Estimating the Significance of Errors in Audio Systems
A companion paper evolved a method for estimating the detectability, audibility and loudness of noises and distortions by calculating masked excitation and loudness patterns for a signal as well as returning a single-number loudness estimation. This first step concentrated on steady-state measurements. How do we transfer such knowledge to predict the significance of errors on programme material? To succeed even partially would improve the fact/fantasy ratio! With this goal, the author concentrates on several relevant non-linear perceptual phenomena and mechanisms. Guidelines are then developed for minimally audible distortions, noise, noise modulations and other errors. Central to estimating the significance of an error is to account for loudness and spatial context, and to quantify the non-linear perceptual processes involved in the formation of external acoustic objects. The paper addresses difficulties of A/B and nulling tests.
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!
This paper costs $33 for non-members and is temporarily free for AES members.