Hypernetworks build Implicit Neural Representations of Sounds

TL;DR

This paper introduces HyperSound, a novel meta-learning hypernetwork approach that generates implicit neural representations for audio, overcoming visual data biases and achieving high-quality sound reconstruction.

Contribution

The paper presents HyperSound, the first hypernetwork-based method for creating INRs for audio, enabling better generalization and quality in sound representation.

Findings

Achieves comparable quality to state-of-the-art audio models

Provides a new hypernetwork-based approach for audio INRs

Demonstrates effective generalization beyond training samples

Abstract

Implicit Neural Representations (INRs) are nowadays used to represent multimedia signals across various real-life applications, including image super-resolution, image compression, or 3D rendering. Existing methods that leverage INRs are predominantly focused on visual data, as their application to other modalities, such as audio, is nontrivial due to the inductive biases present in architectural attributes of image-based INR models. To address this limitation, we introduce HyperSound, the first meta-learning approach to produce INRs for audio samples that leverages hypernetworks to generalize beyond samples observed in training. Our approach reconstructs audio samples with quality comparable to other state-of-the-art models and provides a viable alternative to contemporary sound representations used in deep neural networks for audio processing, such as spectrograms.