The NU Voice Conversion System for the Voice Conversion Challenge 2020: On the Effectiveness of Sequence-to-sequence Models and Autoregressive Neural Vocoders
Nagoya University（名古屋大学）
Wen-Chin Huang, Patrick Lumban Tobing, Yi-Chiao Wu, Kazuhiro Kobayashi, Tomoki Toda
备注：
Accepted to the ISCA Joint Workshop for the Blizzard Challenge and Voice Conversion Challenge 2020
In this paper, we present the voice conversion (VC) systems developed at Nagoya University (NU) for the Voice Conversion Challenge 2020 (VCC2020). We aim to determine the effectiveness of two recent significant technologies in VC: sequence-to-sequence (seq2seq) models and autoregressive (AR) neural vocoders. Two respective systems were developed for the two tasks in the challenge: for task 1, we adopted the Voice Transformer Network, a Transformer-based seq2seq VC model, and extended it with synthetic parallel data to tackle nonparallel data; for task 2, we used the frame-based cyclic variational autoencoder (CycleVAE) to model the spectral features of a speech waveform and the AR WaveNet vocoder with additional fine-tuning. By comparing with the baseline systems, we confirmed that the seq2seq modeling can improve the conversion similarity and that the use of AR vocoders can improve the naturalness of the converted speech.
Custom voice is to construct a personal speech synthesis system by adapting the source speech synthesis model to the target model through the target few recordings. The solution to constructing a custom voice is to combine an adaptive acoustic model with a robust vocoder. However, training a robust vocoder usually requires a multi-speaker dataset, which should include various age groups and various timbres, so that the trained vocoder can be used for unseen speakers. Collecting such a multi-speaker dataset is difficult, and the dataset distribution always has a mismatch with the distribution o…
Graduate School of Informatics, Nagoya University（名古屋大学信息学研究科）；Graduate School of Information Science, Nagoya University（名古屋大学情报科学研究科）；Information Technology Center, Nagoya University（名古屋大学信息技术中心）
In this paper, we investigate the effectiveness of a quasi-periodic WaveNet (QPNet) vocoder combined with a statistical spectral conversion technique for a voice conversion task. The WaveNet (WN) vocoder has been applied as the waveform generation module in many different voice conversion frameworks and achieves significant improvement over conventional vocoders. However, because of the fixed dilated convolution and generic network architecture, the WN vocoder lacks robustness against unseen input features and often requires a huge network size to achieve acceptable speech quality. Such limita…
This paper presents a refinement framework of WaveNet vocoders for variational autoencoder (VAE) based voice conversion (VC), which reduces the quality distortion caused by the mismatch between the training data and testing data. Conventional WaveNet vocoders are trained with natural acoustic features but conditioned on the converted features in the conversion stage for VC, and such a mismatch often causes significant quality and similarity degradation. In this work, we take advantage of the particular structure of VAEs to refine WaveNet vocoders with the self-reconstructed features generated …
National University of Singapore（新加坡国立大学）；Nanyang Technological University（南洋理工大学）
In a typical voice conversion system, vocoder is commonly used for speech-to-features analysis and features-to-speech synthesis. However, vocoder can be a source of speech quality degradation. This paper presents a vocoder-free voice conversion approach using WaveNet for non-parallel training data. Instead of dealing with the intermediate features, the proposed approach utilizes the WaveNet to map the Phonetic PosteriorGrams (PPGs) to the waveform samples directly. In this way, we avoid the estimation errors caused by vocoder and feature conversion. Additionally, as PPG is assumed to be speake…
Graduate School of Culture Technology, KAIST（韩国科学技术院文化技术研究生院）
Singing voice separation (SVS) is a task that separates singing voice audio from its mixture with instrumental audio. Previous SVS studies have mainly employed the spectrogram masking method which requires a large dimensionality in predicting the binary masks. In addition, they focused on extracting a vocal stem that retains the wet sound with the reverberation effect. This result may hinder the reusability of the isolated singing voice. This paper addresses the issues by predicting mel-spectrogram of dry singing voices from the mixed audio as neural vocoder features and synthesizing the singi…
