Syu-Siang Wang

Syu-Siang Wang, Chi-Te Wang, Chih-Chung Lai et al.

Goal: Numerous studies had successfully differentiated normal and abnormal voice samples. Nevertheless, further classification had rarely been attempted. This study proposes a novel approach, using continuous Mandarin speech instead of a single vowel, to classify four common voice disorders (i.e. functional dysphonia, neoplasm, phonotrauma, and vocal palsy). Methods: In the proposed framework, acoustic signals are transformed into mel-frequency cepstral coefficients, and a bi-directional long-short term memory network (BiLSTM) is adopted to model the sequential features. The experiments were conducted on a large-scale database, wherein 1,045 continuous speech were collected by the speech clinic of a hospital from 2012 to 2019. Results: Experimental results demonstrated that the proposed framework yields significant accuracy and unweighted average recall improvements of 78.12-89.27% and 50.92-80.68%, respectively, compared with systems that use a single vowel. Conclusions: The results are consistent with other machine learning algorithms, including gated recurrent units, random forest, deep neural networks, and LSTM. The sensitivities for each disorder were also analyzed, and the model capabilities were visualized via principal component analysis. An alternative experiment based on a balanced dataset again confirms the advantages of using continuous speech for learning voice disorders.

Wen-Yuan Ting, Syu-Siang Wang, Hsin-Li Chang et al.

Cycle-consistent generative adversarial networks (CycleGAN) were successfully applied to speech enhancement (SE) tasks with unpaired noisy-clean training data. The CycleGAN SE system adopted two generators and two discriminators trained with losses from noisy-to-clean and clean-to-noisy conversions. CycleGAN showed promising results for numerous SE tasks. Herein, we investigate a potential limitation of the clean-to-noisy conversion part and propose a novel noise-informed training (NIT) approach to improve the performance of the original CycleGAN SE system. The main idea of the NIT approach is to incorporate target domain information for clean-to-noisy conversion to facilitate a better training procedure. The experimental results confirmed that the proposed NIT approach improved the generalization capability of the original CycleGAN SE system with a notable margin.

Yun-Ju Chan, Chiang-Jen Peng, Syu-Siang Wang et al.

Numerous voice conversion (VC) techniques have been proposed for the conversion of voices among different speakers. Although good quality of the converted speech can be observed when VC is applied in a clean environment, the quality degrades drastically when the system is run in noisy conditions. In order to address this issue, we propose a novel speech enhancement (SE)-assisted VC system that utilizes the SE techniques for signal pre-processing, where the VC and SE components are optimized in an joint training strategy with the aim to provide high-quality converted speech signals. We adopt a popular model, StarGAN, as the VC component and thus call the combined system as EStarGAN. We test the proposed EStarGAN system using a Mandarin speech corpus. The experimental results first verified the effectiveness of joint training strategy used in EStarGAN. Moreover, EStarGAN demonstrated performance robustness in various unseen noisy environments. The subjective listening test results further showed that EStarGAN can improve the sound quality of speech signals converted from noise-corrupted source utterances.

Chiang-Jen Peng, Yun-Ju Chan, Cheng Yu et al.

Multi-task learning (MTL) and attention mechanism have been proven to effectively extract robust acoustic features for various speech-related tasks in noisy environments. In this study, we propose an attention-based MTL (ATM) approach that integrates MTL and the attention-weighting mechanism to simultaneously realize a multi-model learning structure that performs speech enhancement (SE) and speaker identification (SI). The proposed ATM system consists of three parts: SE, SI, and attention-Net (AttNet). The SE part is composed of a long-short-term memory (LSTM) model, and a deep neural network (DNN) model is used to develop the SI and AttNet parts. The overall ATM system first extracts the representative features and then enhances the speech signals in LSTM-SE and specifies speaker identity in DNN-SI. The AttNet computes weights based on DNN-SI to prepare better representative features for LSTM-SE. We tested the proposed ATM system on Taiwan Mandarin hearing in noise test sentences. The evaluation results confirmed that the proposed system can effectively enhance speech quality and intelligibility of a given noisy input. Moreover, the accuracy of the SI can also be notably improved by using the proposed ATM system.

Yu-Wen Chen, Kuo-Hsuan Hung, You-Jin Li et al.

