Gradient Revolution Active Loudspeaker

I couldn’t have said it any better myself. Actually, that was me— back in Issue 100’s “Best Components of 1994” feature. The Revolution was a great speaker then, and it’s even better now. The revised “active” version brings with it a new mid/tweeter unit, which allows bass levels to be adjusted to match room acoustics, and it accepts the addition of extra bass units for larger rooms and higher playback levels.

People have long appreciated that dipole bass is simply better. With HP leading the critical charge, this was one of the founding observations of high-end audio, and it fit right in with acoustic theory. On the other hand, in higher frequencies the back wave of a dipole radiator is just a nuisance—you have to absorb or deflect it, or transients will bounce right off the back wall, show up at your listening position with a time delay, and make the sound edgy and just plain yucky. Clearly, dipole bass plus forward-radiating everything else is the way to go. And yet how many speakers are just the reverse—dipole mid and treble with box bass?

For more than a decade the Gradient Revolution has stood by (during the endless proliferation of box bottom/dipole top), reminding those who have listened to it how well a speaker can succeed in presenting sound without interference from the room—if the speaker is correctly designed in the first place. In this new “active” version (a passive model is also available), the crossover is done at line level, and the speaker is (necessarily) bi-amplified. The electronic crossover includes a level adjustment both to accommodate differences in amplifier gain (if two non-identical amps are used) and, perhaps more importantly, to match the bass level to different room sizes and acoustics. For larger rooms and higher dynamic levels, you can add on more bass units. You can even put two additional bass units, one atop the other, next to the main speaker for really big bass dynamics. This makes a somewhat imposing but graceful-looking system, which is depicted on Gradient’s Web site. In my modest-sized room, the single-bass-unit-per-channel approach sufficed, but I have heard the system with more bass units and can guarantee that integration is not a problem. (Incidentally, if the woofers are overdriven, they can “bottom” but will not be damaged.)

The top unit is far more than a mini monitor sitting on top of a bass unit. The coaxial mid/tweeter driver (from SEAS) is housed in a specially shaped enclosure, which has slots in the sides to provide—via “flow resistance” time delay—a cancellation of the back wave. Intuition tells us that a box speaker only radiates forward. But except in the very high frequencies, a great deal of sound bends around the speaker and radiates back. Hence, a “forward radiator” is often that in name only. Yet save for in a forward-facing cardioid pattern, the Revolution’s top unit radiates minimally. (I measured this at a normal height outdoors, so that the ground reflection was not eliminated and the sound behind the enclosure was down by around 15dB compared to the front level. In effect, there is no back wave from the mid/tweeter unit.)

This has several consequences. The speakers can be put close to the back wall without introducing the usual colorations, though they still work better out in the room. Moreover, the directional cardioid forward pattern minimizes sidewall reflections so the speaker doesn’t “know” that the room is there.

The forward pattern also has remarkable frequency consistency. As one moves off-axis, the uniformity of response in the mids and highs is superb; the only effect is a smooth, even roll-off in the highs. The general level drops in cardioid fashion, but there aren’t any types of glitches and discontinuities that almost always happen with ordinary enclosures. To say this is textbook behavior understates the case—it rewrites what is possible. To my knowledge, only the Quad 63 and its contemporary descendants have comparable radiation-pattern control.

So, what does this all add up to sonically?

What happens is quite amazing. First, the original acoustics of the recording venue are revealed to an all but unprecedented extent. (The “all but” refers to near-field listening, especially to directional speakers like the McIntosh arrays. But one has a very restricted listening position with the latter, not the wide window of the Revolutions.) Gradient designer Jorma Salmi described his intentions by saying, “Gradient loudspeakers do not bring the orchestra to your home. They take you directly to the concert hall.” Bold words, but they are effectively true, and to a startling extent. This is what high-end audio is supposed to be about, right? HP’s “the sound of real music in real space,” yes?

