Jump to Content
Heiga Zen

Heiga Zen

Heiga Zen received his AE from Suzuka National College of Technology, Suzuka, Japan, in 1999, and PhD from the Nagoya Institute of Technology, Nagoya, Japan, in 2006. He was an Intern/Co-Op researcher at the IBM T.J. Watson Research Center, Yorktown Heights, NY (2004--2005), and a Research Engineer at Toshiba Research Europe Ltd. Cambridge Research Laboratory, Cambridge, UK (2008--2011). At Google, he was with the Speech team from July 2011 to July 2018, then joined the Brain team from August 2018. From June 2023, he is a Principal Scientist at Google DeepMind, Japan. His research interests include speech technology and machine learning. He was one of the original authors and the first maintainer of the HMM-based speech synthesis system (HTS).
Authored Publications
Google Publications
Other Publications
Sort By
  • Title
  • Title, desc
  • Year
  • Year, desc
    Preview abstract Speech restoration (SR) is a task of converting degraded speech signals into high-quality ones. In this study, we propose a robust SR model called Miipher, and apply Miipher to a new SR application: increasing the amount of high-quality training data for speech generation by converting speech samples collected from the web to studio-quality. To make our SR model robust against various degradation, we use (i) a speech representation extracted from w2v-BERT for the input feature, and (ii) linguistic features extracted from transcripts and PnG-BERT for conditioning features. Experiments show that the proposed model (i) is robust against various audio degradation, (ii) can restore samples in the LJspeech dataset and improves the quality of text-to-speech (TTS) outputs without changing the model and hyper-parameters, and (iii) enable us to train a high-quality TTS model from restored speech samples collected from the web. View details
    Preview abstract This paper proposes Virtuoso, a massive multilingual speech–text joint learning framework for text-to-speech synthesis (TTS) models. Existing multilingual TTS typically supports tens of languages, which are a small fraction of thousands of languages in the world. One difficulty to scale multilingual TTS to hundreds of languages is collecting high-quality speech–text paired data in low-resource languages. This study extends Maestro, which is a speech–text semi-supervised joint pretraining framework for automatic speech recognition (ASR), to speech generation tasks. To train a TTS model from various types of speech and text data, different training schemes are designed to handle supervised (paired TTS and ASR data) and unsupervised (untranscribed speech and unspoken text) datasets. Experimental evaluation shows that 1) multilingual TTS models trained on Virtuoso can achieve significantly better naturalness and intelligibility than baseline TTS models in seen languages, and 2) these models can synthesize reasonably good speech for unseen languages where no paired TTS data is available. View details
    Preview abstract This paper introduces a new speech dataset called ``LibriTTS-R'' designed for text-to-speech (TTS) use. It is derived by applying speech restoration to the LibriTTS corpus, which consists of 585 hours of speech data at 24 kHz sampling rate from 2,456 speakers and the corresponding texts. The constituent samples of LibriTTS-R are identical to those of LibriTTS, with only the sound quality improved. Experimental results show that the LibriTTS-R ground-truth samples showed significantly improved sound quality compared to those in LibriTTS. In addition, neural end-to-end TTS trained with LibriTTS-R achieved speech naturalness on par with that of the ground-truth samples. The corpus is freely available for download from [URL-HERE] View details
    Twenty-Five Years of Evolution in Speech and Language Processing
    Michael Picheny
    Dilek Hakkani-Tur
    IEEE Signal Processing Magazine, vol. 40 (2023), pp. 27-39
    Preview
    Preview abstract This paper explores the research question of whether training neural language models using a small subset of representative data selected from a large training dataset can achieve the same level of performance obtained using all the original training data. We explore the likelihood-based scoring for the purpose of obtaining representative subsets, which we call RepSet. Our experiments confirm that the representative subset obtained by a likelihood difference-based score can achieve the 90% performance level even when the dataset is reduced to about 1,000th of the original data. We also show that the performance of the random selection method deteriorates significantly when the amount of data is reduced. View details
    Preview abstract Transfer tasks in text-to-speech (TTS) synthesis — where one or more aspects of the speech of one set of speakers is transferred to another set of speakers that do not feature these aspects originally — remains a challenging task. One of the challenges is that models that have high-quality transfer capabilities can have issues in stability, making them impractical for user-facing critical tasks. This paper demonstrates that transfer can be obtained by training an robust TTS system on data generated by a less robust TTS system designed for a high-quality transfer task; In particular, a CHiVE-BERT monolingual TTS system is trained on the output of a Tacotron model designed for accent transfer. While some quality loss is inevitable with this approach, experimental results show that the models trained on synthetic data this way can produce high quality audio displaying accent transfer, while preserving speaker characteristics such as speaking style. View details
