Generating diverse and natural text-to-speech samples using a quantized fine-grained VAE and auto-regressive prosody prior
This work addresses the issue of prosodic variation in TTS for applications requiring diverse and natural speech synthesis, representing an incremental improvement over existing methods.
The paper tackled the problem of unnatural and discontinuous speech in text-to-speech models using fine-grained VAEs by proposing a sequential prior in a discrete latent space with vector quantization and an autoregressive model, resulting in significantly improved naturalness in random sample generation and potential ASR performance gains through data augmentation.
Recent neural text-to-speech (TTS) models with fine-grained latent features enable precise control of the prosody of synthesized speech. Such models typically incorporate a fine-grained variational autoencoder (VAE) structure, extracting latent features at each input token (e.g., phonemes). However, generating samples with the standard VAE prior often results in unnatural and discontinuous speech, with dramatic prosodic variation between tokens. This paper proposes a sequential prior in a discrete latent space which can generate more naturally sounding samples. This is accomplished by discretizing the latent features using vector quantization (VQ), and separately training an autoregressive (AR) prior model over the result. We evaluate the approach using listening tests, objective metrics of automatic speech recognition (ASR) performance, and measurements of prosody attributes. Experimental results show that the proposed model significantly improves the naturalness in random sample generation. Furthermore, initial experiments demonstrate that randomly sampling from the proposed model can be used as data augmentation to improve the ASR performance.