Although the common procedure for measuring sound quality is based on subjective listening tests, this approach is labor intensive, requiring a large number of trained subjects and specialized facilities. If a model-based approach could approximate subjective tests, quality measurement could be automated. This study examines the degree to which objective measures from auditory models might improve the prediction of quality ratings by normal-hearing subjects with different types of distortion. Using a wide variety of audio data, the authors compared two sound quality assessment models based on PEMO-Q and CASP and one extended version of PEMO-Q to state-of-the-art quality-assessment models. The results revealed that the suggested quality models were able to predict a high range of different distortions and performed best compared to other state-of-the-art quality measures. In general, quality measures employing valid auditory models generalized best across different kinds of distortions.
Harlander, Niklas; Huber, Rainer; Ewert, Stephan D.
Affiliations: Medizinische Physik and Cluster of Excellence Hearing4All, Universität Oldenburg, Oldenburg, Germany; HörTech gGmbH and Cluster of Excellence Hearing4All, Oldenburg, Germany(See document for exact affiliation information.)
JAES Volume 62 Issue 5 pp. 324-336; May 2014
Publication Date: June 2, 2014
No AES members have commented on this paper yet.
If you are not yet an AES member and have something important to say about this paper then we urge you to join the AES today and make your voice heard. You can join online today by clicking here.