This study presents a deep learning-based speech signal-processing mobile application known as CITISEN. The CITISEN provides three functions: speech enhancement (SE), model adaptation (MA), and background noise conversion (BNC), allowing CITISEN to be used as a platform for utilizing and evaluating SE models and flexibly extend the models to address various noise environments and users. For SE, a pretrained SE model downloaded from the cloud server is used to effectively reduce noise components from instant or saved recordings provided by users. For encountering unseen noise or speaker environments, the MA function is applied to promote CITISEN. A few audio samples recording on a noisy environment are uploaded and used to adapt the pretrained SE model on the server. Finally, for BNC, CITISEN first removes the background noises through an SE model and then mixes the processed speech with new background noise. The novel BNC function can evaluate SE performance under specific conditions, cover people's tracks, and provide entertainment. The experimental results confirmed the effectiveness of SE, MA, and BNC functions. Compared with the noisy speech signals, the enhanced speech signals achieved about 6\% and 33\% of improvements, respectively, in terms of short-time objective intelligibility (STOI) and perceptual evaluation of speech quality (PESQ). With MA, the STOI and PESQ could be further improved by approximately 6\% and 11\%, respectively. Finally, the BNC experiment results indicated that the speech signals converted from noisy and silent backgrounds have a close scene identification accuracy and similar embeddings in an acoustic scene classification model. Therefore, the proposed BNC can effectively convert the background noise of a speech signal and be a data augmentation method when clean speech signals are unavailable.

Szu-Wei Fu, Chien-Feng Liao, Tsun-An Hsieh et al.

The Transformer architecture has demonstrated a superior ability compared to recurrent neural networks in many different natural language processing applications. Therefore, our study applies a modified Transformer in a speech enhancement task. Specifically, positional encoding in the Transformer may not be necessary for speech enhancement, and hence, it is replaced by convolutional layers. To further improve the perceptual evaluation of the speech quality (PESQ) scores of enhanced speech, the L_1 pre-trained Transformer is fine-tuned using a MetricGAN framework. The proposed MetricGAN can be treated as a general post-processing module to further boost the objective scores of interest. The experiments were conducted using the data sets provided by the organizer of the Deep Noise Suppression (DNS) challenge. Experimental results demonstrated that the proposed system outperformed the challenge baseline, in both subjective and objective evaluations, with a large margin.

Cheng Yu, Ryandhimas E. Zezario, Syu-Siang Wang et al.

Deep learning-based models have greatly advanced the performance of speech enhancement (SE) systems. However, two problems remain unsolved, which are closely related to model generalizability to noisy conditions: (1) mismatched noisy condition during testing, i.e., the performance is generally sub-optimal when models are tested with unseen noise types that are not involved in the training data; (2) local focus on specific noisy conditions, i.e., models trained using multiple types of noises cannot optimally remove a specific noise type even though the noise type has been involved in the training data. These problems are common in real applications. In this paper, we propose a novel denoising autoencoder with a multi-branched encoder (termed DAEME) model to deal with these two problems. In the DAEME model, two stages are involved: training and testing. In the training stage, we build multiple component models to form a multi-branched encoder based on a decision tree (DSDT). The DSDT is built based on prior knowledge of speech and noisy conditions (the speaker, environment, and signal factors are considered in this paper), where each component of the multi-branched encoder performs a particular mapping from noisy to clean speech along the branch in the DSDT. Finally, a decoder is trained on top of the multi-branched encoder. In the testing stage, noisy speech is first processed by each component model. The multiple outputs from these models are then integrated into the decoder to determine the final enhanced speech. Experimental results show that DAEME is superior to several baseline models in terms of objective evaluation metrics, automatic speech recognition results, and quality in subjective human listening tests.

Yu-Tao Chang, Yuan-Hong Yang, Yu-Huai Peng et al.

With the recent advancements of deep learning technologies, the performance of voice conversion (VC) in terms of quality and similarity has been significantly improved. However, heavy computations are generally required for deep-learning-based VC systems, which can cause notable latency and thus confine their deployments in real-world applications. Therefore, increasing online computation efficiency has become an important task. In this study, we propose a novel mixture-of-experts (MoE) based VC system. The MoE model uses a gating mechanism to specify optimal weights to feature maps to increase VC performance. In addition, assigning sparse constraints on the gating mechanism can accelerate online computation by skipping the convolution process by zeroing out redundant feature maps. Experimental results show that by specifying suitable sparse constraints, we can effectively increase the online computation efficiency with a notable 70% FLOPs (floating-point operations per second) reduction while improving the VC performance in both objective evaluations and human listening tests.

Cheng Yu, Kuo-Hsuan Hung, Syu-Siang Wang et al.

