Deep Long Audio Inpainting

TL;DR

This paper explores deep learning approaches for long audio inpainting, addressing the challenge of recovering missing audio segments over 200 ms, and provides a systematic analysis and benchmark for the task.

Contribution

It pioneers the adaptation of deep learning frameworks from other domains to long audio inpainting and systematically analyzes factors affecting performance.

Findings

Analyzed effects of mask size, receptive field, and audio representation on inpainting quality.

Established a benchmark for long audio inpainting performance.

Proposed a systematic approach to evaluate deep learning models for audio inpainting.

Abstract

Long (> 200 ms) audio inpainting, to recover a long missing part in an audio segment, could be widely applied to audio editing tasks and transmission loss recovery. It is a very challenging problem due to the high dimensional, complex and non-correlated audio features. While deep learning models have made tremendous progress in image and video inpainting, audio inpainting did not attract much attention. In this work, we take a pioneering step, exploring the possibility of adapting deep learning frameworks from various domains inclusive of audio synthesis and image inpainting for audio inpainting. Also, as the first to systematically analyze factors affecting audio inpainting performance, we explore how factors ranging from mask size, receptive field and audio representation could affect the performance. We also set up a benchmark for long audio inpainting. The code will be available on GitHub upon accepted.