On The Effect Of Coding Artifacts On Acoustic Scene Classification

TL;DR

This paper investigates how perceptual audio coding affects acoustic scene classification accuracy, revealing significant degradation with compression and proposing training methods to mitigate this effect for resource-constrained devices.

Contribution

It demonstrates the impact of perceptual audio coding on classification performance and introduces training strategies to improve accuracy on compressed audio signals.

Findings

Classification accuracy degrades up to 57% with audio compression.

Lossy compression techniques during training improve accuracy on compressed audio.

Training with compressed audio enhances robustness across different codecs and bitrates.

Abstract

Previous DCASE challenges contributed to an increase in the performance of acoustic scene classification systems. State-of-the-art classifiers demand significant processing capabilities and memory which is challenging for resource-constrained mobile or IoT edge devices. Thus, it is more likely to deploy these models on more powerful hardware and classify audio recordings previously uploaded (or streamed) from low-power edge devices. In such scenario, the edge device may apply perceptual audio coding to reduce the transmission data rate. This paper explores the effect of perceptual audio coding on the classification performance using a DCASE 2020 challenge contribution [1]. We found that classification accuracy can degrade by up to 57% compared to classifying original (uncompressed) audio. We further demonstrate how lossy audio compression techniques during model training can improve classification accuracy of compressed audio signals even for audio codecs and codec bitrates not included in the training process.