So what do the folks at Revel do to improve upon the already impressive F228Be loudspeaker that I reviewed two years ago? Make a bigger version with a third woofer, a newly refined tweeter/waveguide combination, and add two rear-firing ports. The piano black review sample truly looks like the Darth Vader of loudspeakers. Cue the “Imperial March” please!
Revel PerformaBe F328Be Floor Standing Loudspeaker and C426Be Center Channel Speaker
- The F328Be’s sound is incrementally more refined in the middle and high frequencies.
- While tonally related, the F328Be plays at a noticeably bigger scale in-room over the F228Be.
- Potent bass reach. Even better when not so far in the room as in my studio.
- The C426Be is a seamless mate to both the F328Be and F228Be.
- Heavy and solidly built.
- Measured impedance is closer to 4-Ohms than the claimed 8-Ohms.
I am a firm believer in the speaker design and listener preference principles that have evolved at HARMAN over the past several years. Ever since Floyd Toole brought his years of knowledge and research experience at the Canadian NRC from the Great White North to SoCal, HARMAN has used that knowledge and continued research to successfully inform many of the speaker (and some headphone) designs across its many brands. Like-minded designers and engineers like Kevin Voecks and Mark Glazer (Lead Engineer on Performa Be) have applied and built upon that knowledge at Revel, in particular, to create several very popular and appealing speakers for audiophiles the world over. And while I won’t say that my personal appreciation of speakers is limited to those that only follow the “HARMAN Curve”, I’ve listened to enough of them in my studio to know that I tend to like what I hear from speakers that have been tuned to follow that response model.
I had been introduced to the new Revel F328Be and C426Be at CEDIA 2019. You can see HARMAN’s Jim Garrett walking me through the Revel PerformaBe line in this video starting at the 14-minute mark. Having just recently completed my review of its now smaller brother, I was intrigued by the new refinements and the increased displacement afforded by this new speaker. Would it actually make a noticeable difference to me when used under the same conditions as the F228Be? As luck would have it, I still had the F228Be in house so I would be able to draw some direct comparisons. Revel graciously asked if I wanted to sample the new C426Be as well and experiment with a Revel based home theater while I was at it. Well, why not? So, before I knew it, some very large crates arrived at my home, much to my wife’s chagrin, containing some very precious cargo. A pair of piano black F328Be floor-standing loudspeakers, a C426Be center channel, and a pair of the Concerta M8 On-Wall speakers to use for surround duty. Let the games begin!
Three-Way Bass-Reflex Floor Standing Loudspeaker with dual rear-firing ports
Single 1″ (25mm) Beryllium dome, with acoustic lens waveguide
Single 5-1/4″ (130mm) Deep Ceramic Composite aluminum cone, with a cast frame
Three 8″ (200mm) Deep Ceramic Composite aluminum cones, with cast frames
Recommended Amplifier Power:
50 – 300 watts
26 Hz – 40 kHz (-6dB) (Manufacturer)
240 Hz, 2.1 kHz High Order
Nominal impedance (Manufacturer):
Black, White, Metallic Silver, and Walnut
Dimensions (H x W x D):
50.9″ x 13.5 ” x 17.6″
112.6 lbs. (51kg) (each)
5 Years Parts and Labor
Foam Port Plugs, Carpet Spikes
Three-Way Bass-Reflex Center Channel Loudspeaker with dual rear-firing ports
Single 1″ (25mm) Beryllium dome, with acoustic lens waveguide
Single 5-1/4″ (130mm) Deep Ceramic Composite aluminum cone, with a cast frame
Four 6.5″ (165mm) Deep Ceramic Composite aluminum cones, with cast frames
Recommended Amplifier Power:
50 – 300 watts
38 Hz-40 kHz (-6 dB) (Manufacturer)
210 Hz, 2.1 kHz High Order
Nominal impedance (Manufacturer):
Black, White, Metallic Silver, and Walnut
Dimensions (H x W x D):
10.4″ x 38.6 ” x 14″
61lb (27.7kg) (each)
5 Years Parts and Labor
revel, beryllium, Performa, f328be, c426be, acoustic lens, waveguide, loudspeaker, speaker review, review 2021
Once my son and I gingerly freed each of the F328Be loudspeakers from their ginormous rolling crates and carried them down the stairs to my studio, I was able to stand one of them next to an F228Be that I have here and make some direct physical comparisons. The family resemblance is, of course, unmistakable. While the width remains unchanged, Revel has added 7-plus inches to the height and just over 3 inches more depth over the F228Be. While this doesn’t sound like a lot on paper, in the flesh the F328Be cuts a much more imposing figure. That extra height and internal volume are there to accommodate a third 8-inch-Deep Ceramic Composite (DCC) aluminum cone woofer. On the other side of the coin, in order to limit the height of the tweeter from getting too far above ear level, the F328Be has 2 smaller diameter flared ports relocated to the back of the speaker versus the single, large front-firing pipe of the F228Be. This should theoretically help the F328Be to dig deeper and hit harder than its smaller brother.
