set vs otl, which is the most musical

I'd like to comment on something [email protected] said back on July 3rd:

"Ten percent distortion can't be taken seriously as an audiophile product as it's a profound sonic coloration."

Well you would think so, but that's not necessarily the case. A recent pair of Audio Engineering Society papers presented by Earl Geddes and Lidia Lee showed what essentially amounts to a NEGATIVE correlation between harmonic distortion percentage and listener preference!

Thirty percent second harmonic distortion was found to be statistically undetectable, while very low percentages of high order harmonic distortion such as is generated by global negative feedback in an amplifier were both audible and objectionable. In addition, "crossover distortion" such as is typically produced by amplifiers operating in Class B mode was found to be particularly audible and objectionable. For years amplifier designers have been getting rid of large percentages of low order distortion and replacing it with small percentages of high order distortion, and evidently this is often a step in the wrong direction sonically.

If that ten percent distortion is second order harmonics, it is probably inaudible or barely audible, and is very unlikely to be perceived as an objectionable coloration.

I don't mean to be picking on you, Webmaster. I'd have said the exact same thing (and probably have many times) had I not been fairly close to the study and taken the listening test myself. My scores were not included in the data used because I knew something about the focus of the trials, so I was not a totally blind participant.

The papers were presented about two years ago, and are only available through the Audio Engineering Society.

Getting back to the question at hand, I'd say that with appropriate speaker/amplifier matching both approaches can sound very, very, very good.

Duke

Another very important point is that any SET is only producing 10% 2nd-order HD at its max output or very close to it. For this reason, I do think it's a good idea to plan a system with the SETs not required to go about 50% or so output. At that level, you're normally talking about a few percent THD, 90% 2nd harmonic, proven in testing to be entirely inaudible.

One study showed 7th order HD to be around 20x more objectionable than 2nd order.

Thanks for your insightful responses, Audiokinesis and Paulofbrecht. What you said was certainly pretty shocking to me! I did some googling to try to find those papers you were talking about by Geddes and Lee, but I have not been able to come up with anything. Would you happen to have some useful links?

Amfibius, the papers by Geddes and Lee I mentioned were presented at the 115th convention of the Audio Engineering Society, in October 2003. You can order them from the Society. They are entitled "Auditory Perception of Nonlinear Distortion - Theory" and "Auditory Perception of Nonlinear Distortion", and can be ordered as Preprint Number 5890 and Preprint Number 5891.

If you are an Audio Engineerng Society member, you can read them online or print them out.

Duke