in the MAP framework shows additional performance improvement over the ML counterpart when the number of adaptation parameters is large but the amount of available adaptation data is small. Furthermore, the performance of the approach in the MAP framework approach is less sensitive to the choice of model order than the ML counterpart.

Joint speaker and environment adaptation using TensorVoice for robust speech recognition

We present an adaptation of a hidden Markov model (HMM)-based automatic speech recognition system to the target speaker and noise environment. Given HMMs built from various speakers and noise conditions, we build tensorvoices that capture the interaction between the speaker and noise by using a tensor decomposition. We express the updated model for the target speaker and noise environment as a product of the tensorvoices and two weight vectors, one each for the speaker and noise. An iterative algorithm is presented to determine the weight vectors in the maximum likelihood (ML) framework. With the use of separate weight vectors, the tensorvoice approach can adapt to the target speaker and noise environment differentially, whereas the eigenvoice approach, which is based on a matrix decomposition technique, cannot differentially adapt to those two factors. In supervised adaptation tests using the AURORA4 corpus, the relative improvement of performance obtained by the tensorvoice method over the eigenvoice method is approximately 10% on average for adaptation data of 6–24s in length, and the relative improvement of performance obtained by the tensorvoice method over the maximum likelihood linear regression (MLLR) method is approximately 5.4% on average for adaptation data of 6–18s in length. Therefore, the tensorvoice approach is an efficient method for speaker and noise adaptation.

Speaker-adaptive speech recognition using speaker diarization for improved transcription of large spoken archives

This paper deals with speaker-adaptive speech recognition for large spoken archives. The goal is to improve the recognition accuracy of an automatic speech recognition (ASR) system that is being deployed for transcription of a large archive of Czech radio. This archive represents a significant part of Czech cultural heritage, as it contains recordings covering 90years of broadcasting. A large portion of these documents (100,000h) is to be transcribed and made public for browsing. To improve the transcription results, an efficient speaker-adaptive scheme is proposed. The scheme is based on integration of speaker diarization and adaptation methods and is designed to achieve a low Real-Time Factor (RTF) of the entire adaptation process, because the archive’s size is enormous. It thus employs just two decoding passes, where the first one is carried out using the lexicon with a reduced number of items. Moreover, the transcripts from the first pass serve not only for adaptation, but also as the input to the speaker diarization module, which employs two-stage clustering. The output of diarization is then utilized for a cluster-based unsupervised Speaker Adaptation (SA) approach that also utilizes information based on the gender of each individual speaker. Presented experimental results on various types of programs show that our adaptation scheme yields a significant Word Error Rate (WER) reduction from 22.24% to 18.85% over the Speaker Independent (SI) system while operating at a reasonable RTF.

Rapid speaker adaptation using compressive sensing

Speaker-space-based speaker adaptation methods can obtain good performance even if the amount of adaptation data is limited. However, it is difficult to determine the optimal dimension and basis vectors of the subspace for a particular unknown speaker. Conventional methods, such as eigenvoice (EV) and reference speaker weighting (RSW), can only obtain a sub-optimal speaker subspace. In this paper, we present a new speaker-space-based speaker adaptation framework using compressive sensing. The mean vectors of all mixture components of a conventional Gaussian-Mixture-Model-Hidden-Markov-Model (GMM-HMM)-based speech recognition system are concatenated to form a supervector. The speaker adaptation problem is viewed as recovering the speaker-dependent supervector from limited speech signal observations. A redundant speaker dictionary is constructed by a combination of all the training speaker supervectors and the supervectors derived from the EV method. Given the adaptation data, the best subspace for a particular speaker is constructed in a maximum a posterior manner by selecting a proper set of items from this dictionary. Two algorithms, i.e. matching pursuit and l1 regularized optimization, are adapted to solve this problem. With an efficient redundant basis vector removal mechanism and an iterative updating of the speaker coordinate, the matching pursuit based speaker adaptation method is fast and efficient. The matching pursuit algorithm is greedy and sub-optimal, while direct optimization of the likelihood of the adaptation data with an explicit l1 regularization term can obtain better approximation of the unknown speaker model. The projected gradient optimization algorithm is adopted and a few iterations of the matching pursuit algorithm can provide a good initial value. Experimental results show that matching pursuit algorithm outperforms the conventional testing methods under all testing conditions. Better performance is obtained when direct l1 regularized optimization is applied. Both methods can select a proper mixed set of the eigenvoice and reference speaker supervectors automatically for estimation of the unknown speaker models.

Speaker adaptation in the maximum a posteriori framework based on the probabilistic 2-mode analysis of training models

In this article, we describe a speaker adaptation method based on the probabilistic 2-mode analysis of training models. Probabilistic 2-mode analysis is a probabilistic extension of multilinear analysis. We apply probabilistic 2-mode analysis to speaker adaptation by representing each of the hidden Markov model mean vectors of training speakers as a matrix, and derive the speaker adaptation equation in the maximum a posteriori (MAP) framework. The adaptation equation becomes similar to the speaker adaptation equation using the MAP linear regression adaptation. In the experiments, the adapted models based on probabilistic 2-mode analysis showed performance improvement over the adapted models based on Tucker decomposition, which is a representative multilinear decomposition technique, for small amounts of adaptation data while maintaining good performance for large amounts of adaptation data.