Right above the 3 woofers is the same 5.25-inch DCC midrange driver as found in the F228Be. In general, this Be series midrange differs from the comparable drive unit in the Performa3 range by having a larger voice coil and magnet structure. Also, the width of the surround is reduced providing for a larger cone diameter. The Deep Ceramic Composite treatment that is applied to the woofer and midrange drivers is, essentially, a specially formulated ceramic coating (called Alumina) deposited on both sides of the aluminum driver cone creating a sandwich. Revel (and HARMAN) have used and refined various versions of this cone material for over a decade under a variety of names. This coating helps dampen the aluminum and push the driver’s breakup point to outside of its defined operating range within the speaker.
HARMAN has reported that Alumina has the highest first bending mode frequency of any common cone material, but it is brittle, hence the need for the aluminum center. In the Revel Be speakers, HARMAN uses a new deposition system to make a thicker Alumina layer over the aluminum cone base. The results more closely approach a cone made just of pure Alumina. This is said to result in a smoother and more uniform response with lower distortion all around.
The F328Be’s tweeter, on the other hand, is said to be completely new. The Beryllium dome is now motivated by a much larger magnet structure for improved power handling while being mated to a new 6th generation Acoustic Lens waveguide. Not that the new waveguide looks noticeably different from the previous generation when eyeballed, but the folks at Revel say that it has gone through extensive computer modeling and refinement which our measurements show improves the horizontal radiation performance above 5 kHz.
Crossover points are at 240 Hz and 2100 Hz using Revel’s typical high order acoustic slopes. While nominal impedance is claimed to be 8 ohms, our bench test results show a nominal impedance closer to 4 ohms.
Turning to the C426Be, at 61 pounds it is a substantially sized center channel speaker with 4 short integrated legs that can be tweaked to angle the speaker according to your needs. Both the tweeter and midrange drivers are taken from the F228Be but oriented closer together to fit within the height of the baffle. Two 6.5-inch DCC aluminum woofers flank either side of the tweeter and mid, for a total of four. Two tuned ports exit out the rear of the speaker. The crossover design is in the same vein as the F328Be except the midrange crosses to the woofers at 210 Hz. Again, Revel claims an easy to drive 8-ohm impedance, but our bench tests found it to be noticeably lower. Both this speaker and the towers have Revel’s familiar four gold-plated binding posts that are suitable for bi-wiring/amping. Both speaker enclosures also feel respectably solid when exposed to random knuckle-wrapping indicating good bracing throughout.
You could be forgiven for thinking that the Concerta M8 On-Wall speakers that Revel sent along for surround use might seem a little undersized for the job, compared to the front three speakers. Revel however says that the M8’s 1-inch ceramic composite tweeter and dual 3-inch ceramic composite woofers are designed to favorably blend with many of their larger, more sophisticated kin. And, factoring in the moderate size of my home theater, Revel was confident that they would fit the bill as on-wall surrounds. The M8 have solid aluminum enclosures and the included wall mounts allow the speakers to be precisely angled after mounting. High order crossovers round out the feature set.
The Revel PerfomaBe F328Be Loudspeakers were set up for stereo listening in my studio. The speakers were spaced approximately 9-feet apart with my seating position located 10-feet away from the center point between the speakers. The speakers were toed-in approximately 15 degrees. An OPPO BDP-105D Universal Player served as a main audio source along with an ELAC Discovery Music server (sending its SPDIF output to a Benchmark Media DAC3 B) using ROON to manage my music files. A Benchmark Media HPA4 preamp and AHB2 power amplifier along with an Anthem STR preamp and power amp were used for control and drive. For a vinyl source, I used my Technics SL1200 Mk 6 modified by KAB Electroacoustic, along with an Audio Technica OC9ML/II Moving Coil phono cartridge. This was hooked up to the Pass Labs XP-25 Phono preamp. Both the Benchmark Media HPA4 and Anthem STR preamplifiers were used to manage the inputs to the speakers at various points. The interconnects and speaker cables used were from Blue Jeans Cables.