Previous studies have proven that integrating video signals, as a complementary modality, can facilitate improved performance for speech enhancement (SE). However, video clips usually contain large amounts of data and pose a high cost in terms of computational resources and thus may complicate the SE system. As an alternative source, a bone-conducted speech signal has a moderate data size while manifesting speech-phoneme structures, and thus complements its air-conducted counterpart. In this study, we propose a novel multi-modal SE structure in the time domain that leverages bone- and air-conducted signals. In addition, we examine two ensemble-learning-based strategies, early fusion (EF) and late fusion (LF), to integrate the two types of speech signals, and adopt a deep learning-based fully convolutional network to conduct the enhancement. The experiment results on the Mandarin corpus indicate that this newly presented multi-modal (integrating bone- and air-conducted signals) SE structure significantly outperforms the single-source SE counterparts (with a bone- or air-conducted signal only) in various speech evaluation metrics. In addition, the adoption of an LF strategy other than an EF in this novel SE multi-modal structure achieves better results.

Syu-Siang Wang, Yu-You Liang, Jeih-weih Hung et al.

Speech-related applications deliver inferior performance in complex noise environments. Therefore, this study primarily addresses this problem by introducing speech-enhancement (SE) systems based on deep neural networks (DNNs) applied to a distributed microphone architecture, and then investigates the effectiveness of three different DNN-model structures. The first system constructs a DNN model for each microphone to enhance the recorded noisy speech signal, and the second system combines all the noisy recordings into a large feature structure that is then enhanced through a DNN model. As for the third system, a channel-dependent DNN is first used to enhance the corresponding noisy input, and all the channel-wise enhanced outputs are fed into a DNN fusion model to construct a nearly clean signal. All the three DNN SE systems are operated in the acoustic frequency domain of speech signals in a diffuse-noise field environment. Evaluation experiments were conducted on the Taiwan Mandarin Hearing in Noise Test (TMHINT) database, and the results indicate that all the three DNN-based SE systems provide the original noise-corrupted signals with improved speech quality and intelligibility, whereas the third system delivers the highest signal-to-noise ratio (SNR) improvement and optimal speech intelligibility.

Yih-Liang Shen, Chao-Yuan Huang, Syu-Siang Wang et al.

Conventional deep neural network (DNN)-based speech enhancement (SE) approaches aim to minimize the mean square error (MSE) between enhanced speech and clean reference. The MSE-optimized model may not directly improve the performance of an automatic speech recognition (ASR) system. If the target is to minimize the recognition error, the recognition results should be used to design the objective function for optimizing the SE model. However, the structure of an ASR system, which consists of multiple units, such as acoustic and language models, is usually complex and not differentiable. In this study, we proposed to adopt the reinforcement learning algorithm to optimize the SE model based on the recognition results. We evaluated the propsoed SE system on the Mandarin Chinese broadcast news corpus (MATBN). Experimental results demonstrate that the proposed method can effectively improve the ASR results with a notable 12.40% and 19.23% error rate reductions for signal to noise ratio at 0 dB and 5 dB conditions, respectively.

Shih-kuang Lee, Syu-Siang Wang, Yu Tsao et al.

In this paper, we propose a novel speech enhancement (SE) method by exploiting the discrete wavelet transform (DWT). This new method reduces the amount of fast time-varying portion, viz. the DWT-wise detail component, in the spectrogram of speech signals so as to highlight the speech-dominant component and achieves better speech quality. A particularity of this new method is that it is completely unsupervised and requires no prior information about the clean speech and noise in the processed utterance. The presented DWT-based SE method with various scaling factors for the detail part is evaluated with a subset of Aurora-2 database, and the PESQ metric is used to indicate the quality of processed speech signals. The preliminary results show that the processed speech signals reveal a higher PESQ score in comparison with the original counterparts. Furthermore, we show that this method can still enhance the signal by totally discarding the detail part (setting the respective scaling factor to zero), revealing that the spectrogram can be down-sampled and thus compressed without the cost of lowered quality. In addition, we integrate this new method with conventional speech enhancement algorithms, including spectral subtraction, Wiener filtering, and spectral MMSE estimation, and show that the resulting integration behaves better than the respective component method. As a result, this new method is quite effective in improving the speech quality and well additive to the other SE methods.

Wei-Jen Lee, Syu-Siang Wang, Fei Chen et al.