    Preview abstract We present Maestro, a self-supervised training method to unify representations learnt from speech and text modalities. Self-supervised learning from speech signals aims to learn the latent structure inherent in the signal, while self-supervised learning from text attempts to capture lexical information. Learning aligned representations from unpaired speech and text sequences is a challenging task. Previous work either implicitly enforced the representations learnt from these two modalities to be aligned in the latent space through multi- tasking and parameter sharing or explicitly through conversion of modalities via speech synthesis. While the former suffers from interference between the two modalities, the latter introduces additional complexity. In this paper, we propose Maestro, a novel algorithm to learn unified representations from both these modalities simultaneously that can transfer to diverse downstream tasks such as Automated Speech Recognition (ASR) and Speech Translation (ST). Maestro learns unified representations through sequence alignment, duration predic- tion and matching embeddings in the learned space through an aligned masked-language model loss. We establish a new state-of-the-art (SOTA) on VoxPopuli multilingual ASR with a 8% relative reduction in Word Error Rate (WER), multi- domain SpeechStew ASR (3.7% relative) and 21 languages to English multilingual ST on CoVoST 2 with an improvement of 2.8 BLEU averaged over 21 languages. View details
    Preview abstract Neural vocoder using denoising diffusion probabilistic model (DDPM) has been improved by adaptation of the diffusion noise distribution to given acoustic features. In this study, we propose SpecGrad that adapts the diffusion noise so that its time-varying spectral envelope becomes close to the conditioning log-mel spectrogram. This adaptation by time-varying filtering improves the sound quality especially in the high-frequency bands. It is processed in the time-frequency domain to keep the computational cost almost the same as the conventional DDPM-based neural vocoders. Experimental results showed that SpecGrad generates higher-fidelity speech waveform than conventional DDPM-based neural vocoders in both analysis-synthesis and speech enhancement scenarios. Audio demos are available at [wavegrad.github.io/specgrad/]. View details
    Preview abstract We introduce CVSS, a massively multilingual-to-English speech-to-speech translation (S2ST) corpus, covering sentence-level parallel S2ST pairs from 21 languages into English. CVSS is derived from the Common Voice speech corpus and the CoVoST 2 speech-to-text translation (ST) corpus, by synthesizing the translation text from CoVoST 2 into speech using state-of-the-art TTS systems. Two versions of translation speeches are provided: 1) CVSS-C: All the translation speeches are in a single high-quality canonical voice; 2) CVSS-T: The translation speeches are in voices transferred from the corresponding source speeches. In addition, CVSS provides normalized translation text which matches the pronunciation in the translation speech. On each version of CVSS, we built baseline multilingual direct S2ST models and cascade S2ST models, verifying the effectiveness of the corpus. To build strong cascade S2ST baselines, we trained an ST model on CoVoST 2, which outperforms the previous state-of-the-art trained on the corpus without extra data by 5.8 BLEU. Nevertheless, the performance of the direct S2ST models approaches the strong cascade baselines when trained from scratch, and with only 0.1 or 0.7 BLEU difference on ASR transcribed translation when initialized from matching ST models. View details
    Preview abstract This paper explores the research question of whether training neural language models using a small subset of representative data selected from a large training dataset can achieve the same level of performance that obtained using all the original training data. In our experiments, we confirm that the representative subset obtained by the likelihood-difference-based method can maintain the same performance level even when the dataset is reduced to about 10th or 100th of the original data. We also show that the performance of the random selection method deteriorates significantly when the amount of data is reduced. View details
    Preview abstract Denoising diffusion probabilistic models (DDPMs) and generative adversarial networks (GANs) are popular generative models for neural vocoders. The DDPMs and GANs can be characterized by the iterative denoising framework and adversarial training, respectively. This study proposes a fast and high-quality neural vocoder called WaveFit, which integrates the essence of GANs into a DDPM-like iterative framework based on fixed-point iteration. WaveFit iteratively denoises an input signal, and trains a deep neural network (DNN) for minimizing an adversarial loss calculated from intermediate outputs at all iterations. Subjective (side-by-side) listening tests showed no statistically significant differences in naturalness between human natural speech and those synthesized by WaveFit with five iterations. Furthermore, the inference speed of WaveFit was more than 240 times faster than WaveRNN. Audio demos are available at google.github.io/df-conformer/wavefit/. View details