I elected to use the Revel PerformaBe C436Be center channel and Concerta M8 On-Wall surrounds in conjunction with the Revel F228Be that I have for my home theater listening. The F328Be were just too large for my HT space and the F228Be’s tweeter height was able to get closer to the center channel’s tweeter height so they made for the ideal stand-in. Connected equipment included an Anthem MRX1120 AVR, an OPPO BDP-103 Universal Player, a Microsoft XBOX One, a Sony PS3, a Pioneer 50” KURO Plasma TV, a Pioneer CLD-D704 Laserdisc player, my dual sealed 15” Dayton Reference HF subwoofers, and an APC H15 Power Conditioner. The interconnects and speaker cables used were from Blue Jeans Cables. The speakers were used with Anthem’s ARC Genesis Room correction system. Maximum EQ was limited to 300 Hz on the mains, 600 Hz for the center, and 1.5 kHz for the surrounds. The mains and center were crossed over to the subwoofers at 80 Hz while the surrounds were crossed over at 170 Hz.
So, let’s get the first question out of the way right now, that being does the F328Be sound noticeably better than the F228Be? In reality, “Better” is not the operative word, but “bigger” is. When set up the same way, in the same room, the Revel F328Be plays at a scale that is noticeably larger than what the F228Be are capable of. In a room as large as my art studio (33’ x 23.5’) that makes a tangible difference. It just makes everything played through them sound that much more effortless. Both speakers share the same tonal character which I thoroughly enjoy, yet the F328Be to my ears seemed to have a slightly more refined handling of the upper midrange to treble region in most of the music that I would listen to. Not miles worth of difference mind you, but enough to get my attention when compared back-to-back.
The bass response and quality of these two speakers were quite close. In the bench tests, you’ll see that the F228Be actually reach a little lower than the F328Be did in my room, but in actual listening that wasn’t immediately obvious. What I did notice regarding the bass quality of the F328Be was that it was cleaner and more detailed when playing the same tracks back at roughly the same volume. Keep in mind that my listening setup necessitates that I place speakers well into the room, and away from the front wall giving the front ported speaker a slight bass advantage. Placing the rear-ported F328Be closer to a wall would most likely give it a significant room gain advantage over the front-ported F228Be. So, in a nutshell, if you have the room and you set the speakers up fairly traditionally, the F328Be should have an easier time bending the space to its will.
All the other sonic qualities that I noted in my F228Be review apply here as well. There is a sense of rightness to the sound of this speaker that few others can approach. Its reproduction abilities are complete and seamless from top to bottom. Vocals were always precisely imaged and natural in tone. High frequencies were smooth and called attention to themselves only when it was demanded by the material. Bass response was solid with plenty of drive and impact such that a subwoofer need not be considered. And while I’ve mentioned how good the qualities of the individual operating bands of this speaker are, it’s how well all the drivers are integrated that is really the icing on the cake. I could discern no gaps in where each of the drivers handed off to one another. It is so seamless that the F328Be’s sound approaches the effortlessness of a panel speaker yet with a visceral gut-punch when needed that most of those types of speakers sorely lack unassisted. There are other loudspeakers that are tuned to get your attention and they will sound fun and agreeable for a time, but the Revel F328Be sounds correct, balanced, and in the end completely captivating with the widest array of music that I could throw at it. Off-axis casual listening was also very good, better in some respects than the F228Be. There were more of those moments where I’d be drawn away from whatever else I was doing in my studio to sit in the sweet spot and listen again than with some other speakers. Unless Revel ever decided to take this speaker active, with onboard DSP, which would be a potential game-changer in my mind, this is as good an execution of a passive loudspeaker as one is likely to encounter. The engineers at Revel have refined this speaker to such a point that it actually replaced a speaker that was its closest competition, ironically higher up in Revel’s own pecking order. More on that a little later.
Regardless of what Revel claims, I would also be remiss if I didn’t advise you that the F328Be should be used with amplifiers that can solidly feed a 4-Ohm load. Both my Benchmark AHB2 and Anthem STR power amplifiers had no issues driving the big Revels with plenty of clean power to sound their best. One expects that if you are ponying up the cash to buy these speakers to begin with, you will have the commensurate amplification in place. That means that underpowered receivers and AVRs need not apply.
