ViT-TTS: Visual Text-to-Speech with Scalable Diffusion Transformer

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Liu, Huadai, Huang, Rongjie, Lin, Xuan, Xu, Wenqiang, Zheng, Maozong, Chen, Hong, He, Jinzheng, Zhao, Zhou
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866911846927171584
author Liu, Huadai
Huang, Rongjie
Lin, Xuan
Xu, Wenqiang
Zheng, Maozong
Chen, Hong
He, Jinzheng
Zhao, Zhou
author_facet Liu, Huadai
Huang, Rongjie
Lin, Xuan
Xu, Wenqiang
Zheng, Maozong
Chen, Hong
He, Jinzheng
Zhao, Zhou
contents Text-to-speech(TTS) has undergone remarkable improvements in performance, particularly with the advent of Denoising Diffusion Probabilistic Models (DDPMs). However, the perceived quality of audio depends not solely on its content, pitch, rhythm, and energy, but also on the physical environment. In this work, we propose ViT-TTS, the first visual TTS model with scalable diffusion transformers. ViT-TTS complement the phoneme sequence with the visual information to generate high-perceived audio, opening up new avenues for practical applications of AR and VR to allow a more immersive and realistic audio experience. To mitigate the data scarcity in learning visual acoustic information, we 1) introduce a self-supervised learning framework to enhance both the visual-text encoder and denoiser decoder; 2) leverage the diffusion transformer scalable in terms of parameters and capacity to learn visual scene information. Experimental results demonstrate that ViT-TTS achieves new state-of-the-art results, outperforming cascaded systems and other baselines regardless of the visibility of the scene. With low-resource data (1h, 2h, 5h), ViT-TTS achieves comparative results with rich-resource baselines.~\footnote{Audio samples are available at \url{https://ViT-TTS.github.io/.}}
format Preprint
id arxiv_https___arxiv_org_abs_2305_12708
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle ViT-TTS: Visual Text-to-Speech with Scalable Diffusion Transformer
Liu, Huadai
Huang, Rongjie
Lin, Xuan
Xu, Wenqiang
Zheng, Maozong
Chen, Hong
He, Jinzheng
Zhao, Zhou
Audio and Speech Processing
Sound
Text-to-speech(TTS) has undergone remarkable improvements in performance, particularly with the advent of Denoising Diffusion Probabilistic Models (DDPMs). However, the perceived quality of audio depends not solely on its content, pitch, rhythm, and energy, but also on the physical environment. In this work, we propose ViT-TTS, the first visual TTS model with scalable diffusion transformers. ViT-TTS complement the phoneme sequence with the visual information to generate high-perceived audio, opening up new avenues for practical applications of AR and VR to allow a more immersive and realistic audio experience. To mitigate the data scarcity in learning visual acoustic information, we 1) introduce a self-supervised learning framework to enhance both the visual-text encoder and denoiser decoder; 2) leverage the diffusion transformer scalable in terms of parameters and capacity to learn visual scene information. Experimental results demonstrate that ViT-TTS achieves new state-of-the-art results, outperforming cascaded systems and other baselines regardless of the visibility of the scene. With low-resource data (1h, 2h, 5h), ViT-TTS achieves comparative results with rich-resource baselines.~\footnote{Audio samples are available at \url{https://ViT-TTS.github.io/.}}
title ViT-TTS: Visual Text-to-Speech with Scalable Diffusion Transformer
topic Audio and Speech Processing
Sound
url https://arxiv.org/abs/2305.12708