    深層学習によるテキスト音声合成の飛躍的発展
    電子情報通信学会誌, vol. 105-5 (2022), pp. 413-417
    Preview abstract テキスト音声合成では、音声波形を自動で切り貼りして所望するテキストに対応する音声を合成する、波形接続型音声合成が主流であった。一方、条件付き生成モデルを用いてテキストと音声の関係を学習し、これより任意のテキストより音声を合成する生成モデル型音声合成は、声色を少量の音声で変換できる等の利点があるが、合成音の自然性に課題があった。過去約10年間に深層学習が生成モデル型に導入され、性能が飛躍的に向上した結果、高い自然性を保ちつつ柔軟に話者性や韻律を制御できるようになった。本稿では、深層生成モデルの導入がテキスト音声合成に与えた影響について考察する。 View details
    Preview abstract We present a state-of-the-art non-autoregressive Text-To-Speech model. The model called Parallel Tacotron 2 learns to synthesize speech with good quality without supervised duration signals and other assumptions about the token-to-frame mapping. Specifically, we introduce a novel learned attention mechanism and an iterative reconstruction loss based on Soft Dynamic Time Warping. We show that this new unsupervised model outperforms the baselines in naturalness in several diverse multi speaker evaluations. Further, we show that the explicit duration model that the model has learned can be used to control the synthesized speech. View details
    Preview abstract This paper introduces WaveGrad, a conditional model for waveform generation which estimates gradients of the data density. The model is built on prior work on score matching and diffusion probabilistic models. It starts from a Gaussian white noise signal and iteratively refines the signal via a gradient-based sampler conditioned on the mel-spectrogram. WaveGrad offers a natural way to trade inference speed for sample quality by adjusting the number of refinement steps, and bridges the gap between non-autoregressive and autoregressive models in terms of audio quality. We find that it can generate high fidelity audio samples using as few as six iterations. Experiments reveal WaveGrad to generate high fidelity audio, outperforming adversarial non-autoregressive baselines and matching a strong likelihood-based autoregressive baseline using fewer sequential operations. Audio samples are available at https://wavegrad.github.io/ View details
    Preview abstract Although neural end-to-end text-to-speech models can synthesizehighly natural speech, there is still a room for improvements in itsefficiency during inference. This paper proposes a non-autoregressiveneural text-to-speech model augmented with a variational autoencoder-based residual encoder. This model, calledParallel Tacotron, is highlyparallelizable during both training and inference, allowing efficientsynthesis on modern parallel hardware. The use of the variationalautoencoder helps to relax the one-to-many mapping nature of thetext-to-speech problem. To further improve the naturalness, weintroduce an iterative spectrogram loss, which is inspired by iterativerefinement, and lightweight convolution, which can efficiently capturelocal contexts. Experimental results show that Parallel Tacotronmatches a strong autoregressive baseline in subjective naturalnesswith significantly decreased inference time. View details
    Preview abstract This paper introduces WaveGrad 2, an end-to-end non-autoregressive generative model for text-to-speech synthesis trained to estimate the gradients of the data density. Unlike recent TTS systems which are a cascade of separately learned models, during training the proposed model requires only text or phoneme sequence, learns all parameters end-to-end without intermediate features, and can generate natural speech audio with great varieties. This is achieved by the score matching objective, which optimizes the network to model the score function of the real data distribution. Output waveforms are generated using an iterative refinement process beginning from a random noise sample. Like our prior work, WaveGrad 2 offers a natural way to trade inference speed for sample quality by adjusting the number of refinement steps. Experiments reveal that the model can generate high fidelity audio, closing the gap between end-to-end and contemporary systems, approaching the performance of a state-of-the-art neural TTS system. We further carry out various ablations to study the impact of different model configurations. View details
    Preview abstract This paper introduces a new encoder model for neural TTS. The proposed model, called PnG BERT, is augmented from the original BERT model, but taking both phoneme and grapheme representation of a text, as well as the word-level alignment between them, as its input. It can be pre-trained on a large text corpus in a self-supervised manner then fine-tuned in a TTS task. The experimental results suggest that PnG BERT can significantly further improve the performance of a state-of-the-art neural TTS model, by producing more appropriate prosody and more accurate pronunciation. Subjective side-by-side preference evaluation showed that raters had no statistically significant preference between the synthesized speech and the ground truth recordings from professional speakers. View details
