arXiv subjects

Junichi Yamagishi

Publications and source records attributed to Junichi Yamagishi.

At least 181 records · Page 10Linked to original sources

arXiv2019

Transferring neural speech waveform synthesizers to musical instrument sounds generation

Recent neural waveform synthesizers such as WaveNet, WaveGlow, and the neural-source-filter (NSF) model have shown good performance in speech synthesis despite their different methods of waveform generation. The similarity between speech and music audio synthesis techniques suggests interesting avenues to explore in terms of the best way to apply speech synthesizers in the music domain. This work compares three neural synthesizers used for musical instrument sounds generation under three scenarios: training from scratch on music data, zero-shot learning from the speech domain, and fine-tuning-based adaptation from the speech to the music domain. The results of a large-scale perceptual test demonstrated that the performance of three synthesizers improved when they were pre-trained on speech data and fine-tuned on music data, which indicates the usefulness of knowledge from speech data for music audio generation. Among the synthesizers, WaveGlow showed the best potential in zero-shot learning while NSF performed best in the other scenarios and could generate samples that were perceptually close to natural audio.

eess.AS↗

arXiv2019

Neural source-filter waveform models for statistical parametric speech synthesis

Neural waveform models such as WaveNet have demonstrated better performance than conventional vocoders for statistical parametric speech synthesis. As an autoregressive (AR) model, WaveNet is limited by a slow sequential waveform generation process. Some new models that use the inverse-autoregressive flow (IAF) can generate a whole waveform in a one-shot manner. However, these IAF-based models require sequential transformation during training, which severely slows down the training speed. Other models such as Parallel WaveNet and ClariNet bring together the benefits of AR and IAF-based models and train an IAF model by transferring the knowledge from a pre-trained AR teacher to an IAF student without any sequential transformation. However, both models require additional training criteria, and their implementation is prohibitively complicated. We propose a framework for neural source-filter (NSF) waveform modeling without AR nor IAF-based approaches. This framework requires only three components for waveform generation: a source module that generates a sine-based signal as excitation, a non-AR dilated-convolution-based filter module that transforms the excitation into a waveform, and a conditional module that pre-processes the acoustic features for the source and filer modules. This framework minimizes spectral-amplitude distances for model training, which can be efficiently implemented by using short-time Fourier transform routines. Under this framework, we designed three NSF models and compared them with WaveNet. It was demonstrated that the NSF models generated waveforms at least 100 times faster than WaveNet, and the quality of the synthetic speech from the best NSF model was better than or equally good as that from WaveNet.

eess.AS↗

arXiv2019

A Method for Identifying Origin of Digital Images Using a Convolution Neural Network

The rapid development of deep learning techniques has created new challenges in identifying the origin of digital images because generative adversarial networks and variational autoencoders can create plausible digital images whose contents are not present in natural scenes. In this paper, we consider the origin that can be broken down into three categories: natural photographic image (NPI), computer generated graphic (CGG), and deep network generated image (DGI). A method is presented for effectively identifying the origin of digital images that is based on a convolutional neural network (CNN) and uses a local-to-global framework to reduce training complexity. By feeding labeled data, the CNN is trained to predict the origin of local patches cropped from an image. The origin of the full-size image is then determined by majority voting. Unlike previous forensic methods, the CNN takes the raw pixels as input without the aid of "residual map". Experimental results revealed that not only the high-frequency components but also the middle-frequency ones contribute to origin identification. The proposed method achieved up to 95.21% identification accuracy and behaved robustly against several common post-processing operations including JPEG compression, scaling, geometric transformation, and contrast stretching. The quantitative results demonstrate that the proposed method is more effective than handcrafted feature-based methods.

