Originally Posted by SunshineReggae
Not so much about doubting why. Was just wondering if it was also researched (I'm guessing logically it has to be since it's what science is based on). Was just curious if there was any published research which backed that up. Thanks for sharing.
David demonstrated the distortion chamber of horrors to AES chapters and other audio groups around the country. I'm not sure if an AES paper on distortion audibility came out of it or not, but the use of a distortion generating circuit along with Clark's ABX test system in an effort to determine the threshold of distortion audibility in music is mentioned in Mr. Clark's paper, "High Resolution Subjective Testing", Journal of the AES, Vol. 30, Number 5, May, 1982. With his specific distortion circuit, the best lowest level that could be reliably detected was 3%. The paper did not focus on distortion audibility, however, and a preference for lower distortion was not mentioned. There are many papers written about the audible thresholds of specific types of distortions. But there's no simple way to generalize, since every distortion mechanism and resulting spectra is different.
It should be noted that audio devices have always been designed with lower distortion as a goal, and those with higher distortion have it as an undesired or unintentional by-product of a design compromise.
A cursory search through the AES database revealed dozens of papers about measuring distortion, reducing distortion, correcting distortion, the audible thresholds of different types of distortion, correlation (or lack thereof) of mathematical predictions of distortion vs measured levels, and so on. I found not one paper that attempted to analyzed if higher or lower distortion was preferred by listeners. It's almost as if it is a given that lower distortion is a goal because it's always preferred. There is no logical reason why this should be otherwise, as a distortion product adds spectra that does not exist in the original.
This doesn't specifically isolate distortion, but the findings seem to support the general preference for neutral undistorted sound:
Edited by jaddie - 5/5/13 at 11:51pm