Thank you. That's very interesting, especially the description of how traffic studies have improved with technology.
We can all agree there's an enormous amount of bullshit in hi-fi. I believe in subjectivism because at the end of the day it's the sound I listen to. I don't believe in sighted, uncontrolled subjectivism. And I definitely don't believe in snake oil. I just happen to think the interface between subjective performance and technical performance is complex and murky and not as pat-and-dry as this forum makes it seem sometimes.
Henry, you make a good point. We’ve got the terminology screwed up. Preference testing is by definition subjective testing, but when done right it controls for sighted bias to find out 1.) the ability to distinguish what cannot be seen, and 2.) a resulting preference for one or the other.
Without the first, the second has no meaning. And without both, there is no basis for recommending one’s experience as being instructive to another. The control against sighted biases does not make preference testing any less subjective. I try to use terms like “controlled testing” when taking about preferences.
But lots of the preference studies have been done already, and with amps above a rather low threshold, the first hypothesis—that there is a difference—often can’t be demonstrated. Which makes the second hypothesis void. The limitation is operating amps in their linear range, which is a problem with amps of very low power like this one.
If we are sure we can experience a phenomenon repeatably, by all means let’s construct a controlled test. If we can detect a difference, then we can explore what measurements will explain it. But it usually goes the other way, objective measurements vary by greater amounts than we expect given that we can’t tell the difference in controlled subjective testing.
Example: the Directivity Index resulted from Spinorama speaker testing and is the product of trying to explain controlled subjective preference results. That testing explained why speakers that were flat on-axis didn’t always perform well in a real room, and that it wasn’t always the room’s fault (or the amp’s). Objective and controlled subjective testing worked synergistically to improve products.
With amps, though, the only feature that I think marks the good from the great is quiescent output, and that is masked when music is playing. The very best amps are silent even with an ear to the speaker. Amps that are indistinguishable in preference testing but have a slight hiss when quiescent may call attention to themselves in a sensitive system and quiet room. My B&K Reference 125.2 amps hiss when I’m within a foot or two—maybe 80 or 90 dB down and inaudible at any normal listening distance. A similarly powered modern Benchmark AHB2 won’t hiss like that.
In fact, the amp of this thread was very quiet in that dimension. It’s problem wasn’t noise floor but rather distortion, and the low power requires pushing it to the limit. It might be unable to achieve 30 dB RMS above the noise floor a normal listening environment would require without pushing the peaks past lineari. Testing with very sensitive speakers would answer that. And that may be why many like the thing in spite of itself.
But we both know that if it was a Rick Denney kit, threads full of people here and elsewhere would be laughing at it rather than praising and defending it. Are these people trusting their ears? Doesn’t seem so.
Rick “distinguishing between subjective opinion and controlled subjective testing” Denney