no code implementations • 25 Mar 2024 • Takuhiro Kaneko, Hirokazu Kameoka, Kou Tanaka
A generative adversarial network (GAN)-based vocoder trained with an adversarial discriminator is commonly used for speech synthesis because of its fast, lightweight, and high-quality characteristics.
no code implementations • 14 Aug 2023 • Takuhiro Kaneko, Hirokazu Kameoka, Kou Tanaka, Shogo Seki
Owing to the difficulty of a 1D CNN to model high-dimensional spectrograms, the frequency dimension is reduced via temporal upsampling.
no code implementations • 24 Mar 2023 • Takuhiro Kaneko, Hirokazu Kameoka, Kou Tanaka, Shogo Seki
This architecture provides a generator with sufficiently rich information for the synthesized speech to be closely matched to the real speech.
no code implementations • 20 Oct 2022 • Chihiro Watanabe, Hirokazu Kameoka
In this paper, we propose a new variational-autoencoder-based voice conversion model accompanied by an auxiliary network, which ensures that the conversion result correctly reflects the specified F0/timbre information.
1 code implementation • 9 Jun 2022 • Kohei Suzuki, Shoki Sakamoto, Tadahiro Taniguchi, Hirokazu Kameoka
This paper proposes a new voice conversion (VC) task from human speech to dog-like speech while preserving linguistic information as an example of human to non-human creature voice conversion (H2NH-VC) tasks.
1 code implementation • 4 Mar 2022 • Takuhiro Kaneko, Kou Tanaka, Hirokazu Kameoka, Shogo Seki
In recent text-to-speech synthesis and voice conversion systems, a mel-spectrogram is commonly applied as an intermediate representation, and the necessity for a mel-spectrogram vocoder is increasing.
no code implementations • 10 Aug 2021 • Shoki Sakamoto, Akira Taniguchi, Tadahiro Taniguchi, Hirokazu Kameoka
Although this method is powerful, it can fail to preserve the linguistic content of input speech when the number of available training samples is extremely small.
Automatic Speech Recognition Automatic Speech Recognition (ASR) +3
no code implementations • 5 Apr 2021 • Asuka Moritani, Ryo Ozaki, Shoki Sakamoto, Hirokazu Kameoka, Tadahiro Taniguchi
Through subjective evaluation experiments, we evaluated the performance of our StarGAN-EVC system in terms of its ability to achieve EVC for Japanese phrases.
3 code implementations • 25 Feb 2021 • Takuhiro Kaneko, Hirokazu Kameoka, Kou Tanaka, Nobukatsu Hojo
With FIF, we apply a temporal mask to the input mel-spectrogram and encourage the converter to fill in missing frames based on surrounding frames.
2 code implementations • 22 Oct 2020 • Takuhiro Kaneko, Hirokazu Kameoka, Kou Tanaka, Nobukatsu Hojo
To address this, we examined the applicability of CycleGAN-VC/VC2 to mel-spectrogram conversion.
no code implementations • 18 Sep 2020 • Chihiro Watanabe, Hirokazu Kameoka
Particularly, it has been shown that a monaural speech separation task can be successfully solved with a DNN-based method called deep clustering (DC), which uses a DNN to describe the process of assigning a continuous vector to each time-frequency (TF) bin and measure how likely each pair of TF bins is to be dominated by the same speaker.
1 code implementation • 27 Aug 2020 • Hirokazu Kameoka, Takuhiro Kaneko, Kou Tanaka, Nobukatsu Hojo
We previously proposed a method that allows for nonparallel voice conversion (VC) by using a variant of generative adversarial networks (GANs) called StarGAN.
1 code implementation • 7 Aug 2020 • Wen-Chin Huang, Tomoki Hayashi, Yi-Chiao Wu, Hirokazu Kameoka, Tomoki Toda
Sequence-to-sequence (seq2seq) voice conversion (VC) models are attractive owing to their ability to convert prosody.
Automatic Speech Recognition Automatic Speech Recognition (ASR) +2
no code implementations • 18 May 2020 • Hirokazu Kameoka, Wen-Chin Huang, Kou Tanaka, Takuhiro Kaneko, Nobukatsu Hojo, Tomoki Toda
The main idea we propose is an extension of the original VTN that can simultaneously learn mappings among multiple speakers.
1 code implementation • 14 Dec 2019 • Wen-Chin Huang, Tomoki Hayashi, Yi-Chiao Wu, Hirokazu Kameoka, Tomoki Toda
We introduce a novel sequence-to-sequence (seq2seq) voice conversion (VC) model based on the Transformer architecture with text-to-speech (TTS) pretraining.
no code implementations • 5 Nov 2019 • Xin Wang, Junichi Yamagishi, Massimiliano Todisco, Hector Delgado, Andreas Nautsch, Nicholas Evans, Md Sahidullah, Ville Vestman, Tomi Kinnunen, Kong Aik Lee, Lauri Juvela, Paavo Alku, Yu-Huai Peng, Hsin-Te Hwang, Yu Tsao, Hsin-Min Wang, Sebastien Le Maguer, Markus Becker, Fergus Henderson, Rob Clark, Yu Zhang, Quan Wang, Ye Jia, Kai Onuma, Koji Mushika, Takashi Kaneda, Yuan Jiang, Li-Juan Liu, Yi-Chiao Wu, Wen-Chin Huang, Tomoki Toda, Kou Tanaka, Hirokazu Kameoka, Ingmar Steiner, Driss Matrouf, Jean-Francois Bonastre, Avashna Govender, Srikanth Ronanki, Jing-Xuan Zhang, Zhen-Hua Ling
Spoofing attacks within a logical access (LA) scenario are generated with the latest speech synthesis and voice conversion technologies, including state-of-the-art neural acoustic and waveform model techniques.