The Revolution is devoid of any hint of enclosure-induced coloration. It sounds not like a box or a panel or anything else except a perfect point source. Instruments show their own colors without any enclosure-induced overlay. I’m not saying one cannot get similar effects with the best box speakers, but to do so one needs to listen close up and/or do a lot of acoustic-treatment work to rid early room reflections. By contrast, the Revolution can be put down in a spot that’s good for the bass, and voila.

The speaker itself is not balanced “monitor flat.” A little more midrange relative to the treble gives it some extra body, making it somewhat forgiving. But the smoothness and controlled radiation pattern give instruments their real timbres, albeit slightly fuller and rounder. Although the Revolution needs no help at sounding realistic, it’s the ideal speaker to which to apply DSP room correction, à la TacT. Such DSP systems work best when there is little to change, and the agreement between the direct sound and the room sound means that there is no need to choose between which sound to smooth. The uniformity of the response over the listening window means that a perfectly corrected sound at one position will also be perfect at nearby listening positions.

But nothing is quite perfect, and the most obvious difficulty with concentric drivers is that the mid driver surrounding the tweeter may adversely influence the tweeter’s operation. While the top end of the Revolution is very smooth overall, the “fine structure” of the tweeter’s sound is less smooth in response than the best stand-alone tweeters. (Moving very slightly off-axis evens out the irregularities.) How noticeable this will be to you will be determined by your own listening. If you’re used to the ±0dB smoothness of, say, the SEAS Excel tweeter, then you may notice a few little ups and downs.

The only other real limitation that comes to mind is not a failing, but a consequence of virtue. The Revolutions image more correctly than virtually any other speaker. If there is legitimate outside- the-speaker information, it will be presented. On the Waterlily Mahler Fifth with Saint Petersburg and Temirkanov, the trumpet that sits to the far right sounds really far right, as it should. But the Revolution will not synthesize the proverbial wall-to-wall soundstage automatically, independent of what is on the recording. Has not a small voice suggested to you that if the soundstage in your room is wall-to-wall, that it must somehow be a fake? How would the recording “know” how big your room is? If recorded soundstage means anything, it has to mean something that does not depend on the playback room. Soundstage is real, but a synthetic wall-to-wall soundstage, arising on every recording from the reflected backwave, is suspect.

This brings to mind another remarkable aspect of the Revolution’s performance. A while back, while reading a Web site for a competitor’s widedispersion speaker, I came across a claim that the resulting reflections had the supposedly positive effect of localizing the speaker in a front-to-back plane. This statement is outrageous in its admiration for the effect, but correct in its analysis. The low level of reflected sound from the Revolution has the positive effect of not localizing the speakers in the front-to-back sense. Many speakers disappear sideways—as do the Revolutions—but few let the front of the soundstage be determined by the recording alone, without the speaker setting up the front of the stage in its own plane. Well-known theorist Gunther Theile has pointed out that in general, stereo does not really work, precisely because the plane of the speakers is perceived and locates the sound from front-to-back in a way not actually recorded. The Revolutions defeat this defect as well as any speaker I’m aware of, and far better than almost all others. (Suggestion: Ask yourself exactly where in the front-to-back plane a mono, no-reverberation, closemiked vocalist ought to be front-toback and then listen to where the position is with most speakers.)

I suppose it’s clear that I really like these speakers, but my admiration goes far beyond simply liking how they sound. To me this design represents a fundamental improvement over most others by directly and successfully treating the problems of room and speaker interaction. In my experience, the resulting immersion in the originally recorded soundfield is otherwise only available in “near-field” listening. Anyone with serious audio interests ought to listen carefully to the Gradient Revolution for the pleasure of it, and also to hear what is possible when a speaker can “ignore” the room around it. You can spend far more money, but you are unlikely to find a speaker (without any special signal processing) that does anywhere near as good a job of taking you to the concert hall. And that is where you want to go, isn’t it?