Turning to the Revel PerformaBe C426Be center channel speaker and its implementation in my home theater space. Once all the speakers were dialed in with Anthem’s ARC Genesis room correction (for the front 3 channels the room correction was limited to 600 Hz and below) my now Revel-based home theater became a rather exceptional place to enjoy a movie and listen to music. The C426Be was a seamless sonic match to the F228Be Left and Right channels. It also has to be one of the best sounding and consistent center channel speakers that I’ve ever used as there was practically no variation in its sound as I switched between my four main HT seats. Its sound had a very similar coherence to the ELAC Uni-Fi 2.0 center channel’s coax that I had here recently, but with an accuracy and impact to its performance that was turned up by several notches. Suffice it to say that anything that came out of the C426Be sounded as balanced and full as what was coming out of the F228Be. The little Concerta M8 on-wall surrounds did an admirable job keeping up with the 3 big dogs up front for most of my movie and music enjoyment and only came up a little short when I was pushing volumes a bit. But that center channel spoiled me such that I was wishing that Revel made an on-wall equivalent of it. A PerformaBe version of the Ultima Gem2 flex speaker would be ideal. (Hello? Kevin Voecks, Jim Garrett…you hearing this? I charge you nothing for this free product advice!)
Moving now to some choice music and movie cuts that the Revels found most agreeable:
Larkin Poe, Peach, Tricki-Woo Records, 16/44.1 FLAC via Qobuz, 2017
A southern blues/rock band fronted by sisters Rebecca Lovell (vocals) and Megan Lovell (guitars) produce some gritty and power-chord heavy tracks for this their third official album. The F328Be do all the things that you would expect a well-designed speaker to do, like render all the nuances of Rebecca Lovell’s full and rebellious vocals dead center in space on “Come On In My Kitchen” while nicely pegging the distorted slide guitar that Megan is skillfully playing off to the side.
It’s when you spin up the duo’s version of Ram Jam’s “Black Betty” that the Revels really light up the room with their tight yet relentless low-end drive, as evidenced by the punch-in-the-gut kick drum that keeps time through the track. Even here though, that drive doesn’t come at the expense of the little details like the metallic sheen of the tambourine cymbals or, the sounds of the maracas out past the speaker boundaries, or even the overdriven guitar amplifier buzzing throughout the song. Cap it off by how the F328Be play that hollow distorted wail of the slide guitar like some mad banshee klaxon and you get an inkling of what these speakers are capable of. Detail, imaging, and boogie in spades!
Ike Quebec, Soul Samba, Blue Note Records, 16/44.1 FLAC via Qobuz, 1962
Turning down the raucousness a bit to something more subtle and tasteful, the Revel F328Be do a superb job at imaging Ike Quebec’s bluesy tenor saxophone on the track “Loie”. It comes off sounding huge and spacious on the left side while at the same time relaying all the slightest breath and reed sounds while Quebec plays. The deep acoustic bass lines have a nice and satisfying punch to their tone and the beautiful clean ringing of Kenny Burrell’s guitar playing fills up the right side of the room.
“Lioro Tu Despedida” continues the bluesy-Bossa-Nova vibe with both Quebec and Burrell imaged, left and right again respectively, with such clarity and dimension that it came across as almost palpable. Again, great little details abounded in both their solos, but there was an added sense of depth that the Revel’s relayed on this track that made it extra rewarding to listen to. Crisp sounding little background cymbals and some nice stick work on the edge of the drums round out the “acoustic treats” that the F328Be made sure not to gloss over.
Leopold Stokowski, Rhapsodies-Liszt-Enescu-Smetana, RCA Red Seal, 24/88.2 via Qobuz, 2010
And speaking of the judicious balancing of detail with sheer power, Franz Liszt’s “Hungarian Rhapsody No.2” as conducted by Leopold Stokowski allow the F328Be to demonstrate their utter mastery of the air molecules in a big room. The dynamic swings and contrasts in this music are huge and they can shift on a dime. This track most clearly illustrated the scale difference between the Revel F328Be and the F228Be. The F228Be did handle this music well but when you combined the massed strings, with the woodwinds, horns, drum hits, and ancillary percussion and things got loud, the F328Be had decidedly better command of the imaging in the space.
The F228Be were a little more constrained sounding at volume. And I don’t know that it’s something that I would have otherwise noticed without having both here at the same time to compare. It goes without saying that the F328Be painted a sumptuous orchestral picture with this material. The instrument sections were nicely spread out with an excellent sense of depth and positioning of each section. I repeat: detail, imaging, and drive for days.
John Williams Live in Vienna with the Wiener Philharmoniker and Anne Sophie Mutter, Deutsche Grammophon, Blu-Ray-Dolby ATMOS, 2020
Switching to the home theater setup of the F228Be, C426Be, the Concerta M8, and my dual subs, I was eager to try out this new live John Williams Blu-Ray and it most certainly did not disappoint! The entire presentation was expansive giving the impression that I was sitting in the 5th or 6th row of the concert hall but with a much better view with all the various camera angles. Strings, woodwinds, and horns all sounded incredibly detailed and the front three speakers, in particular, sounded tonally identical as the music transitioned between them.
