On the Semantic Latent Space of Diffusion-Based Text-to-Speech Models
This work addresses the challenge of semantic control in TTS for users needing editable speech synthesis, but it is incremental as it builds on existing diffusion models and latent space methods from image synthesis.
The paper tackles the problem of controlling vocal properties in diffusion-based text-to-speech models by exploring their latent space, identifying semantic directions that enable off-the-shelf audio editing without additional training or data.
The incorporation of Denoising Diffusion Models (DDMs) in the Text-to-Speech (TTS) domain is rising, providing great value in synthesizing high quality speech. Although they exhibit impressive audio quality, the extent of their semantic capabilities is unknown, and controlling their synthesized speech's vocal properties remains a challenge. Inspired by recent advances in image synthesis, we explore the latent space of frozen TTS models, which is composed of the latent bottleneck activations of the DDM's denoiser. We identify that this space contains rich semantic information, and outline several novel methods for finding semantic directions within it, both supervised and unsupervised. We then demonstrate how these enable off-the-shelf audio editing, without any further training, architectural changes or data requirements. We present evidence of the semantic and acoustic qualities of the edited audio, and provide supplemental samples: https://latent-analysis-grad-tts.github.io/speech-samples/.