Saved in:
Bibliographic Details
Main Authors: Raj, Vishnu, KV, Gouthaman, Gehlot, Shiv, Villemoes, Lars, Biswas, Arijit
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2509.21463
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912842344562688
author Raj, Vishnu
KV, Gouthaman
Gehlot, Shiv
Villemoes, Lars
Biswas, Arijit
author_facet Raj, Vishnu
KV, Gouthaman
Gehlot, Shiv
Villemoes, Lars
Biswas, Arijit
contents We present GMLv2, a reference-based model designed for the prediction of subjective audio quality as measured by MUSHRA scores. GMLv2 introduces a Beta distribution-based loss to model the listener ratings and incorporates additional neural audio coding (NAC) subjective datasets to extend its generalization and applicability. Extensive evaluations on diverse testset demonstrate that proposed GMLv2 consistently outperforms widely used metrics, such as PEAQ and ViSQOL, both in terms of correlation with subjective scores and in reliably predicting these scores across diverse content types and codec configurations. Consequently, GMLv2 offers a scalable and automated framework for perceptual audio quality evaluation, poised to accelerate research and development in modern audio coding technologies.
format Preprint
id arxiv_https___arxiv_org_abs_2509_21463
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Enhanced Generative Machine Listener
Raj, Vishnu
KV, Gouthaman
Gehlot, Shiv
Villemoes, Lars
Biswas, Arijit
Audio and Speech Processing
Artificial Intelligence
Machine Learning
We present GMLv2, a reference-based model designed for the prediction of subjective audio quality as measured by MUSHRA scores. GMLv2 introduces a Beta distribution-based loss to model the listener ratings and incorporates additional neural audio coding (NAC) subjective datasets to extend its generalization and applicability. Extensive evaluations on diverse testset demonstrate that proposed GMLv2 consistently outperforms widely used metrics, such as PEAQ and ViSQOL, both in terms of correlation with subjective scores and in reliably predicting these scores across diverse content types and codec configurations. Consequently, GMLv2 offers a scalable and automated framework for perceptual audio quality evaluation, poised to accelerate research and development in modern audio coding technologies.
title Enhanced Generative Machine Listener
topic Audio and Speech Processing
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2509.21463