Thanks very much for that link. I think I might've scanned that piece earlier, but now I read it more carefully. I certainly think Gallo is on the right track and I agree with everything that he says. I did not know he was absorbing so many extra costs in order to maintain the price point of the original Reference Threes and 3.1. It makes sense unfortunately that the speaker cost has virtually doubled. He clearly is a gifted designer and very thoughtful about design targets.
I do have some major questions though about the widespread assumption that absolute phase integrity (a virtually physically impossible design target in a discrete three-way system except perhaps in a tiny listening window) is nearly as important to the audible illusion of a music source compared to flat frequency response and some other things. In sources that are time coherent but not time aligned (such as newer BMW and KEF designs) the amount of delay between the arrival times of wave fronts coming from woofer versus mid range versus tweeter is on the order of 1 to 3 ms or so. Although virtually everyone would agree that it is theoretically desirable to have complete phase integrity, I don't believe anyone has demonstrated in double blind testing that a 1 to 3 ms phase smear is really truly audible. What is clearly audible on the other hand are things like transient response, and frequency response and intermodular distortion (many times more audible than classic harmonic distortion). Additionally, designs that prioritize phase alignment have to make sacrifices in other things that are audible (read the excellent Stereophile review of one of Thiel's speakers at http://stereophile.com/floorloudspeakers/1208thi/.
There is pretty good physical evidence that transient response and these other things are intrinsically related, in other words that great transient response predicts a flat frequency response and even a good phase response and relatively low IM distortion.
I think human beings are great at constructing plausible story-lines to explain something, but many of our plausible stories turn out not to be true. Science is literally littered with the corpses of dead highly plausible theories and we still have to test even our favorite ideas that seem totally commonsensical against some kind of empirical prediction. I have not seen a truly careful test double blinded of phase smear at very small time frames (have enough literature in my own area to review and stay on top of).
Part of the reason I'm skeptical is that I don't think there's much reason why evolution would've carved the ability to detect these things because they have very little application in the real world to basic mammalian survival tasks. Unless phase smear at very small time scales meant that we could not identify either the spatial location or the source of a noise (and these are clearly unaffected), is not clear to me that there would be any selection process for such abilities. I honestly don't know what the work is in this area, and I'm sure that phase smear becomes audible at some point (my guess is somewhere probably around 100 to 200 ms). But I believe that very few speakers have that degree of phase delay. Thiel has made a killing selling the idea that extremely small phase delays or mismatches make a big audible difference, and although I have considerable respect for his speakers and for him personally, I don't agree that there is compelling evidence for such an assumption. Of course it's possible that some people's brains can hear small degrees of phase misalignment while other people cannot.
Anybody know of any carefully done research in psychoacoustics to address this question? I'd be interested.
best, Doug