My personal favorite pieces were from the Star Wars and Indiana Jones franchises but anytime that Anne-Sophie Mutter came to the stage to play and solo it gave the Revel C426Be an extra reason to shine. Specifically, on “Devil’s Dance from The Witches of Eastwick” and “Hedwig’s Theme from Harry Potter”, the C426Be relayed every maniacal sweep of her bow or tensioned pluck of the stings with an almost fanatical precision. Layers of detail were served up from that center channel during her performances and then when the rest of the orchestra comes in, the C426Be delivered the same level of weight and drive to the music to keep up with the bigger F228Bes.
The House With a Clock in its Walls, Amblin/Universal, Blu-Ray-Dolby ATMOS, 2018
This movie has some of the most enveloping sounds that I have come across on disc and here again, the front three channels of this Revel system made a hugely positive impression. The C426Be continued to match the F228Be sound for sound and several audio pans across the front three speakers sounded completely seamless. But this film’s soundtrack also gave the Concerta M8 surrounds a serious workout with all the varied and interesting magical effects and spatial cues from inside the main character’s enchanted house, magic spells in the garden, and so on. The little M8s also made sound pans from back to front, and back again sound surprisingly robust, even, and believable. Overall, this Revel speaker combination gave a first-class immersive presentation of this fun magical adventure movie.
Measurements by Carlo Lo Raso, Technical Analysis by Carlo Lo Raso and David A. Rich
Bench tests were performed with a Cross Spectrum Labs calibrated UMIK-1 USB microphone connected to my Surface 3 PRO tablet and using Room EQ Wizard acoustic measurement software. On and off-axis measurements were taken with the mic at a distance of 1-meter from the center point of the tweeter. Unlike measurements taken outdoors or in an anechoic chamber that measure a speaker in isolation (and assess straight-line engineering), in-room measurements give a sense of how a speaker behaves in the actual environment that it’s used. Both types of measurements are valuable, I just find in-room measurements interesting as they help shed light on what I am hearing and why. That and I don’t own an anechoic chamber!
This in-room measurement of the Revel F328Be is an 18-point spatial average. A 9-point measurement was made for the right speaker and then repeated for the left speaker. Averaging both speakers is an approach that others have been using and publishing. Using data from both speakers reduces the modal response of the room by averaging the differences in the response of the two speakers from room asymmetries. The technique cannot eliminate them however and, as such, the response below 300Hz remains room dominated. The results show a fairly consistent response from 600 Hz on up, exhibiting a mild trough between 200 Hz – 400 Hz and a milder bump from 400 Hz – 600 Hz. The dip at 160 Hz is a floor bounce (the distance from the woofer to the floor and back to the speaker) which appears in all my speaker measurements in this room. The bounce frequency will vary depending on whether I am measuring a tower or a bookshelf speaker, but it is always there and for this F328Be measurement it appears at 160 Hz. Bass response is solid down to about 30 Hz before rolling off. Treble response seems to be linear with a very gradual decline from 6 kHz – 20 kHz.
Here is the same in-room response laid over another one taken from the Revel F228Be in approximately the same position for comparison. In the figure above the room modes act differently on the larger F328Be with its 3 woofers than on the F228Be.
What causes the difference at 700 Hz – 1kHz is not clear. Perhaps the height difference of the tweeter and midrange. The F328Be’s waveguide is improved but the results are not clear in this figure, we will see the improvements in the anechoic measurements presented later.
Above is the sound power of the F228BE (blue trace) and F328BE (red trace) taken in the Harman anechoic chamber. Since this is sound power, I limited it to 300Hz. Note this graph is normalized. The F328Be is 1dB more efficient. As can be seen, the F328BE does indeed have more low-end extension.
I do not have distortion measurements, but with 3 woofers in the F328BE, I would expect it to produce less distortion below 200Hz.
It is possible to produce the in-room response just from the anechoic data used to make the CEA spin curves. We will show the full data down a few graphs but this is the easiest to understand. Below 100 Hz the anechoic chamber may not properly show the full bass response of a speaker under test. This predicted in-room response (PIR) provides an easy way to take the whole speaker in with one graph but does not show if the direct sound from the speaker matches the early reflection and sound power as the full spin curve does.
It can be seen that the F328Be is very smooth with HARMAN’s target curve which is a slow decrease in the amplitude as the frequency increases. The F328Be’s PIR matches our in-room response graph above 500 Hz.
Above is the predicted in-room response for the F228Be. Harman used a different display tool with the more ragged-looking response, likely chalked up to a difference in smoothing. The high end below 5 kHz drops more in the F228Be which is a result of the waveguide difference.
