Objective Measures of Perceptual Audio Quality Reviewed: An Evaluation of Their Application Domain Dependence

TL;DR

This study evaluates how well various objective audio quality measures correlate with human perception across different domains, highlighting the importance of training data and auditory models for universal assessment.

Contribution

It introduces a new objective measure, SI-SA2f, and compares multiple existing measures across audio coding and source separation domains.

Findings

2f-model outperforms other measures in both domains

Training dataset and auditory model robustness are key for universal measures

Proposed SI-SA2f offers a new approach based on existing models

Abstract

Over the past few decades, computational methods have been developed to estimate perceptual audio quality. These methods, also referred to as objective quality measures, are usually developed and intended for a specific application domain. Because of their convenience, they are often used outside their original intended domain, even if it is unclear whether they provide reliable quality estimates in this case. This work studies the correlation of well-known state-of-the-art objective measures with human perceptual scores in two different domains: audio coding and source separation. The following objective measures are considered: fwSNRseg, dLLR, PESQ, PEAQ, POLQA, PEMO-Q, ViSQOLAudio, (SI-)BSSEval, PEASS, LKR-PI, 2f-model, and HAAQI. Additionally, a novel measure (SI-SA2f) is presented, based on the 2f-model and a BSSEval-based signal decomposition. We use perceptual scores from 7 listening tests about audio coding and 7 listening tests about source separation as ground-truth data for the correlation analysis. The results show that one method (2f-model) performs significantly better than the others on both domains and indicate that the dataset for training the method and a robust underlying auditory model are crucial factors towards a universal, domain-independent objective measure.