Revel Ultima Salon2 loudspeaker Jim Austin May 2019

Jim Austin wrote about the Revel Ultima Salon2 in May 2019 (Vol.42 No.5):

Loudspeakers are the most diverse of hi-fi components, varying widely in conception, gestation, and morphology. This variety reflects the peculiar nature of our hobby, which spans the widest possible range of tastes, and of disciplines from science to art. Some loudspeakers are conceived to meet specific technical criteria, while others begin with a designer's loose idea and then evolve, via the trial and error of listening tests, to realize the designer's desired sound.

Far along toward the scientific side is Revel, a division of Harman International, which is now owned by Samsung. For years, Harman has done its own rigorous research using blind tests with listening panels to map technical parameters—frequency response, dispersion patterns, and so on—to the tastes of listeners. They've found that listeners have similar tastes in sound—in particular, that casual and highly trained listeners tend to most prefer an extended and even frequency response and uniform horizontal dispersion (footnote 1).

At the top of Revel's current line of models stands the Salon2, part of their Ultima2 series, a tall (53.25"), ported, six-driver, four-way loudspeaker weighing 146 lb and costing $21,998/pair. The Salon2 isn't new: It shared Stereophile's Loudspeaker of the Year award in 2008, shortly after its release. In his review in the June 2008 issue, Larry Greenhill concluded that, in the Salon2, Revel had "created a new reference standard in floorstanding loudspeakers."

John Atkinson, too, was impressed. In measuring the Salon2 for LG's review, he found it a reasonable amplifier load (4 ohms) with a dead-quiet, nonresonant cabinet and a remarkably flat frequency response, both anechoic and in-room. Its off-axis response, which is critical for optimal imaging, was textbook. The speaker "offers superb measured performance," he concluded. In a Follow-Up in March 2009, JA wrote that "Revel's design team has taken the conventional concept of a moving-coil box loudspeaker to the limit of what is currently possible." High praise.

It's now been more than 10 years since that Follow-Up, and not only is the Salon2 still in production, Revel still sells it for the same price: $21,998/pair. Factoring in inflation, that makes it about 14% cheaper than it was 10 years ago. Factoring in the raves—measurements and reviews superior to those of many loudspeakers costing much more—that makes it a relative bargain.

Aiming to experience for myself "the limit of what's currently possible" in a moving-coil box loudspeaker, I requested review samples.

Setting up the Salon2s in my listening room wasn't especially easy, if only because of their size and weight (146 lb each, with a shipping weight of 178 lb). Sonically, it was a breeze. Moving the Revels somewhat changed the character of their sound, and optimizing their positions improved it. But apart from altering the interaction of the Salon2s' woofers with standing-wave room modes in the bass frequencies—always an issue with full-range loudspeakers—the changes in sound were modest compared to most other loudspeakers I've set up.

I ended up with the Salon2s about 10' apart, and my listening chair 10–12' (see below) from each speaker. Both speakers were more than 5' from the front wall, but because the right speaker was quite close (about 18") to a record shelf, I randomly riffled the LPs to diffuse reflected sound. I know that isn't optimal—I'd have preferred to shift the whole setup farther from that sidewall—but I like the idea of testing loudspeakers in real-world conditions. I also briefly tried the Salon2s with the right speaker given more breathing room, and while the compromises of the near-wall setup were easier to hear, I found them musically unimportant, except for a slight excess of room gain.

I did hear some differences in imaging between the right and left speakers, but these were surprisingly modest and inconsistent: with some recordings, the open space to my left resulted in a wider soundstage; with others, reflections from the record shelf to my right seemed to do the same. Go figure (footnote 2).

One of the first things I noticed after I got the Salon2s set up was how good they sounded at low volumes: round and rich and full, with a smaller soundstage than at higher volumes but still with satisfying imaging. This was likely due, in large part, to their deep, generous bass: the ear is less sensitive to bass at low volumes, so more of it is needed to make an impression. Another likely reason: Even at very low volumes, the distortion remained low. The Salons sounded especially wonderful at low volumes with Pass Laboratories' XA60.8 class-A monoblocks (120W into 4 ohms). Played louder, and with more demanding music—especially orchestral—the Revels seemed to slightly prefer the far more powerful PS Audio BHK Signature 300 monoblocks (600W into 4 ohms).