    Preview abstract Semi- and self-supervised training techniques have the potential to improve performance of speech recognition systems without additional transcribed speech data. In this work, we demonstrate the efficacy of two approaches to semi-supervision for automated speech recognition. The two approaches lever-age vast amounts of available unspoken text and untranscribed audio. First, we present factorized multilingual speech synthesis to improve data augmentation on unspoken text. Next, we present an online implementation of Noisy Student Training to incorporate untranscribed audio. We propose a modified Sequential MixMatch algorithm with iterative learning to learn from untranscribed speech. We demonstrate the compatibility of these techniques yielding a relative reduction of word error rate of up to 14.4% on the voice search task. View details
    Preview abstract We propose a hierarchical, fine-grained and interpretable latent model for prosody based on the Tacotron~2. This model achieves multi-resolution modeling by conditioning finer level prosody representations on coarser level ones. In addition, the hierarchical conditioning is also imposed across all latent dimensions using a conditional VAE structure which exploits an auto-regressive structure. Reconstruction performance is evaluated with the $F_0$ frame error (FFE) and the mel-cepstral distortion (MCD) which illustrates the new structure does not degrade the model. Interpretations of prosody attributes are provided together with the comparison between word-level and phone-level prosody representations. Moreover, both qualitative and quantitative evaluations are used to demonstrate the improvement in the disentanglement of the latent dimensions. View details
    Preview abstract This paper presents Non-Attentive Tacotron based on the Tacotron 2 text-to-speech model, where the attention mechanism is replaced with an explicit duration predictor. This improves robustness significantly as measured by unaligned duration ratio and word deletion rate, two new metrics introduced in this paper for large-scale robustness evaluation using a pre-trained speech recognition model. With the use of Gaussian upsampling, Non-Attentive Tacotron achieves a 5-scale mean opinion score in naturalness of 4.41, slightly outperforming Tacotron 2. The duration predictor enables both utterance-wide and per-phoneme control of duration at inference time. If accurate target duration are scarce or unavailable, it is still possible to train a duration predictor in a semi-supervised or unsupervised manner, with results almost as good as supervised training. View details
    Preview abstract Recently proposed approaches for fine-grained prosody control of end-to-end text-to-speech samples enable precise control of the prosody of synthesized speech. Such models incorporate a fine-grained variational autoencoder (VAE) structure into a sequence-to-sequence model, extracting latent prosody features for each input token (e.g.\ phonemes). Generating samples using the standard VAE prior, an independent Gaussian at each time step, results in very unnatural and discontinuous speech, with dramatic variation between phonemes. In this paper we propose a sequential prior in the discrete latent space which can be used to generate more natural samples. This is accomplished by discretizing the latent prosody features using vector quantization, and training an autoregressive (AR) prior model over the result. The AR prior is learned separately from the training of the posterior. We evaluate the approach using subjective listening tests, objective metrics of automatic speech recognition (ASR) performance, as well as measurements of prosody attributes including volume, pitch, and phoneme duration. Compared to the fine-grained VAE baseline, the proposed model achieves equally good copy synthesis reconstruction performance, but significantly improves naturalness in sample generation. The diversity of the prosody in random samples better matches that of the real speech. Furthermore, initial experiments demonstrate that samples generated from the quantized latent sapce can be used as an effective data augmentation strategy to improve ASR performance. View details
    Preview abstract Speech synthesis has advanced to the point of being close to indistinguishable from human speech. However, efforts to train speech recognition systems on synthesized utterances have not been able to show that synthesized data can be effectively used to augment or replace human speech. In this work, we demonstrate that promoting consistent predictions in response to real and synthesized speech enables significantly improved speech recognition performance. We also find that training on 460 hours of LibriSpeech augmented with 500 hours of transcripts (without audio) performance is within 0.2\% WER of a system trained on 960 hours of transcribed audio. This suggests that with this approach, when there is sufficient text available, reliance on transcribed audio can be cut nearly in half. View details
    Hierarchical Generative Modeling for Controllable Speech Synthesis
    Wei-Ning Hsu
    Yu Zhang
    Yuxuan Wang
    Ye Jia
    Jonathan Shen
    Patrick Nguyen
    Ruoming Pang
    International Conference on Learning Representations (2019)