This is a plot of the NRC listening window for the F328Be scaled from 300 Hz on up which is averaged from the following measurements: 0-degree on-axis, +-15 vertical, and +-15 horizontal. Here we see a nicely controlled and uniform response in the portion of the frequency range that is dominated by the speaker (response below 300 Hz tends to be dominated by the room and can vary greatly). This summation differs from the ANSI/CEA-2034 listening window which we will see shortly. The NRC still presents its anechoic curves with this window. Early room reflections, and attempts to window them out, likely produces the broad bump below 800 Hz. The small dip around the crossover is the only thing that is noticeable in an otherwise “flat as pebbled curling ice response.”
When comparing the F328Be to our listening window measurement of the F228Be, we picked up the strange oscillations not seen in the preceding F328Be curve. We do not know if this is a measurement difference in our data or if it is the speakers. We will see these oscillations again in the wide-angle horizontal radiation pattern of the F228Be below. On the plus side, this curve is more monotonic, although it slopes downward while the F328Be tends to be flat. All of this is likely a result of the waveguide change.
HARMAN has supplied us with the true anechoic data for the F328Be, the ANSI/CEA-2034 curve set, which is shown above. The individual plots on the graph are best understood with a textbook written by Dr. Toole: Floyd Toole, “Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms.” 3rd edition (Taylor and Francis, 2018). Chapters 5 of the text covers these measurements and interpretation of the curve set.
The listening window (green trace) shows a flat response with a slight rise at the transition to the tweeter. The early reflection response which reflects the summation of the reflections that can occur in most rooms on the floor, wall, and ceiling is the top red trace. The researchers who developed the early reflection curve examined 15 different rooms of varying square area and aspect ratios to account for most of the possible first reflections.
The key attribute of this curve is that it should be monotonically decreasing. The other key attribute of the curve is the ratio of the listening window to the early reflection response, which should be monotonically increasing. This ratio is the Directive Index (DI) of the early reflection curve (bottom orange trace). If the speaker’s DI were not monotonic it would introduce an undesired, room-dependent, sonic signature.
A Directivity Index for the power response (bottom purple trace) is also presented. Ideally, it will have the same properties as the early reflection curve and directivity index. The difference is this curve shows reflections that start at any point of the speaker’s response and the reflection may experience several bounces before reaching the ear.
Our NRC listening window curve done with the free REW quasi-anechoic measurement tool looks closer to this six-figure measurement setup than one would expect.
This is the CEA 2034 curve set for the F228Be. Like this speaker’s PIR seen earlier, it uses a different display tool. The ridges seen on the curves look to be smoothing differences. The roll-off above 5 kHz is from the earlier generation waveguide than what’s on the F328Be, which we confirm with our own measurements below.
The early reflection curve looks more monotonic on the F328Be which is likely also a result of the waveguide. On the other side, the directivity index curves do not have the bump in the 1.5 kHz – 3.0 kHz area seen on the F328Be.
And bringing in an even earlier generation comparison, the Revel F208 shows that the older waveguide does not match the directivity characteristics of the midrange and tweeter as well, and the tweeter shows a significant roll-off above 8 kHz. Also, note the dip around 6 kHz which is gone with the Be tweeter.
What we are not seeing in the spin curve is the non-linear performance between the F208’s tweeter and the F228Be’s tweeter.
You can see the improvement in the tweeter linearity with drive in these plots produced by the Canadian NRC for another website here.
The tests are for the Revel M106 and M126Be. With respect to distortion, both tweeters show no distortion of significance at 90dB SPL at 2 meters to the lower limit of the graph of 0.5%. The F328BE tweeter would be expected to produce better results since it is more efficient. The F328BE efficiency is 2.5dB higher than the F208 and 1dB higher than the F228BE.
Above are the horizontal radiation curves of the Revel F328Be from 0 to 30 degrees, scaled from 300 Hz on up. Looking at the angles most closely related to the direct response we see smooth curves with no obvious resonances at the 1/6 octave smoothing used here.
The above F228Be curve set from Carlo’s review 2 years ago looks very similar to the F328Be.
Above are the horizontal radiation curves from 45 to 90 degrees, scaled from 300 Hz on up. Again, we continue to see generally smooth response curves with no obvious resonances and the expected rate of extended off-axis decline.
The F228Be shows a larger drop above 8 kHz which shows that changes did occur with the changes in waveguides. The small oscillation seen in this curve also showed up in the NRC listening window we measured for the F228Be (earlier). We do not know if it is the speaker or the measurement.
