In listening, Sony’s array works very well indeed. Cymbals and other high percussion have an unusual realism. Schedrin’s Carmen arrangement on Delos was positively stunning in its combination of delicacy and purity in the high percussion. The top notes of the piano have the glow of reality (cf. the Rachmaninoff on BIS already mentioned). And the treble altogether is very clean and clear. In fixed position measurements, there can be some small irregularities from the spaced multiple tweeters and a rise at the extreme top, but in listening, the irregularities are unobtrusive and the top integrates well, with no sense of separateness, just a little extra “air.” I count this tweeter array a major success in creating a convincing replication of the treble character of instruments with substantial content up top.
The Tonal Road Not Taken by Monitor Speakers
So far, the description outside of the tweeter array probably sounds just like what one would expect of a fairly small speaker using modern drivers. But there are aspects of this speaker that seem to me to enter another world entirely from the typical results seen by getting some good drivers, running a computer program to implement a crossover circuit, making a rigid cabinet, and then off to manufacturing. Thinking of this speaker in those terms would be a little like supposing that one could 3-D print a violin to good effect. For the Sonys partake of the character of a work of art, not just a work of technology.
Part of their unique sound is a contouring of frequency response. They are very smooth and flat up to around 1kHz. Then there is a slope down to a minimum around 3kHz, with a gradual return to level above that and finally quite high up, too high to affect tonal character as such, there is a rise above midrange level. The effect of all this is complex, especially as the combined upper-octave level fluctuates a bit from the spaced tweeters. The overall sound is somewhat subdued, decidedly non-aggressive. But as noted earlier, there is a lot of air in the upper octaves but no brightness in the region in between presence and real top. The effect on human speech is interesting. Voices are articulate but not especially present. And sometimes there is a little emphasis not on sibilance but on fricatives. This could be noted on a recording of the BBC Sherlock Holmes Boscombe Valley Mystery, for example. But the overall subdued sweetness, as I would call it, was evident too on all music material with any energy above the midrange.
The relationship of this to reality is not a simple matter. Everyone supposes that speakers that are exactly flat will reproduce recordings as accurately as possible. But this is only true in the sense of being true by definition. In actuality, the relationship between microphone pick-up and the real sound heard is made less than straightforward by the fact that microphone pickup typically detect diffuse field sound as well as direct arrival, but in playback the sound becomes direct arrival from the speakers with rather little diffuse field. Since the ear’s response to diffuse field and to frontal direct arrival are quite different, especially around 3kHz, to reproduce the natural, real sound via microphone into speakers involves, as it happens, pushing down 3kHz. Siegfried Linkwitz on his website suggests that as much as 4dB is appropriate. As I understand it, the Sonys were optimized for naturalness by listening, but as it happened listening produced a response dip around 3kHz to roughly the theoretical extent that the difference in diffuse field and direct arrival would give (which is roughly the same as Linkwitz’s figure as well).
Of course this is the kind of thing that makes people who approach audio in an ultra-conformist, almost fascist, spirit really livid. “Flat or else” is a motto that appeals to some people. I have to admit that I myself am rather partial to flat speakers in general and find that playing around with frequency response usually leads to colored sound that does not seem real. But the 3kHz issue is always to be dealt with. My own “reference” system is quite flat, with just a little warmth in the lower mids/upper bass, cf. the room response here http://www.regonaudio.com/Harbeth%20Monitor%2040.html, which has a little (but only a little) relaxation in the 3kHz area—not as much as the Sonys. But I found adjusting to the Sonys’ response was easy. Though I was always conscious of the contour—that sort of consciousness is part of the reviewer’s job—I was not disturbed by it.
In any case, compared to the average nominally techno-correct speaker, the Sonys are on the subdued side. Everything tends to sound a little more beautiful than one expects, would be one way of putting it. Or a little bit more like concert sound would be another way, this being effectively the same thing. (Concert hall sound really is beautiful—in a good concert hall, anyway.)
The overall effect can be and often is really marvelous. The slow movement of the Rachmaninoff Cello Sonata with Starker and Neriki (RCA) sounded as beautiful as I have ever heard, except for when I heard the duo perform it live. Their live performance, alas, will never happen again—Starker is deceased. But if you were so unfortunate as to have missed them live, this recording played back on the Sonys will give you a sound very close to the magical beauty of the real experience.