    Preview abstract This paper proposes a neural end-to-end text-to-speech model which can control latent attributes in the generation of speech, that are rarely annotated in the training data (e.g. speaking styles, accents, background noise level, and recording conditions). The model is formulated as a conditional generative model with two levels of hierarchical latent variables. The first level is a categorical variable, which represents attribute groups (e.g. clean/noisy) and provides interpretability. The second level, conditioned on the first, is a multivariate Gaussian variable, which characterizes specific attribute configurations (e.g. noise level, speaking rate) and enables disentangled fine-grained control over these attributes. This amounts to using a Gaussian mixture model (GMM) for the latent distribution. Extensive evaluation of the proposed model demonstrates its ability to control the aforementioned attributes. In particular, it is capable of consistently synthesizing high-quality clean speech regardless of the quality of the training data for the target speaker. View details
    Preview abstract We present a multispeaker, multilingual text-to-speech (TTS) synthesis model based on Tacotron that is able to produce high quality speech in multiple languages. Moreover, the model is able to transfer voices across languages, e.g. synthesize fluent Spanish speech using an English speaker's voice, without training on any bilingual or parallel examples. Such transfer works across distantly related languages, e.g. English and Mandarin. Critical to achieving this result are: 1. using a phonemic input representation to encourage sharing of model capacity across languages, and 2. incorporating an adversarial loss term to encourage the model to disentangle its representation of speaker identity (which is perfectly correlated with language in the training data) from the speech content. Further scaling up the model by training on multiple speakers of each language, and incorporating an autoencoding input to help stabilize attention during training, results in a model which can be used to consistently synthesize intelligible speech for training speakers in all languages seen during training, and in native or foreign accents. View details
    Preview abstract This paper introduces a new speech corpus called ``LibriTTS'' designed for text-to-speech use. It is derived from the original audio and text materials of the LibriSpeech corpus, which has been used for training and evaluating automatic speech recognition systems. The new corpus inherits desired properties of the LibriSpeech corpus while addressing a number of issues which make LibriSpeech less than ideal for text-to-speech work. The released corpus consists of 585 hours of speech data at 24kHz sampling rate from 2,456 speakers and the corresponding texts. Experimental results show that neural end-to-end TTS models trained from the LibriTTS corpus achieved above 4.0 in mean opinion scores in naturalness in five out of six evaluation speakers. The corpus is freely available for download from http://www.openslr.org/60/. View details
    Preview abstract Many Japanese text-to-speech (TTS) systems use word-level pitch accents as one of their prosodic features. Combination of a pronunciation dictionary including lexical pitch accents and a statistical model representing the word accent sandhi is often used to predict pitch accents from a text. However, using human transcribers to build the dictionary and training data for the model is tedious and expensive. This paper proposes a neural pitch accent recognition model. This model combines the information from audio, and its transcription (word sequence in hiragana characters) via two-dimensional attention and outputs word-level pitch accents. Experimental results show a reduction in the word pitch accent prediction error rate over that with text only. It lowers the load of human annotators when building a pronunciation dictionary. As the approach is general, it can be used to do pronunciation learning in other languages as well. View details
    テキスト音声合成技術の変遷と最先端
    日本音響学会誌, vol. 74-7 (2018), pp. 387-393
    Preview abstract テキスト音声合成 (Text-to-Speech Synthesis; TTS) とは,任意の文章 (テキスト) に対応する音声波形を合成することである.テキスト音声合成技術は,計算機資源の増大とともに,専門家による音声生成の先験的知識に基づいた規則的手法から,大規模データベースに基づいた統計的手法に移行してきた. 統計的手法には,音声データベース内の自然音声の波形を接続することで合成音を得る波形接続型 (Concatenative TTS) と,データより統計モデルを学習し,ここから直接合成音を出力する生成モデル型 (Generative TTS) がある.近年,機械学習・特に深層学習が生成モデル型に導入され,合成音の自然性が大きく向上した.また,高い自然性を保ったまま話者性の変更や感情の付与が可能になり,応用範囲が大きく広がっている.本稿では,テキスト音声合成技術の変遷及び最先端の動向と,著者が考える今後の研究課題を報告する. View details
    Parallel WaveNet: Fast High-Fidelity Speech Synthesis
    Aäron van den Oord
    Yazhe Li
    Igor Babuschkin
    Karen Simonyan
    Koray Kavukcuoglu
    George van den Driessche
    Luis Carlos Cobo Rus
    Florian Stimberg
    Norman Casagrande
    Dominik Grewe
    Seb Noury
    Sander Dieleman
    Erich Elsen
    Alexander Graves
    Helen King
    Thomas Walters
    Demis Hassabis
    NA, Google Deepmind, NA (2017)
    Preview abstract The recently-developed WaveNet architecture [27] is the current state of the art in realistic speech synthesis, consistently rated as more natural sounding for many different languages than any previous system. However, because WaveNet relies on sequential generation of one audio sample at a time, it is poorly suited to today’s massively parallel computers, and therefore hard to deploy in a real-time production setting. This paper introduces Probability Density Distillation, a new method for training a parallel feed-forward network from a trained WaveNet with no significant difference in quality. The resulting system is capable of generating high-fidelity speech samples at more than 20 times faster than real-time, and is deployed online by Google Assistant, including serving multiple English and Japanese voices. View details
