Sorry, you need to enable JavaScript to visit this website.

This paper presents a technique to interpret and visualize intermediate layers in generative CNNs trained on raw speech data in an unsupervised manner. We argue that averaging over feature maps after ReLU activation in each transpose convolutional layer yields interpretable time-series data. This technique allows for acoustic analysis of intermediate layers that parallels the acoustic analysis of human speech data: we can extract F0, intensity, duration, formants, and other acoustic properties from intermediate layers in order to test where and how CNNs encode various types of information.

Generative deep neural networks are widely used for speech synthesis, but most existing models directly generate waveforms or spectral outputs. Humans, however, produce speech by controlling articulators, which results in the production of speech sounds through physical properties of sound propagation. We introduce the Articulatory Generator to the Generative Adversarial Network paradigm, a new unsupervised generative model of speech production/synthesis.

Listening to spoken content often requires modifying the speech rate while preserving the timbre and pitch of the speaker. To date, advanced signal processing techniques are used to address this task, but it still remains a challenge to maintain a high speech quality at all time-scales. Inspired by the success of speech generation using Generative Adversarial Networks (GANs), we propose a novel unsupervised learning algorithm for time-scale modification (TSM) of speech, called ScalerGAN. The model is trained using a set of speech utterances, where no time-scales are provided.

This study proposes a Wave-U-Net discriminator, which is a single but expressive discriminator that assesses a waveform in a sample-wise manner with the same resolution as the input signal while extracting multilevel features via an encoder and decoder with skip connections. The experimental results demonstrate that a Wave-U-Net discriminator can be used as an alternative to a typical ensemble of discriminators while maintaining speech quality, reducing the model size, and accelerating the training speed.

In recent years, the quality of text-to-speech (TTS) synthesis vastly improved due to deep-learning techniques, with parallel architectures, in particular, providing excellent synthesis quality at fast inference. Training these models usually requires speech recordings, corresponding phoneme-level transcripts, and the temporal alignment of each phoneme to the utterances. Since manually creating such fine-grained alignments requires expert knowledge and is time-consuming, it is common practice to estimate them using automatic speech–phoneme alignment methods.

We propose novel deep speaker representation learning that considers perceptual similarity among speakers for multi-speaker generative modeling. Following its success in accurate discriminative modeling of speaker individuality, knowledge of deep speaker representation learning (i.e., speaker representation learning using deep neural networks) has been introduced to multi-speaker generative modeling.

Speech-to-text alignment is a critical component of neural text-to-speech (TTS) models. Autoregressive TTS models typically use an attention mechanism to learn these alignments on-line. However, these alignments tend to be brittle and often fail to generalize to long utterances and out-of-domain text, leading to missing or repeating words. Most non-autoregressive end-to-end TTS models rely on durations extracted from external sources. In this paper we leverage the alignment mechanism proposed in RAD-TTS and demonstrate its applicability to wide variety of neural TTS models.

We propose iSTFTNet, which replaces some output-side layers of the mel-spectrogram vocoder with the inverse short-time Fourier transform (iSTFT) after sufficiently reducing the frequency dimension using upsampling layers, reducing the computational cost from black-box modeling and avoiding redundant estimations of high-dimensional spectrograms. During our experiments, we applied our ideas to three HiFi-GAN variants and made the models faster and more lightweight with a reasonable speech quality.

To address the issue of one-to-many mapping from phoneme sequences to acoustic features in expressive speech synthesis, this paper proposes a method of discourse-level prosody modeling with a variational autoencoder (VAE) based on the non-autoregressive architecture of FastSpeech. In this method, phone-level prosody codes are extracted from prosody features by combining VAE with FastSpeech, and are predicted using discourse-level text features together with BERT embeddings. The continuous wavelet transform (CWT) in FastSpeech2 for F0 representation is not necessary anymore.