cs.CV↗

arXiv2019

Use of a Capsule Network to Detect Fake Images and Videos

The revolution in computer hardware, especially in graphics processing units and tensor processing units, has enabled significant advances in computer graphics and artificial intelligence algorithms. In addition to their many beneficial applications in daily life and business, computer-generated/manipulated images and videos can be used for malicious purposes that violate security systems, privacy, and social trust. The deepfake phenomenon and its variations enable a normal user to use his or her personal computer to easily create fake videos of anybody from a short real online video. Several countermeasures have been introduced to deal with attacks using such videos. However, most of them are targeted at certain domains and are ineffective when applied to other domains or new attacks. In this paper, we introduce a capsule network that can detect various kinds of attacks, from presentation attacks using printed images and replayed videos to attacks using fake videos created using deep learning. It uses many fewer parameters than traditional convolutional neural networks with similar performance. Moreover, we explain, for the first time ever in the literature, the theory behind the application of capsule networks to the forensics problem through detailed analysis and visualization.

cs.CV↗

arXiv2019

Effect of choice of probability distribution, randomness, and search methods for alignment modeling in sequence-to-sequence text-to-speech synthesis using hard alignment

Sequence-to-sequence text-to-speech (TTS) is dominated by soft-attention-based methods. Recently, hard-attention-based methods have been proposed to prevent fatal alignment errors, but their sampling method of discrete alignment is poorly investigated. This research investigates various combinations of sampling methods and probability distributions for alignment transition modeling in a hard-alignment-based sequence-to-sequence TTS method called SSNT-TTS. We clarify the common sampling methods of discrete variables including greedy search, beam search, and random sampling from a Bernoulli distribution in a more general way. Furthermore, we introduce the binary Concrete distribution to model discrete variables more properly. The results of a listening test shows that deterministic search is more preferable than stochastic search, and the binary Concrete distribution is robust with stochastic search for natural alignment transition.

eess.AS↗

arXiv2019

A Unified Speaker Adaptation Method for Speech Synthesis using Transcribed and Untranscribed Speech with Backpropagation

By representing speaker characteristic as a single fixed-length vector extracted solely from speech, we can train a neural multi-speaker speech synthesis model by conditioning the model on those vectors. This model can also be adapted to unseen speakers regardless of whether the transcript of adaptation data is available or not. However, this setup restricts the speaker component to just a single bias vector, which in turn limits the performance of adaptation process. In this study, we propose a novel speech synthesis model, which can be adapted to unseen speakers by fine-tuning part of or all of the network using either transcribed or untranscribed speech. Our methodology essentially consists of two steps: first, we split the conventional acoustic model into a speaker-independent (SI) linguistic encoder and a speaker-adaptive (SA) acoustic decoder; second, we train an auxiliary acoustic encoder that can be used as a substitute for the linguistic encoder whenever linguistic features are unobtainable. The results of objective and subjective evaluations show that adaptation using either transcribed or untranscribed speech with our methodology achieved a reasonable level of performance with an extremely limited amount of data and greatly improved performance with more data. Surprisingly, adaptation with untranscribed speech surpassed the transcribed counterpart in the subjective test, which reveals the limitations of the conventional acoustic model and hints at potential directions for improvements.

eess.AS↗

arXiv2019

Bootstrapping non-parallel voice conversion from speaker-adaptive text-to-speech

Voice conversion (VC) and text-to-speech (TTS) are two tasks that share a similar objective, generating speech with a target voice. However, they are usually developed independently under vastly different frameworks. In this paper, we propose a methodology to bootstrap a VC system from a pretrained speaker-adaptive TTS model and unify the techniques as well as the interpretations of these two tasks. Moreover by offloading the heavy data demand to the training stage of the TTS model, our VC system can be built using a small amount of target speaker speech data. It also opens up the possibility of using speech in a foreign unseen language to build the system. Our subjective evaluations show that the proposed framework is able to not only achieve competitive performance in the standard intra-language scenario but also adapt and convert using speech utterances in an unseen language.

eess.AS↗

arXiv2019

Initial investigation of an encoder-decoder end-to-end TTS framework using marginalization of monotonic hard latent alignments

