Neural Spectral Band Generation for Audio Coding

TL;DR

This paper introduces a neural network-based method for high-frequency audio band generation that improves perceptual quality and reduces side information compared to traditional spectral band replication techniques.

Contribution

It proposes a deep neural network framework for spectral band generation in audio coding, optimized with adversarial training for perceptual plausibility.

Findings

Outperforms HE-AAC-v1 in perceptual quality

Uses less side information than traditional methods

Effective neural generative approach for high-frequency audio synthesis

Abstract

Spectral band replication (SBR) enables bit-efficient coding by generating high-frequency bands from the low-frequency ones. However, it only utilizes coarse spectral features upon a subband-wise signal replication, limiting adaptability to diverse acoustic signals. In this paper, we explore the efficacy of a deep neural network (DNN)-based generative approach for coding the high-frequency bands, which we call neural spectral band generation (n-SBG). Specifically, we propose a DNN-based encoder-decoder structure to extract and quantize the side information related to the high-frequency components and generate the components given both the side information and the decoded core-band signals. The whole coding pipeline is optimized with generative adversarial criteria to enable the generation of perceptually plausible sound. From experiments using AAC as the core codec, we show that the proposed method achieves a better perceptual quality than HE-AAC-v1 with much less side information.