Gradient designer Jorma Salmi is one of the most inventive of all speaker designers in the modern era. His first speaker, the Gradient 1.1/1.3 (the two models were essentially the same design), appearing in the mid-1980s, combined a floor-firing woofer with a 12" dipole midrange and a line array of four tweeters. It looked unusual, but it sounded extraordinarily good and in particular it set new standards for differentiating against the room around it. Its sound was unaffected by room acoustics to an unprecedented degree.
Next came the Revolution in the 1990s. More conventional visually, it was equally unusual in its acoustic design, having dynamic-driver dipole bass combined with a concentric mid/tweeter unit in a flow-resistance enclosure that attenuated the back wave of the mids and highs almost entirely. Again, this was a speaker with a remarkable indifference to room conditions. Later, the Revolution was made available in an active version with adjustment of bass level possible along with possible addition of extra bass units. (You can find my TAS reviews of the Gradient 1.3 and the original Revolution speakers reprinted on regonaudio.com.)
Now in the first decade of the 2000s, Salmi’s latest thinking on room independence is embodied in a completely new model, the Helsinki 1.5, which is even more strikingly different from other speakers than were its predecessors. The visual design is unusual and so is the sonic nature of the speaker. It pushes certain aspects of speaker performance to what I think are literally unprecedented levels. If these aspects lie in the directions of your own major sonic goals, it may be that no other speaker will do, once you have heard these.
The specific area in which the Helsinki 1.5s truly excel is the reproduction of spatial information. This is a long-standing goal of high-end audio, and indeed one might go so far as to say that the desire to push this as far as it could be pushed was the main thing that historically separated the high end in its beginnings from ordinary “high fidelity,” largely via the emphasis by HP and his writers, almost from the beginning of TAS, on “soundstage.”
Attention has been given to the spatial aspect of speaker design ever since, e.g., the minimization of diffraction. But the Helsinki 1.5s explore space in a way that goes far beyond such standard considerations. The Helsinki 1.5s not only disappear as sources themselves, they make in effect your listening room disappear from the sonic picture, or at least remarkably close to that.
Truth to tell, the expansive soundstage of quite a few speakers does not partake of this latter kind of disappearance at all. Rather, the sense of large space is obtained by reflections off the boundary surfaces of the listening room, and, however large-scaled the results, the space produced is not really the space recorded. The Helsinki’s take another path entirely: They try to minimize, indeed almost to eliminate, the early reflections in the listening room so that the listener is, during the crucial early-arrival period, hearing only the sound recorded, not the sound of the listening room at all.
This might seem a bit theoretical. But the sonic impression is anything but theoretical—it is overwhelming and immediate. Music recorded in a large space sounds enormous. Music recorded close up in smaller spaces sounds to that scale. And everything in between is similarly reproduced as it is in spatial terms.
It is worth having a look at how this is done. The Helsinki is a three-way speaker, but without a box. The bass is produced by a dynamic driver operated in an open baffle, but the baffle is oriented along the nominal axis of the speaker so the dipole is firing crossways, as it were. The midrange driver is a forward-firing dynamic driver mounted in a 2-½"-thick disc that is rounded at the edges and slightly scooped, the part of the disc not occupied by the driver being filled with absorbing material. On account of the absorption, the whole midrange unit has a cardioid forward pattern and effectively no back wave radiation at all. And the disc is tilted upwards at an angle to point at the listener but away from the floor so that the floor reflection is minimal. Finally, the tweeter is mounted in another scooped-out disc that functions as a waveguide so that the dispersion angle of forward radiation of the highs is controlled. A picture will make all this clearer than words can!
The overall effect is that it is a long time before any reflections arrive at the listening position. Sidewall reflections are minimal on account of the directed forward radiation. Floor reflection is minimal because of the upward tilt of the midrange.
The wave-guided high frequencies are too directional to bounce off the floor to any substantial extent. And the midrange is sent to the ceiling at an angle that makes the reflection arrive behind the listener at an ordinary listening distance, while again the high-frequency driver is too directional to reflect off the ceiling to any extent. Some aspects of this were anticipated in the Gradient 1.3—hence the 1.5 model designation, one supposes.
The listening effect is spectacular. But I decided that in addition to just listening, I would try to verify what was going on in terms of measurements. Stereo depends on symmetry: The channels need really to match, and, while pair-matching of speakers is routinely tested, pair matching of channels in the listening room is another story. Rooms are almost never completely symmetric, so the match between the speakers themselves can be obfuscated by room effects. The idea is that for perfect stereo, a mono signal fed to one speaker and the same signal with the reverse polarity fed to the other should produce at the listening position an absolute null—no signal at all—but this does not happen in general, and often not even to good approximation. But with the Helsinki 1.5s, even in a situation where the room setup involved an alcove on one side not matched on the other and so on, remarkable in-room matching occurred. I checked the sum of the two channels, both channels “in phase,” versus the difference, one channel out of phase relative to the other, with the first twenty milliseconds of an impulse fed to the speakers. (Twenty milliseconds is essentially the “integration time” of the ear as far as stereo imaging is concerned). The difference, from 300Hz on up, was 15dB down from the sum and at some frequencies even more than that. With typical wide-radiation speakers, the highs do well on this test, if one undertakes the easy task of absorbing the early high-frequency reflections. But the in-room midrange “channel matching” in the sense indicated is lucky to be on the order of 5dB. With the Helsinkis, the strong matching went down in frequency much further.