At these more natural volumes, the Salon2s forcefully and explicitly reproduced every note my music contains, from top to bottom. The only notes they wouldn't reproduce were the very lowest pedal tones of the biggest pipe organs. (A 64'-long organ pipe vibrates air at a frequency of 8Hz—which surely is more felt than heard.) The Salon2s had no problem reproducing the lowest note of a five-string double bass, or C0 (16.351Hz) in the bottom octave or the fundamentals of the very lowest notes of an Imperial Concert Grand made by Bösendorfer—who, on the suggestion of composer Ferruccio Busoni, added extra keys to extend the piano's range all the way down to the bottom of a pipe organ's 32' register.

According to LG's review of the Salon2, in revising the original Salon, Revel put much effort into enlarging the sweet spot by making the speaker's off-axis and on-axis responses match as closely as possible. They succeeded. When I have company over to listen, I usually give them the good chair, which is precisely centered between the speakers; I sit off to the right, my head a couple of feet to the left of the right speaker. From that position, the imaging I heard with the Salon2s was better than with any other pair of speakers I've had in my room. Sitting shotgun, I heard a well-formed soundstage centered on a point just off-center but still well inside the right speaker, and extending all the way over to just inside and behind the left speaker.

This imaging excellence wasn't limited to off-center seating. The Salon2s imaged better than any other speakers I've heard, except in terms of depth—and even there, they were competitive. Wilson Audio's Alexandria XLFs ($210,000/pair), set up in a showroom at Manhattan's Innovative Audio, had the best soundstage depth of any stereo pair of speakers I've heard, convincingly reproducing the London Symphony Orchestra (apparently seated on risers) at perhaps two-thirds actual size as they played Shostakovich's Symphony 5 under Mstislav Rostropovich (24-bit/96kHz download, LSO Live 550).

The Salon2 is tall, with the tweeter at the top, and my listening chair is low. That put the tweeter axes well above the usual target: my ears. The typical solution is to lengthen the speakers' rear spikes and shorten the front ones, to aim the tweeters slightly down, toward the ears. I didn't do that because, when Revel packed up my review pair, they omitted the spikes. That left me sitting in a position where, according to JA's vertical-dispersion measurements, the speaker's frequency response is still quite flat but the top octave is rolled off more—perhaps 2–3dB lower at 20kHz—than it would be on axis.

The more polite high frequencies resulting from listening below the tweeter axis could be compensated for by turning the Salon2s' tweeter-level controls up a notch, and that worked very well. Such adjustable crossovers are a great idea, and I wish more speakers had them—but I wish the steps on these controls were finer. (There's also a bass-level control, intended to be used when the speakers are placed in a corner or near the front wall.)

The Salon2s look like bass monsters: big cabinets with three 8" woofers per side, each with an inverted aluminum cone. And yet, I thought, well-recorded orchestral music sounded just about right in my room: the balance of the various instrumental sections sounded almost natural. But as JA observed during a short visit, Peter Maag's 1958 recording, with the London Symphony, of the Scherzo of Mendelssohn's A Midsummer Night's Dream (FLAC ripped from 24K gold CD, Classic Compact Discs CSCD 6001), had a little bit too much bass.

Once he'd pointed it out, I had to agree: At realistic listening levels, double basses and low-end percussion, especially the orchestral bass drum, were a little too loud in the mix, though not enough to be distracting. The big Salon2s, then, had a bit too much bass for my room—too much room gain?—but only a bit.

A counterexample: In 2016 I chose, as my Records to Die For selections, two recordings by obscure Icelandic musicians. One was Enter 4, by Hjaltalín (CD, Sena SCD588), which I've always found to have abundant, satisfying bass even through bass-lite speakers. The amount of bass the Salon2s produced was absurd, preposterous, over the top, unhinged. And not in a good way. Presumably, Sveinn Helgi Halldórsson, Enter 4's engineer, never heard his work through full-range speakers. He engineered it wrong.

