A Large-Scale Database and Predictive Model of Listener-Rated Ease of Speech Understanding in Commercial Hearing Aids
Andrew Sabin, Steve Taddei, Abram Bailey · 2026
Read A Large-Scale Database and Predictive Model of Listener-Rated Ease of Speech Understanding in Commercial Hearing Aids by Andrew Sabin, Steve Taddei, Abram Bailey (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-26342.txt.
Description
Andrew Sabin, Steve Taddei, and Abram Bailey build a hearing-aid metric from the listeners who actually wear the aids. HearAdvisor rates hearing aids for consumers. Its speech metric was built to predict lab intelligibility under simulated distortions. Whether it matches what wearers find easy to understand was unclear. Their study replaces guesswork with scale: visitors with self-reported hearing loss rated recordings blind, on a five-point ease scale.
Ratings top one hundred thousand after quality screening. Recordings span eighty-three commercial products across seventy-two realistic scenes, captured on an acoustic manikin. A learned model compares aided audio against clean speech inside a frozen Whisper encoder and predicts the ratings. On unseen devices it beats the old metric clearly, scene by scene. In loud scenes it reaches the reliability ceiling of the listeners themselves. Consumer judgement, captured at scale, becomes the yardstick.
Catalog record
| CREATOR | Andrew Sabin, Steve Taddei, Abram Bailey |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 22 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2606.26342