Revel's Kevin Voecks
Kevin Voecks: In other words, why did it take so long? We didn't bring out a new Salon right away because we didn't have to. We listened to competitors' speakers, and the Salon1 still sounded better in our double-blind listening tests, so we resisted change for the sake of change. Second, it took more than three years of active development by our team to perfect the new beryllium tweeter you see in the Salon2. Third, Stereophile drops a product from "Recommended Components" after three years, even if the product leads the field in sound quality. That got us thinking at Revel about introducing a newer version of the Salon.
Larry Greenhill: Did you simply upgrade the drivers in the Salon2, or were more extensive changes made?
Voecks: The design of the Salon2 was more complicated than a driver update. Rather, our design team started with a clean sheet of paper, as if there had never been a Salon1. What topology would be best for the Salon2? We considered all possibilities, including going to a three-way design. After long thought, it became apparent that the basic design configurations of the Salon1, Studio1, and Voice1 were still our best guidelines, but improvements could be made.
Greenhill: While the Salon2 looks very different from the Salon1, it turns out that both are four-way speakers with drivers of similar sizes, and a similar volume enclosure for the bass drivers. Is the internal functioning of the Salon2 really that different?
Voecks: The design of the Salon2's enclosure is very different, and turned out to involve real serendipity. We asked all our dealers, "What kind of external appearance do you want?" Over and over, we heard the same thing: The new speaker can't look like the Salon1, which some described as "an alien that had just landed." Also, the dealers asked us to get away from the large wooden side panels. The Salon2 should be more subtle, more elegant, more understated, and have lots of beautiful wood, not just in the side panels. We realized that we had the same wish list, from an engineering point of view. So selecting a slim tower with an oval cross section for the Salon2 was a win-win situation.
There are other examples of structural differences. The Salon2's reflex port tube has been redesigned to perform better by changing its shape and making it fire downward rather than to the rear. The new speaker's curved backyou recall that the rear of the Salon1's lower section was flatmeant that we had to move the port's opening from the back of the Salon2, where the porthole would cause an interruption of the speaker's profile when looked at from the side. When we moved the port's opening to the bottom of the Salon2's enclosure, we had to make certain that there was adequate clearance between the bottom of the enclosure and the top of the base plate to allow the air to flow smoothly out of the port without interference from below.
Greenhill: You described the Salon1 as looking like "an alien that had just landed." Thinking back, I had a different impression. A number of reviewers, including me, found the Salon1 daringly modern and postindustrial, and striking out in a new direction from the boxy loudspeakers we had seen before. Its unique design was acknowledged and praised when Stereophile awarded it Joint Best Loudspeaker of the Year in 1998.
Voecks: The Salon1 met both its aesthetic and engineer design goals 10 years ago. The radius on the tweeter head assembly of the original Salon provided a perfect baffle shape for the tweeter and midrange, although it was extremely difficult to manufacture. I personally was very gratified to be able to have the flexibility of being able to modify the Salon1's appearance just by bolting on a different pair of wooden side panels. However, Revel may have provided too many options in the Salon1. There were many different finishes and side panels. This was more burdensome for us to manufacture, for dealers to store, and for the customer. Just matching side-panel wood grains to body color took some work. As a result, we've limited the Salon2 finishes to only two: glossy mahogany and glossy piano black.
Greenhill: What are the major engineering changes in the Salon2?
Voecks: You recall that the Salon1 required a rear tweeter for it to deliver a uniform power response in the room. We were able to eliminate the rear tweeter in the Salon2 because of the increased power output of the new beryllium front tweeter, coupled with benefits from its new waveguide design and baffle shape, to deliver the same response on axis and 60° off axis. Why is that important? The 6070° off-axis response constitutes the first reflection from sidewalls, which is demonstrably audible and psychoacoustically important.
Greenhill: You've often said that Revel's double-blind listening tests allow you to voice loudspeakers in a reliable manner. What did these tests reveal when you compared the Salon1 and Salon2?
Voecks: Yes, we used the double-blind, position-independent listening tests in Revel's multichannel listening lab to compare the old and new Salons. Our panel consistently prefers the Salon2 to the Salon1.
Revel's double-blind listening technology, developed by Sean Olive, has remained the same over the past 10 years. Each loudspeaker tested is moved by a computer-controlled drive belt into the same exact position in the room behind an acoustically transparent scrim cloth. The room is kept dimly illuminated to avoid visual recognition. Sean has trained and certified seven people at Harman to be reliable listeners; ie, to give the same unknown speakers the same ranking, time after time. Mark Glasier and I are the most frequent listeners. Over a thousand hours of listening tests were involved in the design of the Salon2. By the way, using larger numbers of trained listeners, Sean Olive has found that listening preferences are not affected by country of origin, region of country that one lives in, or whether one prefers rock music or country music.