    Preview abstract Recent progress in generative modeling has improved the naturalness of synthesized speech significantly. In this talk I will summarize these generative model-based approaches for speech synthesis and describe possible future directions. View details
    WaveNet: A Generative Model for Raw Audio
    Aäron van den Oord
    Sander Dieleman
    Karen Simonyan
    Alexander Graves
    Koray Kavukcuoglu
    Arxiv (2016)
    Preview abstract This paper introduces WaveNet, a deep generative neural network trained end-to-end to model raw audio waveforms, which can be applied to text-to-speech and music generation. Current approaches to text-to-speech are focused on non-parametric, example-based generation (which stitches together short audio signal segments from a large training set), and parametric, model-based generation (in which a model generates acoustic features synthesized into a waveform with a vocoder). In contrast, we show that directly generating wideband audio signals at tens of thousands of samples per second is not only feasible, but also achieves results that significantly outperform the prior art. A single trained WaveNet can be used to generate different voices by conditioning on the speaker identity. We also show that the same approach can be used for music audio generation and speech recognition. View details
    Directly Modeling Voiced and Unvoiced Components in Speech Waveforms by Neural Networks
    Keiichi Tokuda
    Proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), IEEE (2016), pp. 5640-5644
    Preview abstract This paper proposes a novel acoustic model based on neural networks for statistical parametric speech synthesis. The neural network outputs parameters of a non-zero mean Gaussian process, which defines a probability density function of a speech waveform given linguistic features. The mean and covariance functions of the Gaussian process represent deterministic (voiced) and stochastic (unvoiced) components of a speech waveform, whereas the previous approach considered the unvoiced component only. Experimental results show that the proposed approach can generate speech waveforms approximating natural speech waveforms. View details
    Fast, Compact, and High Quality LSTM-RNN Based Statistical Parametric Speech Synthesizers for Mobile Devices
    Yannis Agiomyrgiannakis
    Niels Egberts
    Przemysław Szczepaniak
    Proc. Interspeech, San Francisco, CA, USA (2016), pp. 2273-2277
    Preview abstract Acoustic models based on long short-term memory recurrent neural network (LSTM-RNN) were applied to statistical parametric speech synthesis (SPSS) and showed significant improvements in naturalness and latency over those based on hidden Markov models (HMMs). This paper describes further optimizations of LSTM-RNN-based SPSS to deploy it to mobile devices; weight quantization, multi-frame inference, and robust inference using an ε-contaminated Gaussian loss function. Experimental results in subjective listening tests show that these optimizations can make LSTM-RNN-based SPSS comparable to HMM-based SPSS in runtime speed while maintaining naturalness. Evaluations between LSTM-RNN-based SPSS and HMM-driven unit selection speech synthesis are also presented. View details
    Preview abstract This paper introduces a general and flexible framework for FO and aperiodicity analysis, specifically intended for high-quality speech synthesis and modification applications. The proposed framework consists of three subsystems: instantaneous frequency estimator and initial aperiodicity detector, FO trajectory tracker, and FO refinement and aperiodicity extractor. A preliminary implementation of the proposed framework substantially outperformed (1/5 to 1/10 in terms of RMS FO estimation error) existing FO extractors in tracking ability of temporally varying FO trajectories. The front end aperiodicity detector consists of a complex-valued wavelet analysis filter with a highly selective temporal and spectral envelope. This front end aperiodicity detector uses a new measure that quantifies the deviation from periodicity. The measure is less sensitive to slow FM and AM and closely correlates with the signal to noise ratio. The front end combines instantaneous frequency information over a set of filter outputs using the measure to yield an observation probability map. The second stage generates the initial FO trajectory using this map and signal power information. The final stage uses the deviation measure of each harmonic component and FO adaptive time warping to refine the FO estimate and aperiodicity estimation. The proposed framework is flexible to integrate other sources of instantaneous frequency when they provide relevant information. View details
    Preview abstract Building text-to-speech (TTS) systems requires large amounts of high quality speech recordings and annotations, which is a challenge to collect especially considering the variation in spoken languages around the world. Acoustic modeling techniques that could utilize inhomogeneous data are hence important as they allow us to pool more data for training. This paper presents a long short-term memory (LSTM) recurrent neural network (RNN) based statistical parametric speech synthesis system that uses data from multiple languages and speakers. It models language variation through cluster adaptive training and speaker variation with speaker dependent output layers. Experimental results have shown that the proposed multilingual TTS system can synthesize speech in multiple languages from a single model while maintaining naturalness. Furthermore, it can be adapted to new languages with only a small amount of data. View details