Reverberation, which is generally caused by sound reflections from walls, ceilings, and floors, can result in severe performance degradation of acoustic applications. Due to a complicated combination of attenuation and time-delay effects, the reverberation property is difficult to characterize, and it remains a challenging task to effectively retrieve the anechoic speech signals from reverberation ones. In the present study, we proposed a novel integrated deep and ensemble learning algorithm (IDEA) for speech dereverberation. The IDEA consists of offline and online phases. In the offline phase, we train multiple dereverberation models, each aiming to precisely dereverb speech signals in a particular acoustic environment; then a unified fusion function is estimated that aims to integrate the information of multiple dereverberation models. In the online phase, an input utterance is first processed by each of the dereverberation models. The outputs of all models are integrated accordingly to generate the final anechoic signal. We evaluated the IDEA on designed acoustic environments, including both matched and mismatched conditions of the training and testing data. Experimental results confirm that the proposed IDEA outperforms single deep-neural-network-based dereverberation model with the same model architecture and training data.

Jen-Cheng Hou, Syu-Siang Wang, Ying-Hui Lai et al.

Speech enhancement (SE) aims to reduce noise in speech signals. Most SE techniques focus only on addressing audio information. In this work, inspired by multimodal learning, which utilizes data from different modalities, and the recent success of convolutional neural networks (CNNs) in SE, we propose an audio-visual deep CNNs (AVDCNN) SE model, which incorporates audio and visual streams into a unified network model. We also propose a multi-task learning framework for reconstructing audio and visual signals at the output layer. Precisely speaking, the proposed AVDCNN model is structured as an audio-visual encoder-decoder network, in which audio and visual data are first processed using individual CNNs, and then fused into a joint network to generate enhanced speech (the primary task) and reconstructed images (the secondary task) at the output layer. The model is trained in an end-to-end manner, and parameters are jointly learned through back-propagation. We evaluate enhanced speech using five instrumental criteria. Results show that the AVDCNN model yields a notably superior performance compared with an audio-only CNN-based SE model and two conventional SE approaches, confirming the effectiveness of integrating visual information into the SE process. In addition, the AVDCNN model also outperforms an existing audio-visual SE model, confirming its capability of effectively combining audio and visual information in SE.

Jen-Cheng Hou, Syu-Siang Wang, Ying-Hui Lai et al.

Speech enhancement (SE) aims to reduce noise in speech signals. Most SE techniques focus only on addressing audio information. In this work, inspired by multimodal learning, which utilizes data from different modalities, and the recent success of convolutional neural networks (CNNs) in SE, we propose an audio-visual deep CNNs (AVDCNN) SE model, which incorporates audio and visual streams into a unified network model. We also propose a multi-task learning framework for reconstructing audio and visual signals at the output layer. Precisely speaking, the proposed AVDCNN model is structured as an audio-visual encoder-decoder network, in which audio and visual data are first processed using individual CNNs, and then fused into a joint network to generate enhanced speech (the primary task) and reconstructed images (the secondary task) at the output layer. The model is trained in an end-to-end manner, and parameters are jointly learned through back-propagation. We evaluate enhanced speech using five instrumental criteria. Results show that the AVDCNN model yields a notably superior performance compared with an audio-only CNN-based SE model and two conventional SE approaches, confirming the effectiveness of integrating visual information into the SE process. In addition, the AVDCNN model also outperforms an existing audio-visual SE model, confirming its capability of effectively combining audio and visual information in SE.

Syu-Siang Wang, Alan Chern, Yu Tsao et al.

For most of the state-of-the-art speech enhancement techniques, a spectrogram is usually preferred than the respective time-domain raw data since it reveals more compact presentation together with conspicuous temporal information over a long time span. However, the short-time Fourier transform (STFT) that creates the spectrogram in general distorts the original signal and thereby limits the capability of the associated speech enhancement techniques. In this study, we propose a novel speech enhancement method that adopts the algorithms of discrete wavelet packet transform (DWPT) and nonnegative matrix factorization (NMF) in order to conquer the aforementioned limitation. In brief, the DWPT is first applied to split a time-domain speech signal into a series of subband signals without introducing any distortion. Then we exploit NMF to highlight the speech component for each subband. Finally, the enhanced subband signals are joined together via the inverse DWPT to reconstruct a noise-reduced signal in time domain. We evaluate the proposed DWPT-NMF based speech enhancement method on the MHINT task. Experimental results show that this new method behaves very well in prompting speech quality and intelligibility and it outperforms the convnenitional STFT-NMF based method.