Revel Salon loudspeaker A Visit, Continued 1
Voecks: Back at NRC, we thought off-axis measurements should be held to standards similar to on-axis ones, but we could not specify how. Now that process is much more refined here at Revel. We have a single curve with so much information. Second, the listening process is very much improved. We can place a speaker in the same spot in the room. Even at Canada's NRC we had a problem doing that. You'd have to switch back and forth to see what was coming from the speaker and what was different about the room placement. Third, back then I was a one-man band. Now we have a big engineering department. We have so many experts. But the process takes longer than it used to. That's because I didn't spend as much time on the industrial design as I do now.
We used Madrigal's in-house industrial design department for the Salon. When designing the Salon, I began by specifying the size of the tweeter and midrange. The engineering department used CAD to calculate the radius of the front baffle required to minimize diffraction effects, and identified the type of router needed to carve the radius from the raw cabinet. The job was outsourced to a special firm that had a huge router. This allows us to achieve a unique appearance, so our loudspeakers are not like every other box. We've found it hard to convince people that there is something special inside this speaker unless its fit'n'finish and materials are extraordinary. For example, we had to develop a proprietary high-gloss paint process that is applied over the primer and that will smooth over all the seams in the MDF construction material.
Greenhill: Why does each Salon weigh 250 lbs?
Voecks: We had the budget to make a bigger cabinet. The cabinet volume we selected is a straightforward tradeoff between sensitivity, low frequency extension, and size, so that the Salon can be driven by a moderate-powered amplifier in a large room and still have powerful deep bass with a very low frequency cutoff. We use laser interferometry to look at the cabinet and make it quiet, and this information led us to a very heavy cabinet. Part of its mass is the need to brace and damp a large cabinet. We used a constrained-layer damping technique: You compress a mass—the panels—through the hardware onto a somewhat flexible layer of the rubber material we use. We got all kinds of density of rubber material, then, using laser interferometry, looked at the side panels' velocity and acceleration when attached, checking to see which combination of materials gave the greatest reduction in cabinet-panel velocity or output.
Greenhill: How did you achieve the unusual extension and heft in the bass response I heard in the Salon?
Voecks: The Salon uses three 8" woofers per enclosure, giving more power handling than a single 15" subwoofer. In addition, we spread the heat buildup over six voice-coils rather than one or two. As heat builds up, the impedance goes up. Then the speaker doesn't react linearly with level. We took all our combined expertise at Revel and JBL and put it to work in getting the best bass response we could from the Salon. The Salon's reflex design is tuned to 24Hz (-3dB downpoint). Also, we designed the mouth of the port to play at high levels with very low noise from air movement and low distortion. This is due to the port's large flare and its effective length. We calculated the wind speed in the port, and calculated the output capability so that it would handle high levels. All of these have been optimized.
Greenhill: Which Harman listening room did you use to develop the Salon?
Voecks: Most important was the 5355-cubic-foot Multichannel Listening Laboratory [MLL] (footnote 1). Its speaker mover—a custom-built floor at the front of the room—allows us to perform positional substitutions of up to nine different loudspeakers. Loudspeakers are attached to one of nine pallets that move in 1" increments through a range of 4' forward and backward, while the entire array moves 4' to the left or right of the listener. When tested, each speaker is moved to the exact same place.
That's when the truth comes out. It is essential that different loudspeakers, to be compared, get moved into the same location. A slight difference in location can change the sound of a speaker as it interacts with room boundaries. Even if we have two prototypes of the same speaker with a slight difference in the wiring of the crossover, one is never certain if the differences heard are from the internal wiring, or from differences in the position of the speakers vis-à-vis the listener. If several different speakers are compared without moving them, then one needs a much larger number of tests just to control for position.
Footnote 1: See S.E. Olive, B. Castro, F.E. Toole, "A New Laboratory for Evaluating Multichannel Audio Components and Systems," presented at the 105th Convention of the Audio Engineering Society (September 1998). Preprint 4842 (H-1).