    Preview abstract Statistical parametric speech synthesis (SPSS) combines an acoustic model and a vocoder to render speech given a text. Typically decision tree-clustered context-dependent hidden Markov models (HMMs) are employed as the acoustic model, which represent a relationship between linguistic and acoustic features. There have been attempts to replace the HMMs by alternative acoustic models, which provide trajectory and context modeling. Recently, artificial neural network-based acoustic models, such as deep neural networks, mixture density networks, and recurrent neural networks (RNNs), showed significant improvements over the HMM-based one. This talk reviews the progress of acoustic modeling in SPSS from the HMM to the RNN. View details
    Preview abstract Statistical parametric speech synthesis (SPSS) combines an acoustic model and a vocoder to render speech given a text. Typically decision tree-clustered context-dependent hidden Markov models (HMMs) are employed as the acoustic model, which represent a relationship between linguistic and acoustic features. Recently, artificial neural network-based acoustic models, such as deep neural networks, mixture density networks, and long short-term memory recurrent neural networks (LSTM-RNNs), showed significant improvements over the HMM-based approach. This paper reviews the progress of acoustic modeling in SPSS from the HMM to the LSTM-RNN. View details
    Statistical parametric speech synthesis: from HMM to LSTM-RNN
    RTTH Summer School on Speech Technology -- A Deep Learning Perspective, Barcelona, Spain (2015)
    Preview abstract This talk will present progress of acoustic modeling in statistical parametric speech synthesis from the conventional hidden Markov model HMM to the state-of-the-art long short-term memory recurrent neural network. The details of implementation and applications of statistical parametric speech synthesis are also included. View details
    Directly Modeling Speech Waveforms by Neural Networks for Statistical Parametric Speech Synthesis
    Keiichi Tokuda
    Proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), IEEE (2015), pp. 4215-4219
    Preview abstract This paper proposes a novel approach for directly-modeling speech at the waveform level using a neural network. This approach uses the neural network-based statistical parametric speech synthesis framework with a specially designed output layer. As acoustic feature extraction is integrated to acoustic model training, it can overcome the limitations of conventional approaches, such as two-step (feature extraction and acoustic modeling) optimization, use of spectra rather than waveforms as targets, use of overlapping and shifting frames as unit, and fixed decision tree structure. Experimental results show that the proposed approach can directly maximize the likelihood defined at the waveform domain. View details
    Deep Learning for Acoustic Modeling in Parametric Speech Generation: A systematic review of existing techniques and future trends
    Zhen-Hua Ling
    Shiyin Kang
    Mike Schuster
    Xiao-Jun Qian
    Helen Meng
    Li Deng
    IEEE Signal Processing Magazine, vol. 32 (2015), pp. 35-52
    Preview abstract Hidden Markov models (HMMs) and Gaussian mixture models (GMMs) are the two most common types of acoustic models used in statistical parametric approaches for generating low-level speech waveforms from high-level symbolic inputs via intermediate acoustic feature sequences. However, these models have their limitations in representing complex, nonlinear relationships between the speech generation inputs and the acoustic features. Inspired by the intrinsically hierarchical process of human speech production and by the successful application of deep neural networks (DNNs) to automatic speech recognition (ASR), deep learning techniques have also been applied successfully to speech generation, as reported in recent literature. View details
    Unidirectional Long Short-Term Memory Recurrent Neural Network with Recurrent Output Layer for Low-Latency Speech Synthesis
    Proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), IEEE (2015), pp. 4470-4474
    Preview abstract Long short-term memory recurrent neural networks (LSTM-RNNs) have been applied to various speech applications including acoustic modeling for statistical parametric speech synthesis. One of the concerns for applying them to text-to-speech applications is its effect on latency. To address this concern, this paper proposes a low-latency, streaming speech synthesis architecture using unidirectional LSTM-RNNs with a recurrent output layer. The use of unidirectional RNN architecture allows frame-synchronous streaming inference of output acoustic features given input linguistic features. The recurrent output layer further encourages smooth transition between acoustic features at consecutive frames. Experimental results in subjective listening tests show that the proposed architecture can synthesize natural sounding speech without requiring utterance-level batch processing. View details