End-to-end text-to-speech (TTS) synthesis is a method that directly converts input text to output acoustic features using a single network. A recent advance of end-to-end TTS is due to a key technique called attention mechanisms, and all successful methods proposed so far have been based on soft attention mechanisms. However, although network structures are becoming increasingly complex, end-to-end TTS systems with soft attention mechanisms may still fail to learn and to predict accurate alignment between the input and output. This may be because the soft attention mechanisms are too flexible. Therefore, we propose an approach that has more explicit but natural constraints suitable for speech signals to make alignment learning and prediction of end-to-end TTS systems more robust. The proposed system, with the constrained alignment scheme borrowed from segment-to-segment neural transduction (SSNT), directly calculates the joint probability of acoustic features and alignment given an input text. The alignment is designed to be hard and monotonically increase by considering the speech nature, and it is treated as a latent variable and marginalized during training. During prediction, both the alignment and acoustic features can be generated from the probabilistic distributions. The advantages of our approach are that we can simplify many modules for the soft attention and that we can train the end-to-end TTS model using a single likelihood function. As far as we know, our approach is the first end-to-end TTS without a soft attention mechanism.

eess.AS↗

arXiv2019

Neural Harmonic-plus-Noise Waveform Model with Trainable Maximum Voice Frequency for Text-to-Speech Synthesis

Neural source-filter (NSF) models are deep neural networks that produce waveforms given input acoustic features. They use dilated-convolution-based neural filter modules to filter sine-based excitation for waveform generation, which is different from WaveNet and flow-based models. One of the NSF models, called harmonic-plus-noise NSF (h-NSF) model, uses separate pairs of source and neural filters to generate harmonic and noise waveform components. It is close to WaveNet in terms of speech quality while being superior in generation speed. The h-NSF model can be improved even further. While h-NSF merges the harmonic and noise components using pre-defined digital low- and high-pass filters, it is well known that the maximum voice frequency (MVF) that separates the periodic and aperiodic spectral bands are time-variant. Therefore, we propose a new h-NSF model with time-variant and trainable MVF. We parameterize the digital low- and high-pass filters as windowed-sinc filters and predict their cut-off frequency (i.e., MVF) from the input acoustic features. Our experiments demonstrated that the new model can predict a good trajectory of the MVF and produce high-quality speech for a text-to-speech synthesis system.

eess.AS↗

arXiv2019

GELP: GAN-Excited Linear Prediction for Speech Synthesis from Mel-spectrogram

Recent advances in neural network -based text-to-speech have reached human level naturalness in synthetic speech. The present sequence-to-sequence models can directly map text to mel-spectrogram acoustic features, which are convenient for modeling, but present additional challenges for vocoding (i.e., waveform generation from the acoustic features). High-quality synthesis can be achieved with neural vocoders, such as WaveNet, but such autoregressive models suffer from slow sequential inference. Meanwhile, their existing parallel inference counterparts are difficult to train and require increasingly large model sizes. In this paper, we propose an alternative training strategy for a parallel neural vocoder utilizing generative adversarial networks, and integrate a linear predictive synthesis filter into the model. Results show that the proposed model achieves significant improvement in inference speed, while outperforming a WaveNet in copy-synthesis quality.

eess.AS↗

arXiv2019

Multi-task Learning For Detecting and Segmenting Manipulated Facial Images and Videos

Detecting manipulated images and videos is an important topic in digital media forensics. Most detection methods use binary classification to determine the probability of a query being manipulated. Another important topic is locating manipulated regions (i.e., performing segmentation), which are mostly created by three commonly used attacks: removal, copy-move, and splicing. We have designed a convolutional neural network that uses the multi-task learning approach to simultaneously detect manipulated images and videos and locate the manipulated regions for each query. Information gained by performing one task is shared with the other task and thereby enhance the performance of both tasks. A semi-supervised learning approach is used to improve the network's generability. The network includes an encoder and a Y-shaped decoder. Activation of the encoded features is used for the binary classification. The output of one branch of the decoder is used for segmenting the manipulated regions while that of the other branch is used for reconstructing the input, which helps improve overall performance. Experiments using the FaceForensics and FaceForensics++ databases demonstrated the network's effectiveness against facial reenactment attacks and face swapping attacks as well as its ability to deal with the mismatch condition for previously seen attacks. Moreover, fine-tuning using just a small amount of data enables the network to deal with unseen attacks.