So far, all news has been good. But of course all speakers have limitations. The most obvious limitation of the Helsinkis is that in the bass they are a dipole in a fairly small baffle. So low bass is not available, on account of dipole cancellation. There is not much usable output below 50Hz, and a subwoofer is needed to make music that involves deep bass convincing. Indeed, I would recommend a sub for almost all music.
The Helsinkis also require some special consideration in room placement. The sonic results can be adjusted by placement, especially in tonal balance, to a surprising extent. Because the woofer is dipole and edge-on to an on-axis listener, the on-axis bass is effectively missing. For an on-axis listener, there is bass only from the room sound, and the bass level is likely to be too low. On the other hand, the speaker is very smooth, uncolored, and quite flat over a fairly wide angle off-axis, so that the speaker can be rotated away from being aimed directly at the listener with minimal change to the higher frequencies; indeed the speakers sound more neutral from the mids on up 15 to 20 degrees off-axis than on-axis. (The final production model has approximately 1dB less top end than the version I had, something I compensated for in my auditioning using the Z Systems digital equalizer, but which was not compensated for in Paul Seydor’s audition. But even with the production curve with its slightly cut highs, I found that the top end was more natural fairly far off-axis. Such a rotation brings up the direct-arrival bass, the axis of the cross-firing bass driver being more nearly pointed at the listener.)
Putting the speaker close to a wall can bring up bass further. Usually such placement would cancel dipole bass but since the dipole is crossways, the woofer is at right angles to the wall and the wall reinforces bass. And of course one can tilt the speaker to an angle to the wall—the possibilities approach the proverbially limitless. And the changes can be large. The downside of wall placement is that the magical stereo is somewhat diminished—not a lot, given that the back wave is so deeply attenuated in the mid and higher frequencies, but somewhat.
I got the best bass with positioning the speakers along a wall, tilted toward the wall at a considerable angle, about 30 degrees from perpendicular. But the most gratifying results overall were with the speakers out from the wall about four feet, tilted inwards about 20 degrees beyond what would have been aiming them directly at the listening position—a lot of toe-in! In this position, the perceived balance was very natural and good orchestral recordings could be amazingly convincing. The wonderful recording for Delos by the late John Eargle of the orchestral music of the final scene of Wagner’s Ring was truly transporting to another realm, to the point that I was almost speechless at its conclusion, finding it hard to return to Earth, as it were. To me, this is the kind of experience that audio is for.
When one tilts the speakers in free space, the response in effect pivots on approximately 400Hz. Higher frequencies roll down; lower frequencies come up, as far as direct arrival is concerned. Thus, in effect the composite response is oriented somewhat towards the 300–500Hz region, the pivot point. This is far from unmusical—the speaker always sound attractive. Our own Paul Seydor was kind enough to let me try them in his room, and the sound there was positively seductive, with the speakers out into the room but tilted at a little more than 45 degrees.
And as expected, the soundstage spatial behavior was magical. But the subtle orientation towards the 300–500 Hz range was there in his room, as in my own. The Helsinkis are typically very smooth in in-room response, unusually so, but they are not necessarily exactly flat, and they do have a tendency to be light in the bass and to emphasize the mids. Not necessarily a bad thing! But still something to check out.
The Helsinkis also offer an extraordinary kind of transparency. The highs are crystal clear but in addition, the suppression of early reflections extends the clarity down into a special kind of unraveling of midrange textures. Complex music is extraordinarily revealed. I once described the Gradient 1.3s as the ideal speaker to use if one needed to transcribe an orchestral score from listening alone. The Helsinkis continue the tradition. On top of this, the Helsinkis have an unusual kind of reality of a less easily described sort. Natural sounds recorded for films, for example, have a tendency to sound much less “canned” than usual. One tends on the subconscious level to take the sound as actually real far often more than usual. (My Dobermans noticed this, too, for what that is worth; they were fooled more often than usual into thinking things were real and barking at them and so on.)
The Helsinkis seem to me to be one of those rare speakers that really needs to be heard by anyone seriously interested in audio. They are something truly out of the ordinary. The vast majority of speaker designs are simply attempted improvements of old ideas. Typical floor-standing boxes might impress Rice and Kellogg (inventors of the modern dynamic direct-radiating speaker in the 1920s) with the extent to which their ideas had been perfected. But they would not be surprised by them.
The Helsinkis on the other hand are a surprise: They try to do something quite different from other speakers. You will likely need a subwoofer. You will need to experiment a good bit to get the ideal tonal balance, and you may never get quite the sound truly weighty in the lower mids/upper bass. But this speaker has things to offer that no other does, especially in putting you in the space of the recording venue. There is literally nothing else like it, and it is an indispensable audition if you care about the proverbial “state of the art.”