Dynamic Temporal Alignment of Speech to Lips

TL;DR

This paper introduces an automated deep learning-based method for aligning re-recorded speech with lip movements in movies, improving accuracy especially when original audio quality is poor or ambiguous.

Contribution

It presents a novel audio-visual shared representation for precise lip-sync alignment, outperforming existing methods in challenging audio conditions.

Findings

Successful quantitative and qualitative alignment results

Outperforms state-of-the-art methods in unclear audio scenarios

Effective in cases where traditional shift-based alignment fails

Abstract

Many speech segments in movies are re-recorded in a studio during postproduction, to compensate for poor sound quality as recorded on location. Manual alignment of the newly-recorded speech with the original lip movements is a tedious task. We present an audio-to-video alignment method for automating speech to lips alignment, stretching and compressing the audio signal to match the lip movements. This alignment is based on deep audio-visual features, mapping the lips video and the speech signal to a shared representation. Using this shared representation we compute the lip-sync error between every short speech period and every video frame, followed by the determination of the optimal corresponding frame for each short sound period over the entire video clip. We demonstrate successful alignment both quantitatively, using a human perception-inspired metric, as well as qualitatively. The strongest advantage of our audio-to-video approach is in cases where the original voice in unclear, and where a constant shift of the sound can not give a perfect alignment. In these cases state-of-the-art methods will fail.