cs.CV↗

arXiv2019

Speaker Anonymization Using X-vector and Neural Waveform Models

The social media revolution has produced a plethora of web services to which users can easily upload and share multimedia documents. Despite the popularity and convenience of such services, the sharing of such inherently personal data, including speech data, raises obvious security and privacy concerns. In particular, a user's speech data may be acquired and used with speech synthesis systems to produce high-quality speech utterances which reflect the same user's speaker identity. These utterances may then be used to attack speaker verification systems. One solution to mitigate these concerns involves the concealing of speaker identities before the sharing of speech data. For this purpose, we present a new approach to speaker anonymization. The idea is to extract linguistic and speaker identity features from an utterance and then to use these with neural acoustic and waveform models to synthesize anonymized speech. The original speaker identity, in the form of timbre, is suppressed and replaced with that of an anonymous pseudo identity. The approach exploits state-of-the-art x-vector speaker representations. These are used to derive anonymized pseudo speaker identities through the combination of multiple, random speaker x-vectors. Experimental results show that the proposed approach is effective in concealing speaker identities. It increases the equal error rate of a speaker verification system while maintaining high quality, anonymized speech.

eess.AS↗

arXiv2019

Neural source-filter-based waveform model for statistical parametric speech synthesis

Neural waveform models such as the WaveNet are used in many recent text-to-speech systems, but the original WaveNet is quite slow in waveform generation because of its autoregressive (AR) structure. Although faster non-AR models were recently reported, they may be prohibitively complicated due to the use of a distilling training method and the blend of other disparate training criteria. This study proposes a non-AR neural source-filter waveform model that can be directly trained using spectrum-based training criteria and the stochastic gradient descent method. Given the input acoustic features, the proposed model first uses a source module to generate a sine-based excitation signal and then uses a filter module to transform the excitation signal into the output speech waveform. Our experiments demonstrated that the proposed model generated waveforms at least 100 times faster than the AR WaveNet and the quality of its synthetic speech is close to that of speech generated by the AR WaveNet. Ablation test results showed that both the sine-wave excitation signal and the spectrum-based training criteria were essential to the performance of the proposed model.

eess.AS↗

arXiv2019

ASVspoof 2019: Future Horizons in Spoofed and Fake Audio Detection

ASVspoof, now in its third edition, is a series of community-led challenges which promote the development of countermeasures to protect automatic speaker verification (ASV) from the threat of spoofing. Advances in the 2019 edition include: (i) a consideration of both logical access (LA) and physical access (PA) scenarios and the three major forms of spoofing attack, namely synthetic, converted and replayed speech; (ii) spoofing attacks generated with state-of-the-art neural acoustic and waveform models; (iii) an improved, controlled simulation of replay attacks; (iv) use of the tandem detection cost function (t-DCF) that reflects the impact of both spoofing and countermeasures upon ASV reliability. Even if ASV remains the core focus, in retaining the equal error rate (EER) as a secondary metric, ASYspoof also embraces the growing importance of fake audio detection. ASVspoof 2019 attracted the participation of 63 research teams, with more than half of these reporting systems that improve upon the performance of two baseline spoofing countermeasures. This paper describes the 2019 database, protocols and challenge results. It also outlines major findings which demonstrate the real progress made in protecting against the threat of spoofing and fake audio.

eess.AS↗

arXiv2019

t-DCF: a Detection Cost Function for the Tandem Assessment of Spoofing Countermeasures and Automatic Speaker Verification