    Deep Mixture Density Networks for Acoustic Modeling in Statistical Parametric Speech Synthesis
    Proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), IEEE (2014), pp. 3872-3876
    Preview abstract Statistical parametric speech synthesis (SPSS) using deep neural networks (DNNs) has shown its potential to produce naturally-sounding synthesized speech. However, there are limitations in the current implementation of DNN-based acoustic modeling for speech synthesis, such as the unimodal nature of its objective function and its lack of ability to predict variances. To address these limitations, this paper investigates the use of a mixture density output layer. It can estimate full probability density functions over real-valued output features conditioned on the corresponding input features. Experimental results in objective and subjective evaluations show that the use of the mixture density output layer improves the prediction accuracy of acoustic features and the naturalness of the synthesized speech. View details
    Statistical Parametric Speech Synthesis
    UKSpeech Conference, Edinburgh, UK (2014)
    Preview abstract Statistical parametric speech synthesis has grown in popularity over the last years. In this tutorial, its system architecture is outlined, and then basic techniques used in the system, including algorithms for speech parameter generation, are described with simple examples. View details
    Statistical Parametric Speech Synthesis Using Deep Neural Networks
    Mike Schuster
    Proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), IEEE (2013), pp. 7962-7966
    Preview abstract Conventional approaches to statistical parametric speech synthesis typically use decision tree-clustered context-dependent hidden Markov models (HMMs) to represent probability densities of speech parameters given texts. Speech parameters are generated from the probability densities to maximize their output probabilities, then a speech waveform is reconstructed from the generated parameters. This approach is reasonably effective but has a couple of limitations, e.g. decision trees are inef?cient to model complex context dependencies. This paper examines an alternative scheme that is based on a deep neural network (DNN). The relationship between input texts and their acoustic realizations is modeled by a DNN. The use of the DNN can address some limitations of the conventional approach. Experimental results show that the DNN-based systems outperformed the HMM-based systems with similar numbers of parameters. View details
    Deep Learning in Speech Synthesis
    8th ISCA Speech Synthesis Workshop, Barcelona, Spain (2013)
    Preview abstract Deep learning has been a hot research topic in various machine learning related areas including general object recognition and automatic speech recognition. This talk will present recent applications of deep learning to statistical parametric speech synthesis and contrast the deep learning-based approaches to the existing hidden Markov model-based one. View details
    Statistical Parametric Speech Synthesis Based on Speaker and Language Factorization
    Norbert Braunschweiler
    Sabine Buchholz
    Mark J. F. Gales
    Kate Knill
    Sacha Krstulovic
    Javier Latorre
    IEEE Transactions on Audio, Speech, and Language Processing, vol. 20 (2012), pp. 1713-1724
    Product of Experts for Statistical Parametric Speech Synthesis
    Mark J. F. Gales
    Yoshihiko Nankaku
    Keiichi Tokuda
    IEEE Transactions on Audio, Speech, and Language Processing, vol. 20 (2012), pp. 794-805
    Continuous Stochastic Feature Mapping Based on Trajectory HMMs
    Yoshihiko Nankaku
    Keiichi Tokuda
    IEEE Transactions on Audio, Speech, and Language Processing, vol. 19 (2011), pp. 417-430
    The HMM-Based Speech Synthesis System (HTS)
    Keiichi Tokuda
    Computer Processing of Asian Spoken Languages, Americas Group Publications (2010)
    Statistical Parametric Speech Synthesis
    Keiichi Tokuda
    Alan W. Black
    Speech Communication, vol. 51 (2009), pp. 1039-1064
    The Nitech-NAIST HMM-Based Speech Synthesis System for the Blizzard Challenge 2006
    Tomoki Toda
    Keiichi Tokuda
    IEICE Transactions on Information and Systems, vol. E91-D (2008), pp. 1764-1773
    A Hidden Semi-Markov Model-Based Speech Synthesis System
    Keiichi Tokuda
    Takashi Masuko
    Takao Kobayashi
    Tadashi Kitamura
    IEICE Transactions on Information and Systems, vol. E90-D (2007), pp. 825-834
    Details of Nitech HMM-Based Speech Synthesis System for the Blizzard Challenge 2005
    Tomoki Toda
    Masaru Nakamura
    Keiichi Tokuda
    IEICE Transactions on Information and Systems, vol. E90-D (2007), pp. 325-333
    Reformulating the HMM as a Trajectory Model by Imposing Explicit Relationships Between Static and Dynamic Feature Vector Sequences
    Keiichi Tokuda
    Tadashi Kitamura
    Computer Speech and Language, vol. 21 (2007), pp. 153-173
    The HMM-based Speech Synthesis System (HTS) Version 2.0
    Takashi Nose
    Junichi Yamagishi
    Shinji Sako
    Takashi Masuko
    Alan W. Black
    Keiichi Tokuda,
    ISCA SSW7 (2007)
    HMM-Based Approach to Multilingual Speech Synthesis
    Keiichi Tokuda
    Alan W. Black
    Text to Speech Synthesis: New Paradigms and Advances, Prentice Hall (2004)