Shahinian Diapason loudspeaker system JGH Responds
After Dick Shahinian's apoplexy after reading the preprint of the foregoing had subsided, I was able to discuss with him why he had designed the speakers to measure the way they did.
It was intentional. The preliminary design was by the book, with all the right measurements in place. The final design, which produces the anomalous measurements JA got, was tailored to make the speakers sound more like three-dimensional reality on orchestral music than they did with a flat response.
But before launching into this, a brief recap for those who haven't seen my comments about the Diapasons in previous issues:
I have observed several times that large acoustical spaces, and surround-system systems capable of simulating (or reproducing) them, give instrumental sounds a natural richness and body that I find lacking from two-channel stereo reproduction. Prior to my in-home audition, I had heard the Diapasons many times in a number of different CES rooms during the past several years, and while I acknowledged their vague imaging and stage foreshortening, I had always been very impressed by their ability to create (in me) the impression of listening to a real orchestra in a real concert hall. Here was a two-channel system that gave orchestral music the weight and richness that I thought to be the exclusive domain of properly implemented ambient surround systems, but without the treble dullness that would normally result if that tonality was obtained by merely tilting a flat response to roll off progressively with rising frequency.
The warmth and roundness of a properly implemented surround system appears to be a psychoacoustic phenomenon, related to how our ears respond to multi-directional reverberant information. I have run numerous response measurements, from my listening seat, of two front channels alone vs fronts and rears combined, and have never found any difference that could account for the audible difference in tonality. The response stays essentially flat in both modes. But without the rear ambience, there is only one way of obtaining the same kind of spectral balance: by messing with the front speakers' frequency response. Which is what Shahinian has done.
To do it, it was necessary first to design-in the kind of tilted response JA and I had measured. Not surprisingly, this resulted in a dead, muffled high end. The solution to that was, obviously, a treble rise, whose amplitude had to be roughly equal to the system's average lower-range output in order to maintain correct LF/HF balance. But why the broad dip from 700Hz to 3kHz? Because without it, much of the warmth obtained by the tilted lower range would be lost, it would be necessary to further elevate the range below 700 to regain it, and the treble would sound muted again. Back to square one!
In other words, what we have here is a system embodying purposeful distortions which are intended to make orchestral reproduction sound more realistic than is possible from two channels which embody no such corrections. But what is realism anyway? It is whatever reminds a listener of what he hears at a live performance. But what does that for JA and for me are clearly different things. In my case, the first thing I react to is the reproduction of musical timbres. If they don't sound right, nothing else about the reproduction can ring my reality bell. To JA, the recognition that says "reality" is a convincing soundstage. And while tonal accuracy is important to him, just as soundstaging and detail are important to me, it is not, as it is for me, the first thing he responds to. Either is a perfectly valid definition of realism, but when it comes to something as intrinsically (and intentionally) colored as the Diapason, there is no possibility that we could agree about its realism.
Do the Diapason's colorations achieve their intended result? They may not do it for JA's hearing system, but they do it for mine and there is reason to believe they do it for a lot of other music lovers'. Shahinian's demos have traditionally been among the only rooms at CES where visitors who prefer symphonies to rock will sit mesmerized for hours listening to excerpts from every kind of symphonic recording. If it takes a highly colored system to do that for them, so be it.
What particularly bothers me about JA's very negative appraisal of the Diapasons (footnote 5) is that Stereophile would appear to be renouncing what Robert Harley has declared to be the very essence of high-end: the practice of designing a product to sound musical rather than to measure a certain way. The fact that this particular product measures unusually poorly should not automatically disqualify it as a reproducer of music, particularly when so many musical people find it so sonically rewarding.
But wait a minute...Could this writer be the same JGH who scolded the high-end industry some months ago for valuing euphony over accuracy? (footnote 6). Yes, it is. But the euphony I was decrying was the production of pleasant but unrealistic sounds. To my ears, that does not describe the Diapasons.
These speakers aren't for everyone, and I would be the last person to call them audiophile speakers. They don't do what audiophiles expect from speakers, and (dare I say it?) many audiophiles don't relate to realism anyway, particularly from a symphony orchestra. (When was the last time you heard a live symphony orchestra?) But for the music-lover whose main interest is symphonic, these could be the most satisfying loudspeakers available.
All of this begs the question of why the Diapasons sounded so intolerably steely in my listening room, when they have never sounded that way to me anywhere else. JA says above that he heard the same brightness I heard, but did he hear the same amount of it? He did not hear the speakers in my home, and I didn't hear them in his. I have also heard some of that brightness at Shahinian's show demos, but it always sounded much more natural than what I heard at home, which was just 'way over the top. (Brightness, remember, is a normal part of live music; only excessive brightness is a no-no.)
Because of a serious illness in his family, Dick has not as yet had the chance to investigate my original findings. But we're planning to get together soon, in Boulder or Santa Fe, after which I may be able to write a final conclusion to this at present inconclusive review (footnote 7).—J. Gordon Holt
Footnote 5: More than a decade later, I am still puzzled by J. Gordon Holt's characterization of my comments. In effect, my own listening notes parallel his, as reported in the first part of this review.—John Atkinson
Footnote 6: "Just What Is High End?," "As We See It," Vol.14 No.7.—J. Gordon Holt
Footnote 7: Richard Shahinian never did submit further samples of the Diapason system for Gordon or me to audition.—John Atkinson