The ASVspoof challenge series was born to spearhead research in anti-spoofing for automatic speaker verification (ASV). The two challenge editions in 2015 and 2017 involved the assessment of spoofing countermeasures (CMs) in isolation from ASV using an equal error rate (EER) metric. While a strategic approach to assessment at the time, it has certain shortcomings. First, the CM EER is not necessarily a reliable predictor of performance when ASV and CMs are combined. Second, the EER operating point is ill-suited to user authentication applications, e.g. telephone banking, characterised by a high target user prior but a low spoofing attack prior. We aim to migrate from CM- to ASV-centric assessment with the aid of a new tandem detection cost function (t-DCF) metric. It extends the conventional DCF used in ASV research to scenarios involving spoofing attacks. The t-DCF metric has 6 parameters: (i) false alarm and miss costs for both systems, and (ii) prior probabilities of target and spoof trials (with an implied third, nontarget prior). The study is intended to serve as a self-contained, tutorial-like presentation. We analyse with the t-DCF a selection of top-performing CM submissions to the 2015 and 2017 editions of ASVspoof, with a focus on the spoofing attack prior. Whereas there is little to choose between countermeasure systems for lower priors, system rankings derived with the EER and t-DCF show differences for higher priors. We observe some ranking changes. Findings support the adoption of the DCF-based metric into the roadmap for future ASVspoof challenges, and possibly for other biometric anti-spoofing evaluations.

eess.AS↗

arXiv2019

Does the Lombard Effect Improve Emotional Communication in Noise? - Analysis of Emotional Speech Acted in Noise -

Speakers usually adjust their way of talking in noisy environments involuntarily for effective communication. This adaptation is known as the Lombard effect. Although speech accompanying the Lombard effect can improve the intelligibility of a speaker's voice, the changes in acoustic features (e.g. fundamental frequency, speech intensity, and spectral tilt) caused by the Lombard effect may also affect the listener's judgment of emotional content. To the best of our knowledge, there is no published study on the influence of the Lombard effect in emotional speech. Therefore, we recorded parallel emotional speech waveforms uttered by 12 speakers under both quiet and noisy conditions in a professional recording studio in order to explore how the Lombard effect interacts with emotional speech. By analyzing confusion matrices and acoustic features, we aim to answer the following questions: 1) Can speakers express their emotions correctly even under adverse conditions? 2) Can listeners recognize the emotion contained in speech signals even under noise? 3) How does emotional speech uttered in noise differ from emotional speech uttered in quiet conditions in terms of acoustic characteristic?

eess.AS↗

arXiv2019

Joint training framework for text-to-speech and voice conversion using multi-source Tacotron and WaveNet

We investigated the training of a shared model for both text-to-speech (TTS) and voice conversion (VC) tasks. We propose using an extended model architecture of Tacotron, that is a multi-source sequence-to-sequence model with a dual attention mechanism as the shared model for both the TTS and VC tasks. This model can accomplish these two different tasks respectively according to the type of input. An end-to-end speech synthesis task is conducted when the model is given text as the input while a sequence-to-sequence voice conversion task is conducted when it is given the speech of a source speaker as the input. Waveform signals are generated by using WaveNet, which is conditioned by using a predicted mel-spectrogram. We propose jointly training a shared model as a decoder for a target speaker that supports multiple sources. Listening experiments show that our proposed multi-source encoder-decoder model can efficiently achieve both the TTS and VC tasks.

eess.AS↗

arXiv2019

Training a Neural Speech Waveform Model using Spectral Losses of Short-Time Fourier Transform and Continuous Wavelet Transform

Recently, we proposed short-time Fourier transform (STFT)-based loss functions for training a neural speech waveform model. In this paper, we generalize the above framework and propose a training scheme for such models based on spectral amplitude and phase losses obtained by either STFT or continuous wavelet transform (CWT), or both of them. Since CWT is capable of having time and frequency resolutions different from those of STFT and is cable of considering those closer to human auditory scales, the proposed loss functions could provide complementary information on speech signals. Experimental results showed that it is possible to train a high-quality model by using the proposed CWT spectral loss and is as good as one using STFT-based loss.

eess.AS↗