3 code implementations • 29 Jul 2019 • Takuhiro Kaneko, Hirokazu Kameoka, Kou Tanaka, Nobukatsu Hojo
To bridge this gap, we rethink conditional methods of StarGAN-VC, which are key components for achieving non-parallel multi-domain VC in a single model, and propose an improved variant called StarGAN-VC2.
6 code implementations • 9 Apr 2019 • Takuhiro Kaneko, Hirokazu Kameoka, Kou Tanaka, Nobukatsu Hojo
Non-parallel voice conversion (VC) is a technique for learning the mapping from source to target speech without relying on parallel data.
no code implementations • 9 Apr 2019 • Hirokazu Kameoka, Kou Tanaka, Aaron Valero Puche, Yasunori Ohishi, Takuhiro Kaneko
We use the latent code of an input face image encoded by the face encoder as the auxiliary input into the speech converter and train the speech converter so that the original latent code can be recovered from the generated speech by the voice encoder.
no code implementations • 5 Apr 2019 • Kou Tanaka, Hirokazu Kameoka, Takuhiro Kaneko, Nobukatsu Hojo
WaveCycleGAN has recently been proposed to bridge the gap between natural and synthesized speech waveforms in statistical parametric speech synthesis and provides fast inference with a moving average model rather than an autoregressive model and high-quality speech synthesis with the adversarial training.
no code implementations • 29 Mar 2019 • Shinji Takaki, Hirokazu Kameoka, Junichi Yamagishi
Recently, we proposed short-time Fourier transform (STFT)-based loss functions for training a neural speech waveform model.
no code implementations • 16 Dec 2018 • Li Li, Hirokazu Kameoka, Shoji Makino
While MVAE is notable in its impressive source separation performance, the convergence-guaranteed optimization algorithm and that it allows us to estimate source-class labels simultaneously with source separation, there are still two major drawbacks, i. e., the high computational complexity and unsatisfactory source classification accuracy.
no code implementations • 9 Nov 2018 • Kou Tanaka, Hirokazu Kameoka, Takuhiro Kaneko, Nobukatsu Hojo
This paper describes a method based on a sequence-to-sequence learning (Seq2Seq) with attention and context preservation mechanism for voice conversion (VC) tasks.
no code implementations • 5 Nov 2018 • Hirokazu Kameoka, Kou Tanaka, Damian Kwasny, Takuhiro Kaneko, Nobukatsu Hojo
Second, it achieves many-to-many conversion by simultaneously learning mappings among multiple speakers using only a single model instead of separately learning mappings between each speaker pair using a different model.
no code implementations • 29 Sep 2018 • Shogo Seki, Hirokazu Kameoka, Li Li, Tomoki Toda, Kazuya Takeda
This paper deals with a multichannel audio source separation problem under underdetermined conditions.
no code implementations • 25 Sep 2018 • Kou Tanaka, Takuhiro Kaneko, Nobukatsu Hojo, Hirokazu Kameoka
The experimental results demonstrate that our proposed method can 1) alleviate the over-smoothing effect of the acoustic features despite the direct modification method used for the waveform and 2) greatly improve the naturalness of the generated speech sounds.
2 code implementations • 13 Aug 2018 • Hirokazu Kameoka, Takuhiro Kaneko, Kou Tanaka, Nobukatsu Hojo
Such situations can be avoided by introducing an auxiliary classifier and training the encoder and decoder so that the attribute classes of the decoder outputs are correctly predicted by the classifier.
1 code implementation • 2 Aug 2018 • Hirokazu Kameoka, Li Li, Shota Inoue, Shoji Makino
This paper proposes a multichannel source separation technique called the multichannel variational autoencoder (MVAE) method, which uses a conditional VAE (CVAE) to model and estimate the power spectrograms of the sources in a mixture.
13 code implementations • 6 Jun 2018 • Hirokazu Kameoka, Takuhiro Kaneko, Kou Tanaka, Nobukatsu Hojo
This paper proposes a method that allows non-parallel many-to-many voice conversion (VC) by using a variant of a generative adversarial network (GAN) called StarGAN.
no code implementations • 6 Apr 2018 • Keisuke Oyamada, Hirokazu Kameoka, Takuhiro Kaneko, Kou Tanaka, Nobukatsu Hojo, Hiroyasu Ando
In this paper, we address the problem of reconstructing a time-domain signal (or a phase spectrogram) solely from a magnitude spectrogram.
1 code implementation • 3 Apr 2018 • Lauri Juvela, Bajibabu Bollepalli, Xin Wang, Hirokazu Kameoka, Manu Airaksinen, Junichi Yamagishi, Paavo Alku
This paper proposes a method for generating speech from filterbank mel frequency cepstral coefficients (MFCC), which are widely used in speech applications, such as ASR, but are generally considered unusable for speech synthesis.
9 code implementations • 30 Nov 2017 • Takuhiro Kaneko, Hirokazu Kameoka
A subjective evaluation showed that the quality of the converted speech was comparable to that obtained with a Gaussian mixture model-based method under advantageous conditions with parallel and twice the amount of data.