The match between the midrange and tweeter looks excellent and about the same as the F328Be. The HARMAN ANSI/CEA-2034 curve set pointed to some differences made by the waveguide that is likely seen at larger angles than the 90-degree maximum here.
Moving to the vertical radiation pattern for positive angles for the Revel F328Be, we see the plots for 0 to +20 degrees, scaled from 300 Hz on up. These show very little change outside of a 2.3 kHz dip as we move beyond 10-degrees indicating a high order crossover preventing the woofer and tweeter from producing much interaction.
The F228Be vertical radiation pattern for positive angles looks about the same as that of the F328Be which is no surprise. Most Revel speakers look like this. There is a fundamental problem in speakers with offset drivers, especially with passive crossovers. The vertical radiation patterns are very different and not monotonic. Coaxial (or concentric) drivers solve this but bring in a whole different set of issues. Put on some pink noise and move from sitting to standing and you will hear obvious changes, although far less than in a speaker with low order crossovers.
In the vertical radiation pattern for negative angles for the F328Be, we see the plots for 0 to -20 degrees, scaled from 300 Hz on up. Here, the crossover area between the woofer and tweeter shows little change outside of a 2.3 kHz dip as we move beyond 10-degrees. The change at 15 and 20 degrees is a little more pronounced than in the positive angles. Again, this is a fundamental problem with offset drivers and nothing much can be done to improve this in a well-designed 5 figure speaker in comparison to one at a tenth the price. Active crossovers do provide a degree of freedom to clean this up a little.
The vertical radiation pattern for negative angles of the F228Be brings no change as expected.
Here is the impedance sweep for the F328Be. Results show much of the speaker’s operating range takes place between 4 to 6 ohms with the 70 Hz to 400 Hz region dipping below 3 ohms. The low impedance in the woofer range is expected with three drivers. The tweeter area looks to have an 8-ohm impedance. Impedance measurements were taken using the Dayton Audio DATS v3 loudspeaker tester and software.
This is not a speaker to be driven by an amplifier that cannot drive a 4-ohm load 20 Hz – 20 kHz with both channels driven. A spec we have seen disappear in most class AB equipment that lacks the requisite number of output devices, heat sinks, and a robust power transformer.
Here is the impedance sweep for the F228Be for comparison. One might expect the impedance to be higher, below 300 Hz, compared to the F328Be since it has only two woofers but that is not the case. Perhaps the F328Be woofers are different with higher voice coil impedance.
And here is the impedance sweep for the Revel F208. This was taken with an AP we had access to when we did the review of this product. This is truncated to 20 Hz.
The tweeter has an impedance of more than half the Be tweeter. This is a sign the Be tweeter is much more efficient and did not need to draw as much power for the same SPL. Needless to say, the low impedance implies more heat and a significant change in driver linearity with increased level. You are paying half the price, so compromises are expected.
For the sake of completeness and to show what can be done with the judicious use of room correction, here is another in-room response of the Revel F328Be taken with the application of ARC room EQ, courtesy of the Anthem STR preamp. The correction was limited to 600 Hz and below to deal with various room issues, everything above 600 Hz is the speaker’s natural response. Now it looks much closer to the Predicted In-Room response, especially below 600 Hz when compared to Revel’s graph right under it. While I used this profile for some of my listening evaluations, much of my impressions were from listening to the speaker’s “au natural.”
Here is the in-room response plot we measured for the Revel C426Be center channel speaker.
And here is the Predicted In-room Response (PIR) of the C426Be from the HARMAN’s anechoic chamber which appears to be doing a good job of predicting the response in the room shown above. This is very different from the predicted response of the tower speakers above 300 Hz despite what looks like an identical midrange and tweeter.
We suspect for this curve HARMAN placed the speaker below the point that the microphone was on-axis to simulate the speaker near the floor when used as a center.
This is a plot of the NRC listening window for the C426Be scaled from 300 Hz on up which is averaged from the following measurements: 0-degree on-axis, +-15 vertical, and +-15 horizontal. Here we see a less controlled curve than the towers in the portion of the frequency range that is dominated by the speaker (response below 300 Hz tends to be dominated by the room and can vary greatly). Note this and all our measurements below are with the C426Be lifted so the tweeter is on axis with the microphone so differences between our measurements and the HARMAN measurements are larger than for the towers in which our mic positions matched.
Strange looking curve here. It looks much worse than the tower above 300Hz. The midrange and tweeter physically overlap in the C426Be which reduces the interference between the drivers, but the spin curve in the transition area is much worse in all but the on-axis measurement. One suspects they did this plot off-axis to simulate the speaker below the listener’s ear near the floor.