Recent progress in deep generative models has improved the quality of neural vocoders in speech domain. However, generating a high-quality singing voice remains challenging due to a wider variety of musical expressions in pitch, loudness, and pronunciations. In this work, we propose a hierarchical diffusion model for singing voice neural vocoders. The proposed method consists of multiple diffusion models operating in different sampling rates; the model at the lowest sampling rate focuses on generating accurate low-frequency components such as pitch, and other models progressively generate the …
Graduate School of Informatics, Nagoya University（名古屋大学信息学研究科）；Information Technology Center, Nagoya University（名古屋大学信息技术中心）；Graduate School of Information Science, Nagoya University（名古屋大学信息科学研究科）
In this paper, we integrate a simple non-parallel voice conversion (VC) system with a WaveNet (WN) vocoder and a proposed collapsed speech suppression technique. The effectiveness of WN as a vocoder for generating high-fidelity speech waveforms on the basis of acoustic features has been confirmed in recent works. However, when combining the WN vocoder with a VC system, the distorted acoustic features, acoustic and temporal mismatches, and exposure bias usually lead to significant speech quality degradation, making WN generate some very noisy speech segments called collapsed speech. To tackle t…
We present a universal neural vocoder based on Parallel WaveNet, with an additional conditioning network called Audio Encoder. Our universal vocoder offers real-time high-quality speech synthesis on a wide range of use cases. We tested it on 43 internal speakers of diverse age and gender, speaking 20 languages in 17 unique styles, of which 7 voices and 5 styles were not exposed during training. We show that the proposed universal vocoder significantly outperforms speaker-dependent vocoders overall. We also show that the proposed vocoder outperforms several existing neural vocoder architectures…
This paper introduces the Multi-Band Excited WaveNet a neural vocoder for speaking and singing voices. It aims to advance the state of the art towards an universal neural vocoder, which is a model that can generate voice signals from arbitrary mel spectrograms extracted from voice signals. Following the success of the DDSP model and following the development of the recently proposed excitation vocoders we propose a vocoder structure consisting of multiple specialized DNN that are combined with dedicated signal processing components. All components are implemented as differentiable operators an…
Mel-frequency filter bank (MFB) based approaches have the advantage of learning speech compared to raw spectrum since MFB has less feature size. However, speech generator with MFB approaches require additional vocoder that needs a huge amount of computation expense for training process. The additional pre/post processing such as MFB and vocoder is not essential to convert real human speech to others. It is possible to only use the raw spectrum along with the phase to generate different style of voices with clear pronunciation. In this regard, we propose a fast and effective approach to convert…
Nagoya University（名古屋大学）；Meta Reality Labs Research（Meta现实实验室）
Our previous work, the unified source-filter GAN (uSFGAN) vocoder, introduced a novel architecture based on the source-filter theory into the parallel waveform generative adversarial network to achieve high voice quality and pitch controllability. However, the high temporal resolution inputs result in high computation costs. Although the HiFi-GAN vocoder achieves fast high-fidelity voice generation thanks to the efficient upsampling-based generator architecture, the pitch controllability is severely limited. To realize a fast and pitch-controllable high-fidelity neural vocoder, we introduce th…
Academia Sinica（中央研究院）；Taiwan AI Labs（台湾人工智能实验室）；National Tsing Hua Univ.（国立清华大学）；Music Group（470音乐集团）；PARC（帕洛阿尔托研究中心）
A vocoder is a conditional audio generation model that converts acoustic features such as mel-spectrograms into waveforms. Taking inspiration from Differentiable Digital Signal Processing (DDSP), we propose a new vocoder named SawSing for singing voices. SawSing synthesizes the harmonic part of singing voices by filtering a sawtooth source signal with a linear time-variant finite impulse response filter whose coefficients are estimated from the input mel-spectrogram by a neural network. As this approach enforces phase continuity, SawSing can generate singing voices without the phase-discontinu…