Listening tests over the past 10 years have taught us one other thing. Above the midprice range of loudspeakers, there is no correlation between the sound quality and the loudspeaker's price. Although many high-priced loudspeakers do perform adequately in our listening tests, the most expensive speaker in a given double-blind listening test may be the least preferred by our listening panel.
Greenhill: Did the Salon2's listening tests correlate with its laboratory measurements?
Voecks: We can easily predict which loudspeakers will be preferred and which will not. For example, a speaker that measures poorly on axis will never be preferred in our double-blind listening tests. Also, listening tests correlated best with our measurements conducted in an anechoic chamber, like the one we have at the Harman facility in California. In that chamber, we perform 72 measurements from different positions that, all together, make up a sphere surrounding the speaker. This technique of measurement and the software used to process the resulting data are called Spin-o-rama.
We do a traditional on-axis measurement as part of the "Spin-o-rama" measurement, but the "listening window" is the most useful measurement of the direct sound. It is comprised of 10, 20, and 30° horizontal off-axis measurements, and ±10° vertical measurements.
Another listening preference that is reliably reproduced [is the effect of] speaker resonances. Research shows that resonances are audible and influence listener preferences. Speakers with resonances are not preferred by our listening panel. The best way to detect a resonance is to compare a single-point measure with an averaged on-axis response processed with Spin-o-rama. If a peak that is found with a single measurement goes down when you look at the averaged and processed multiple measures, then the peak was due to microphone placement. If the level of the peak does not drop when many on-axis measurements are averaged, then the peak is one coming from the speaker and due to a true resonance.
Spin-o-rama plots also describe a loudspeaker's off-axis response. Up to10kHz, the off-axis response also will predict which speakers are preferred by our listening panel. Those speakers without large resonances and which have an off-axis response similar to the on-axis response are most often preferred. That's why we endeavor to make the off-axis response similar to the on-axis measurement..
Greenhill: Why would a customer purchase a Salon2 if he knows that the less-expensive Studio2 [reviewed by Kalman Rubinson in March 2008Ed.] incorporates the same design principles?
Voecks: The Salon2 moves more air and has greater output, particularly in the bass. The Salon2's three 8" woofers have a combined area equivalent to a 14" woofer, but the heat generated is spread out among three voice-coils. This means that you won't get the heat buildup that leads to dynamic compression. (As voice-coils heat up, impedance goes up and leads to a mismatch in a speaker's filter network.) The Salon2 is more resistant to dynamic compression than the Studio2 because it has more drivers to dissipate the heat. The Salon2 also has a smaller midrange than the Studio2. This leads to a better match between tweeter and midrange drivers, helping control the Salon2's off-axis response.
Greenhill: Why do the rear-panel bass controls of the Salon2 do so much more than those of the Salon1? My in-room measurements suggest that the Salon1's Low-Frequency Compensation control is limited to varying the frequency response in a narrow 10Hz region around 50Hz. The Salon2's Low-Frequency Compensation control depresses the bass response over a wider band between 20 and 100Hz.
Voecks: With a passive filter network like the crossover in the Salon1, there is not much variation you can build in. But putting in a Low-Frequency Compensation control is hard to do in any loudspeaker. With the Salon2, I was able to achieve a more useful low-pass function than I had with the Salon1, primarily due to insights derived from my design work on the much less expensive Performa1 loudspeaker.
Greenhill: I noticed that the Salon2 has a larger optimal listening zone, often called a "sweet spot," than the Salon1. Was that another design goal?
Voecks: Yes, very much so. The response 60° off axis can predict the apparent size of a loudspeaker's sweet spot. The only reviews I've read that measure a loudspeaker's off-axis responses are those done by John Atkinson in Stereophile. I designed the Salon2 to generate a broader sweet spot by employing a small midrange driver, a low crossover point for the tweeter, and steeper, 24dB/octave crossover slopes with a much narrower overlap. Our double-blind listening tests show a broader sweet spot, and our lab measurements show a good off-axis response. Both the Salon2 and Studio2 should show a good off-axis response when John Atkinson measures them.
Greenhill: Did your work in professional audio help in the design and manufacture of two-channel audiophile loudspeakers?
Voecks: Yes. JBL's test and manufacturing facility for pro audio is across the street from Revel, and has really helped. Revel and JBL engineering staff exchange ideas in a productive way at our joint technology meetings, which Floyd Toole started. For example, one design goal for pro-audio loudspeakers is that they never can break, even under the most difficult electrical and performance conditions. JBL manufactures speakers with superb cooling features for their driver voice-coils, and uses multiple drivers and voice-coils to dissipate heat during live concerts. We use the same design approach to limit heat, dynamic compression, and driver damage. JBL was the first to use flat-wire voice-coils, and we have adopted that feature in some of the Revel loudspeakers.