Speech enhancement systems differ in listener preferences and discrimination
Perceptual Differences and Preferences in Speech Enhancement
Sound
Summary
It can be hard to know if speech enhancement tools not only improve sound but also if people can tell the difference or like one better than another. The authors checked how well everyday listeners could tell apart and prefer different enhanced speech samples compared to clean speech. They found that listeners’ ability to spot changes and their preferences depend on the system and background noise. Popular automatic measures sometimes matched these human judgments but not always, suggesting these tests catch different things than usual methods.
What this means in practice
- •For audio engineers: Use perceptual discrimination and preference tests to better evaluate speech enhancement tools beyond standard metrics.
- •For voice assistant developers: Improve speech enhancement features by incorporating listener preferences and discriminability into system tuning processes.
Authors
Simon Dahl Jepsen, Jesper Rindom Jensen
Abstract
Speech enhancement (SE) systems are commonly evaluated with objective metrics, with and without references, and with categorical subjective listening tests, all measuring speech quality and intelligibility. So far, it is unclear from such evaluations whether the processed speech is perceptually different or is preferred by everyday listeners. We investigate discrimination and preference as complementary dimensions of SE evaluation through crowdsourced triangle and pairwise preference tests across four SE systems, with comparisons to clean references. Results show that both listeners' ability to distinguish and whether they exhibit systematic preference vary across systems and substantially across acoustic conditions. Pairwise evaluations of objective metrics showed variable correlation with both discriminability and preference, motivating further studies into these novel aspects in speech enhancement evaluation.