Budapest University of Technology and Economics（布达佩斯技术与经济大学）；MTA-ELTE Lendület Lingual Articulation Research Group（MTA-ELTE伦杜莱特语言发音研究组）
Traditional vocoder-based statistical parametric speech synthesis can be advantageous in applications that require low computational complexity. Recent neural vocoders, which can produce high naturalness, still cannot fulfill the requirement of being real-time during synthesis. In this paper, we experiment with our earlier continuous vocoder, in which the excitation is modeled with two one-dimensional parameters: continuous F0 and Maximum Voiced Frequency. We show on the data of 9 speakers that an average voice can be trained for DNN-TTS, and speaker adaptation is feasible 400 utterances (abou…
In this paper, we propose a parallel WaveGAN (PWG)-like neural vocoder with a quasi-periodic (QP) architecture to improve the pitch controllability of PWG. PWG is a compact non-autoregressive (non-AR) speech generation model, whose generative speed is much faster than real time. While utilizing PWG as a vocoder to generate speech on the basis of acoustic features such as spectral and prosodic features, PWG generates high-fidelity speech. However, when the input acoustic features include unseen pitches, the pitch accuracy of PWG-generated speech degrades because of the fixed and generic network…
College of Electrical Engineering and Computer Science, National Taiwan University（台湾大学电机资讯学院）；Graduate Institute of Communication Engineering, National Taiwan University（台湾大学电信工程学研究所）
Recently, neural vocoders have been widely used in speech synthesis tasks, including text-to-speech and voice conversion. However, when encountering data distribution mismatch between training and inference, neural vocoders trained on real data often degrade in voice quality for unseen scenarios. In this paper, we train four common neural vocoders, including WaveNet, WaveRNN, FFTNet, Parallel WaveGAN alternately on five different datasets. To study the robustness of neural vocoders, we evaluate the models using acoustic features from seen/unseen speakers, seen/unseen languages, a text-to-speec…
Recently, autoregressive neural vocoders have provided remarkable performance in generating high-fidelity speech and have been able to produce synthetic speech in real-time. However, autoregressive neural vocoders such as WaveFlow are capable of modeling waveform signals from mel-spectrogram, its number of parameters is significant to deploy on edge devices. Though NanoFlow, which has a small number of parameters, is a state-of-the-art autoregressive neural vocoder, the performance of NanoFlow is marginally lower than WaveFlow. Therefore, we propose a new type of autoregressive neural vocoder …
Google Research（谷歌研究院）；Tokyo University of Agriculture and Technology（东京农工大学）
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 \textit{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 o…
Institute of Acoustics, Chinese Academy of Sciences（中国科学院声学研究所）；University of Chinese Academy of Sciences（中国科学院大学）
Voice conversion models have developed for decades, and current mainstream research focuses on non-streaming voice conversion. However, streaming voice conversion is more suitable for practical application scenarios than non-streaming voice conversion. In this paper, we propose a streaming any-to-many voice conversion based on fully non-autoregressive model, which includes a streaming transformer based acoustic model and a streaming vocoder. Streaming transformer based acoustic model is composed of a pre-trained encoder from streaming end-to-end based automatic speech recognition model and a d…
Recently, GAN vocoders have seen rapid progress in speech synthesis, starting to outperform autoregressive models in perceptual quality with much higher generation speed. However, autoregressive vocoders are still the common choice for neural generation of speech signals coded at very low bit rates. In this paper, we present a GAN vocoder which is able to generate wideband speech waveforms from parameters coded at 1.6 kbit/s. The proposed model is a modified version of the StyleMelGAN vocoder that can run in frame-by-frame manner, making it suitable for streaming applications. The experimental…
vec2wav 2.0: Advancing Voice Conversion via Discrete Token Vocoders
X-LANCE Lab（X-LANCE实验室）；MoE Key lab of Artificial Intelligence（人工智能关键实验室）；School of Computer Science（计算机科学学院）；Shanghai Jiao Tong University（上海交通大学）；Jiangsu Key Lab of Language Computing（江苏语言计算关键实验室）；Shenzhen Research Institute of Big Data（深圳大数据研究 institute）