This was not the Revels' fault—let's be clear about that—but it did demonstrate a truth about full-range and other ambitious loudspeakers: Their very virtues make them usable with a narrower slice of recorded music. In all my years with the excellent but less ambitious DeVore Fidelity Nines, I never had a music-reproduction failure this catastrophic.

So now I must qualify my 2016 R2D4 recommendation: Hjaltalín's Enter 4 is a wonderful album and sounds really good—except through full-range speakers, when it sounds like crap.

Full-range speakers make other demands, including on your room. That bit of extra bass that JA detected with the Mendelssohn is one example. Here's another: With so much happening over such a wide range of frequencies, complex music can seem overwhelming, disorganized, all over the place. When I sat 12' away, the sound was more coherent than when I sat 10' away. But sitting farther from the speakers requires a bigger room.

Occasionally, with close-miked recordings of aggressively picked guitars, the Salon2s could sound too scratchy. An example was "Whatever It Was," from Greg Brown's Slant 6 Mind (16/44.1 FLAC, Red House RHR CD 98/Qobuz). That sound may well be on the recording; as I've written before, it's hard to know for sure. Still, why tolerate harsh sound? I'm against suffering through sound just because it happens to be recorded that way.

With the Salon2s, though, that harshness could be tamed by turning down the tweeter control. Of course, there were trade-offs: With the tweeter level turned down, very good piano recordings got mellower, losing a bit of the percussive edge that's necessary to reproduce recorded piano sound in a natural way. Life isn't perfect. (The best way to manage this, with digital tracks, is with server-based DSP. But I'm still waiting for Roon or some other server software to make their DSP settings track- or album-specific.) Meanwhile, some piano recordings could occasionally, through the Salon2s, take on a metallic quality in the middle of the instrument's treble—roughly, the octave above C6.

It's been said by some members of the audio fraternity, including our own Herb Reichert, that audio components can sound like the type of material they're made of. The Salon2 uses metal drivers—was that what I was hearing?

I don't think so—these are no ordinary metal drivers. When I gently tapped each inverted cone (though not the Salon2's beryllium tweeter), the only ones that made a sound resembling what I heard with those piano notes were the aluminum woofers—which, unless my review samples were broken (they weren't), produce no sound above about 150Hz, nearly an octave below middle C and three octaves below the notes in question.

What was happening, then, with those metallic-sounding piano notes above 1kHz? I don't know. My best guess is that that's what these recordings actually sound like, and that the Revels were doing a better job than previous speakers of revealing what's on them. And JA's measurements reveal nothing untoward in that range in the Salon2's frequency response.

Revel's Ultima2 Salon2 is a wonderful loudspeaker—by no small measure, the best I've had in my home. Science works. They're explicit, literal, full-sounding, wide-ranging, even in frequency response, and—thanks to their tweeter and woofer controls—versatile. They image beautifully. They play as loud as you'd want speakers to play in any domestic setting in which you care about sound quality. (If you just want to party, get some JBLs.)

On the other hand, like other ambitious, full-range loudspeakers, the Ultima2 Salon2 makes demands—on the source material, on the room, on other equipment (because they're so revealing), and sometimes—when you're forced to hear what is actually on those recordings—on you, the listener. Are you ready for that kind of commitment?—Jim Austin

Footnote 1: See "Differences in Performance and Preference of Trained versus Untrained Listeners in Loudspeaker Tests: A Case Study," Sean E. Olive, J. Audio Eng. Soc., Vol.51, No.9, 2003 September.

Footnote 2: I used Doug Jones's Listening Environment Diagnostic Recording Test (LEDR), which is designed to assess a system's reproduction of recorded space. The "Lateral" image extended about 3' beyond the left speaker (the open side), but stopped at the right speaker (the one near the wall). This test is available at See also Bob Katz, "Take Me to Your LEDR!" in the December 1989 issue.