Let There Be Sound: Reconstructing High Quality Speech from Silent Videos
This work solves the problem of generating accurate and natural speech from lip movements for applications like assistive technologies or video editing, representing a strong specific gain rather than a foundational breakthrough.
The paper tackles the problem of reconstructing high-quality speech from silent videos by addressing the one-to-many mapping challenge in lip-to-speech systems, resulting in speech generation quality close to real human utterance and outperforming existing methods in naturalness and intelligibility by a large margin.
The goal of this work is to reconstruct high quality speech from lip motions alone, a task also known as lip-to-speech. A key challenge of lip-to-speech systems is the one-to-many mapping caused by (1) the existence of homophenes and (2) multiple speech variations, resulting in a mispronounced and over-smoothed speech. In this paper, we propose a novel lip-to-speech system that significantly improves the generation quality by alleviating the one-to-many mapping problem from multiple perspectives. Specifically, we incorporate (1) self-supervised speech representations to disambiguate homophenes, and (2) acoustic variance information to model diverse speech styles. Additionally, to better solve the aforementioned problem, we employ a flow based post-net which captures and refines the details of the generated speech. We perform extensive experiments on two datasets, and demonstrate that our method achieves the generation quality close to that of real human utterance, outperforming existing methods in terms of speech naturalness and intelligibility by a large margin. Synthesised samples are available at our demo page: https://mm.kaist.ac.kr/projects/LTBS.