In this paper, we propose a technique to alleviate the quality degradation caused by collapsed speech segments sometimes generated by the WaveNet vocoder. The effectiveness of the WaveNet vocoder for generating natural speech from acoustic features has been proved in recent works. However, it sometimes generates very noisy speech with collapsed speech segments when only a limited amount of training data is available or significant acoustic mismatches exist between the training and testing data. Such a limitation on the corpus and limited ability of the model can easily occur in some speech gen…
In this work, we propose ParaNet, a non-autoregressive seq2seq model that converts text to spectrogram. It is fully convolutional and brings 46.7 times speed-up over the lightweight Deep Voice 3 at synthesis, while obtaining reasonably good speech quality. ParaNet also produces stable alignment between text and speech on the challenging test sentences by iteratively improving the attention in a layer-by-layer manner. Furthermore, we build the parallel text-to-speech system and test various parallel neural vocoders, which can synthesize speech from text through a single feed-forward pass. We al…
Budapest University of Technology and Economics（布达佩斯技术与经济大学）；MTA-SZTE Research Group on Artificial Intelligence（匈牙利科学院-塞格德大学人工智能研究组）；University of Szeged（塞格德大学）
Recently it was shown that within the Silent Speech Interface (SSI) field, the prediction of F0 is possible from Ultrasound Tongue Images (UTI) as the articulatory input, using Deep Neural Networks for articulatory-to-acoustic mapping. Moreover, text-to-speech synthesizers were shown to produce higher quality speech when using a continuous pitch estimate, which takes non-zero pitch values even when voicing is not present. Therefore, in this paper on UTI-based SSI, we use a simple continuous F0 tracker which does not apply a strict voiced / unvoiced decision. Continuous vocoder parameters (Cont…
Budapest University of Technology and Economics（布达佩斯技术与经济大学）；MTA-ELTE Lendület Lingual Articulation Research Group（MTA-ELTE Lendület 语言发音研究组）
To date, various speech technology systems have adopted the vocoder approach, a method for synthesizing speech waveform that shows a major role in the performance of statistical parametric speech synthesis. WaveNet one of the best models that nearly resembles the human voice, has to generate a waveform in a time consuming sequential manner with an extremely complex structure of its neural networks.
This paper presents a new approach for a vocoder design based on full frequency masking by octaves in addition to a technique for spectral filling via beta probability distribution. Some psycho-acoustic characteristics of human hearing - inaudibility masking in frequency and phase - are used as a basis for the proposed algorithm. The results confirm that this technique may be useful to save bandwidth in applications requiring intelligibility. It is recommended for the legal eavesdropping of long voice conversations.
With recent advancements in voice cloning, the performance of speech synthesis for a target speaker has been rendered similar to the human level. However, autoregressive voice cloning systems still suffer from text alignment failures, resulting in an inability to synthesize long sentences. In this work, we propose a variant of attention-based text-to-speech system that can reproduce a target voice from a few seconds of reference speech and generalize to very long utterances as well. The proposed system is based on three independently trained components: a speaker encoder, synthesizer and unive…
Nowadays more and more applications can benefit from edge-based text-to-speech (TTS). However, most existing TTS models are too computationally expensive and are not flexible enough to be deployed on the diverse variety of edge devices with their equally diverse computational capacities. To address this, we propose FBWave, a family of efficient and scalable neural vocoders that can achieve optimal performance-efficiency trade-offs for different edge devices. FBWave is a hybrid flow-based generative model that combines the advantages of autoregressive and non-autoregressive models. It produces …
This paper proposes a novel voice conversion (VC) method based on non-autoregressive sequence-to-sequence (NAR-S2S) models. Inspired by the great success of NAR-S2S models such as FastSpeech in text-to-speech (TTS), we extend the FastSpeech2 model for the VC problem. We introduce the convolution-augmented Transformer (Conformer) instead of the Transformer, making it possible to capture both local and global context information from the input sequence. Furthermore, we extend variance predictors to variance converters to explicitly convert the source speaker's prosody components such as pitch an…