Above are the horizontal radiation curves of the Revel C426Be from 0 to 30 degrees, scaled from 300 Hz on up. Looking at the angles most closely related to the direct response we see smooth curves with no obvious resonances at the 1/6 octave smoothing used here.
Above are the horizontal radiation curves from 45 to 90 degrees, scaled from 300 Hz on up. These curves are less monotonic than the F228Be and F328Be. We have no understanding why.
The horizontally arrayed woofers transition to the midrange at 220Hz. In theory, this should look just like the tower speakers with our on-axis measurement.
Moving to the vertical radiation pattern for positive angles of the Revel C426Be, we see the plots for 0 to +20 degrees, scaled from 300 Hz on up. These show very little change outside of a 2.3 kHz dip as we move beyond 10-degrees indicating a high order crossover preventing the woofer and tweeter from producing much interaction. This is so strange. On physical observation the only thing we see is the midrange is squeezed as close as it can get to the tweeter, slightly overlapping the waveguide. This reduces the interference between the midrange and tweeter so it should look better, not worse.
In the vertical radiation pattern for negative angles, we see the plots for 0 to -20 degrees, scaled from 300 Hz on up. Here, the crossover area between the woofer and tweeter shows little change outside of a 2.3 kHz dip as we move beyond 10-degrees. The change at 15 and 20 degrees is a little more pronounced than in the positive angles.
Again, the difference between the C426Be and the towers (F228Be and F328Be) is a mystery. It is not that our measurements are not correlating to HARMAN’s. The spin curves show problems in exactly the frequencies we got the big dip. HARMAN appears to have optimized the vertical radiation above the speaker at positive angles, perhaps by adding some time delay between the midrange and tweeter, assuming that this type of speaker would (in most cases) be placed well below ear level. This can change the optimal axis, but the physics of non-coincident drivers does not change.
Again, we note HARMAN is trying everything here including squishing the drivers closer to each other.
Here is the impedance sweep for the C426Be. Results show about half of the speaker’s operating range takes place above 5 Ohms with 80 Hz to 500 Hz dipping below 4 Ohms. This lower than advertised impedance may be a result of four woofers running in parallel. The tweeter impedance looks very different than that of the F328Be / F228Be which indicates it is a different design. A changed tweeter may point to some of the response issues we saw at the midrange tweeter crossover.
The Revel F328Be and C426Be are loudspeakers that every serious audiophile should audition, if just to hear what is possible from a passive design.
- Same fantastic natural Revel sound quality, just more of it.
- Precise imaging.
- Big soundstage with good off-axis performance.
- Deep, detailed bass response.
- Point out that these are 4-Ohm speakers in the marketing materials.
- Active versions, someday.
There is no doubt that both the Revel F328Be and C426Be are exceptional sounding, highly evolved examples of the passive loudspeaker. If your tastes veer towards a more honest and accurate representation, these babies are as good as they get for their given purpose. While we raised some questions with some of our measurements, in actual use I found the C426Be to be a superb sounding center channel speaker that is an ideal match for any of its stablemates in the PerformaBe range. I’d go so far as to say, as important as the center channel speaker is, don’t mess with the smaller ones. Splurge and get this one. You will not regret it. When combined with the pair of Revel F228Be that I had in the house, it created as seamless and accurate a home theater front end as I have ever had in my home.
The PerformaBe F328Be speakers take what makes the F228Be so good and dial it up several levels. If you have the physical space to let them breathe, you will be duly rewarded with equal parts refinement, detail, and sheer drive power. My only real caveat is that, according to our measurements, these are actually 4-Ohm speakers, so plan your amplification resources accordingly. At $16,000.00 for a pair, the F328Be have ostensibly taken the place once occupied by the now discontinued Revel Ultima Studio 2. The Revel F228Be at $10,000.00 per pair, is legitimately close to its bigger brother in capability and is realistically more suited to most average sized listening rooms. Looking a little farther out in the HARMAN field you also have the $15,000.00 JBL 4367 which may be appealing to a slightly different buyer, but I can testify that it follows a similar rationale to its sound and it matches the sheer lung capacity of the F328Be pound for pound. Offering some additional finish choices on the F328Be would also be a nice option at this price point.
In the final analysis, there is no one perfect speaker but if you have the means and space, you will be hard-pressed to find a loudspeaker more rewarding than Revel PerformaBe F328Be. If you will be using them, or any of its siblings in a home theater scenario, the C426Be is almost an essential addition. If this is the pond you are looking to fish in, then these are the fish you should be looking to catch. Highly recommended!
The author wishes to thank David A. Rich for his assistance in this review.