Voice conversion has gained increasing popularity in many applications of speech synthesis. The idea is to change the voice identity from one speaker into another while keeping the linguistic content unchanged. Many voice conversion approaches rely on the use of a vocoder to reconstruct the speech from acoustic features, and as a consequence, the speech quality heavily depends on such a vocoder. In this paper, we propose NVC-Net, an end-to-end adversarial network, which performs voice conversion directly on the raw audio waveform of arbitrary length. By disentangling the speaker identity from …
The packet loss problem seriously affects the quality of service in Voice over IP (VoIP) sceneries. In this paper, we investigated online receiver-based packet loss concealment which is much more portable and applicable. For ensuring the speech naturalness, rather than directly processing time-domain waveforms or separately reconstructing amplitudes and phases in frequency domain, a flow-based neural vocoder is adopted to generate the substitution waveform of lost packet from Mel-spectrogram which is generated from history contents by a well-designed neural predictor. Furthermore, a waveform s…
Evaluation of the Speech Resynthesis Capabilities of the VoicePrivacy Challenge Baseline B1
International Audio Laboratories Erlangen（埃尔朗根国际音频实验室）；Friedrich-Alexander-Universität Erlangen-Nürnberg（弗里德里希-亚历山大-埃尔朗根-纽伦堡大学）；University of Erlangen-Nürnberg（埃尔朗根-纽伦堡大学）；Fraunhofer IIS（弗劳恩霍夫集成电路研究所）
Denoising diffusion probabilistic models (diffusion models for short) require a large number of iterations in inference to achieve the generation quality that matches or surpasses the state-of-the-art generative models, which invariably results in slow inference speed. Previous approaches aim to optimize the choice of inference schedule over a few iterations to speed up inference. However, this results in reduced generation quality, mainly because the inference process is optimized separately, without jointly optimizing with the training process. In this paper, we propose InferGrad, a diffusio…
Recently, the effectiveness of text-to-speech (TTS) systems combined with neural vocoders to generate high-fidelity speech has been shown. However, collecting the required training data and building these advanced systems from scratch are time and resource consuming. An economical approach is to develop a neural vocoder to enhance the speech generated by existing or low-cost TTS systems. Nonetheless, this approach usually suffers from two issues: 1) temporal mismatches between TTS and natural waveforms and 2) acoustic mismatches between training and testing data. To address these issues, we ad…
Aalto University School of Science（阿alto大学科学学院）；Aalto University（阿alto大学）；School of Data Science, The Chinese University of Hong Kong, Shenzhen（香港中文大学（深圳）数据科学学院）；Spellbrush, Akihabara, Tokyo（东京秋叶原Spellbrush）
In this work, we investigate the effectiveness of two techniques for improving variational autoencoder (VAE) based voice conversion (VC). First, we reconsider the relationship between vocoder features extracted using the high quality vocoders adopted in conventional VC systems, and hypothesize that the spectral features are in fact F0 dependent. Such hypothesis implies that during the conversion phase, the latent codes and the converted features in VAE based VC are in fact source F0 dependent. To this end, we propose to utilize the F0 as an additional input of the decoder. The model can learn …
Despite recent progress in generative adversarial network (GAN)-based vocoders, where the model generates raw waveform conditioned on acoustic features, it is challenging to synthesize high-fidelity audio for numerous speakers across various recording environments. In this work, we present BigVGAN, a universal vocoder that generalizes well for various out-of-distribution scenarios without fine-tuning. We introduce periodic activation function and anti-aliased representation into the GAN generator, which brings the desired inductive bias for audio synthesis and significantly improves audio qual…
The use of modern vocoders in an analysis/synthesis pipeline allows us to investigate high-quality voice conversion that can be used for privacy purposes. Here, we propose to transform the speaker embedding and the pitch in order to hide the sex of the speaker. ECAPA-TDNN-based speaker representation fed into a HiFiGAN vocoder is protected using a neural-discriminant analysis approach, which is consistent with the zero-evidence concept of privacy. This approach significantly reduces the information in speech related to the speaker's sex while preserving speech content and some consistency in t…
Northwestern Polytechnical University（西北工业大学）；Tencent AI Lab（腾讯人工智能实验室）
The zero-shot scenario for speech generation aims at synthesizing a novel unseen voice with only one utterance of the target speaker. Although the challenges of adapting new voices in zero-shot scenario exist in both stages -- acoustic modeling and vocoder, previous works usually consider the problem from only one stage. In this paper, we extend our previous Glow-WaveGAN to Glow-WaveGAN 2, aiming to solve the problem from both stages for high-quality zero-shot text-to-speech and any-to-any voice conversion. We first build a universal WaveGAN model for extracting latent distribution $p(z)$ of s…
Several of the latest GAN-based vocoders show remarkable achievements, outperforming autoregressive and flow-based competitors in both qualitative and quantitative measures while synthesizing orders of magnitude faster. In this work, we hypothesize that the common factor underlying their success is the multi-resolution discriminating framework, not the minute details in architecture, loss function, or training strategy. We experimentally test the hypothesis by evaluating six different generators paired with one shared multi-resolution discriminating framework. For all evaluative measures with …
Most neural vocoders employ band-limited mel-spectrograms to generate waveforms. If full-band spectral features are used as the input, the vocoder can be provided with as much acoustic information as possible. However, in some models employing full-band mel-spectrograms, an over-smoothing problem occurs as part of which non-sharp spectrograms are generated. To address this problem, we propose UnivNet, a neural vocoder that synthesizes high-fidelity waveforms in real time. Inspired by works in the field of voice activity detection, we added a multi-resolution spectrogram discriminator that empl…
The generative adversarial network (GAN) has shown its outstanding capability in improving Non-Autoregressive TTS (NAR-TTS) by adversarially training it with an extra model that discriminates between the real and the generated speech. To maximize the benefits of GAN, it is crucial to find a powerful discriminator that can capture rich distinguishable information. In this paper, we propose a multi-scale time-frequency spectrogram discriminator to help NAR-TTS generate high-fidelity Mel-spectrograms. It treats the spectrogram as a 2D image to exploit the correlation among different components in…
Sun Yat-Sen University（中山大学）；MoE Key Lab of Artificial Intelligence, AI Institute（教育部人工智能重点实验室，人工智能研究院）
Recent studies have shown that neural vocoders based on generative adversarial network (GAN) can generate audios with high quality. While GAN based neural vocoders have shown to be computationally much more efficient than those based on autoregressive predictions, the real-time generation of the highest quality audio on CPU is still a very challenging task. One major computation of all GAN-based neural vocoders comes from the stacked upsampling layers, which were designed to match the length of the waveform's length of output and temporal resolution. Meanwhile, the computational complexity of …
This paper proposes a spectral-domain perceptual weighting technique for Parallel WaveGAN-based text-to-speech (TTS) systems. The recently proposed Parallel WaveGAN vocoder successfully generates waveform sequences using a fast non-autoregressive WaveNet model. By employing multi-resolution short-time Fourier transform (MR-STFT) criteria with a generative adversarial network, the light-weight convolutional networks can be effectively trained without any distillation process. To further improve the vocoding performance, we propose the application of frequency-dependent weighting to the MR-STFT …
This paper presents a neural vocoder named HiNet which reconstructs speech waveforms from acoustic features by predicting amplitude and phase spectra hierarchically. Different from existing neural vocoders such as WaveNet, SampleRNN and WaveRNN which directly generate waveform samples using single neural networks, the HiNet vocoder is composed of an amplitude spectrum predictor (ASP) and a phase spectrum predictor (PSP). The ASP is a simple DNN model which predicts log amplitude spectra (LAS) from acoustic features. The predicted LAS are sent into the PSP for phase recovery. Considering the is…
