Sensitivity (SPL at 1 meter/1 watt)
Bass output, subwoofer(CEA-2010A standard)
• Ultra-low bass (20-31.5 Hz) average: 98.5 dB
• Low bass (40-63 Hz) average: 116.4 dB
I would like to have measured the frequency response of the 3D Array with the unit mounted on a wall, but the ersatz wall I attach to my measurement stand for this purpose wasn’t large enough to accommodate the long bar. So I measured it placed atop a 2-meter stand, with the microphone placed at a distance of 2 meters, using quasi-anechoic technique to remove the effects of reflections from nearby objects. On-wall mounting should produce a bass boost of about 6 dB at roughly 200 Hz, which as you can see from the frequency response chart could well have a beneficial effect on the sound.
The crosstalk cancellation circuitry used for the bar’s left and right speakers is designed to work with your two ears rather than a single measurement microphone, so such circuits tend to result in frequency response anomalies when measured by a single microphone. GoldenEar suggested that I could terminate the opposite speaker to eliminate the effects of the circuitry, but told me that it wouldn’t measure significantly better that way. The results I got were above average for a soundbar anyway, so I simply fed the signal from the amplifier straight into the left-speaker terminals.
I measured the left and center channels of the bar, in each case placing the microphone directly in front of the tweeter. I averaged the measurements at 0°, ±10°, ±20°, and ±30°, smoothed to 1/12th octave. I measured the SuperSat 3 satellite the same way. Bass response was measured by close-miking the woofers in the speakers; I then spliced this result to the quasi-anechoic measurements at 180 Hz. Results are normalized to 0 dB at 1 kHz. All frequency response measurements were made with a Clio FW audio analyzer and then imported into a LinearX LMS analyzer for post-processing.
For a soundbar, the 3D Array measures quite well, with fairly smooth response from the left and center channels. Both are remarkably flat on-axis, on the order of ±3 dB, but the interference between the two midwoofers off-axis makes their averaged response across a ±30° window, well, rougher, but still above average for a soundbar. The left was a little smoother, either because of the effects of its different circuitry or because of its position out at the edge of the bar. The center speaker’s response is a little recessed between 1 and 6 kHz. There’s a possibility that this could result in dialogue being down slightly in the mix, but the dip is so broad that you could easily fix the problem, if it occurs, by simply boosting the center channel by 1 dB. (I didn’t notice this issue at all when I checked out the 3D Array in my home theater system.) The interference effects between the midwoofers show up even at ±20°, where there’s a 17-dB dip at 3.3 kHz in the center channel, but the effects are actually milder at more extreme angles of ±45° and ±60°.
The SuperSat 3 measures even better, with flat response on-axis (±2.1 dB — wow!), and smooth off-axis response out past 10 kHz. At 14 kHz, a roughly 1/3rd-octave-wide peak starts to appear at ±30°, and it’s up to about 5 dB at ±45° and ±60°, but the frequency in question is so high that I doubt it will make an audible difference.
Note, though, that the 3D Array and the SuperSat 3 have, respectively, limited and nearly nonexistent bass response, so it will be necessary to use a higher-than-usual crossover point with them. GoldenEar told me it recommends 120 Hz, which seems to work well. I don’t recommend you use these speakers with a receiver that has a fixed 80-Hz subwoofer crossover, because you’ll probably get a “sonic hole” between the sub and the soundbar and satellites.
Impedance (also measured with Clio FW) shouldn’t prove troubling for most receivers and amps. For the left/right channels of the soundbar, minimum impedance is 3.8 ohms at 4 kHz with a phase angle of -3°. The center channel of the bar dips to a low of 5.4 ohms at 6.5 kHz, -7° phase angle. The satellite’s minimum impedance is 3.8 ohms at 370 Hz/-7°. The sensitivity (measured on-axis outdoors, average output from 300 Hz to 10 kHz at 1 meter with a 2.83-volt RMS signal) for the soundbar is about average and for the satellite it’s a few decibels above average, so I wouldn’t hesitate to power this rig even with a $300 bottom-of-the-line receiver.
The subwoofer frequency response measurement was taken using ground plane technique, with the mike placed 2 meters away from the sub so that the contributions of the woofer and the large passive radiator would be properly combined. If there’s anyone out there who actually uses the low-pass filter (i.e., crossover) built into their subwoofer, note that it does not work on the the ForceField 3’s line input.
CEA-2010A output measurements for the ForceField 3 subwoofer were taken at 3 meters and then scaled up 9.54 dB per CEA-2010A requirements so that they are equivalent to 1-meter results. While performing these measurements, I encountered an issue that I haven’t previously seen in a subwoofer. Most subwoofers have built-in limiters that clamp down on the output, usually between 40 and 63 Hz, in order to protect the driver. You don’t notice their effects because they simply limit the maximum volume. The ForceField 3, though, has a protection circuit more like the ones built into many amplifiers. After it encounters an overload condition, it temporarily shuts down until the overload condition is removed. In the CEA-2010 measurements, which use high-level tone bursts spaced 1 second apart, the first bursts at 63 and 50 Hz made it through, but the sub shut down after that. I normally average the results of eight tone bursts, but at these frequencies (marked with an asterisk in the results), I had to use just one tone burst. Typically, though, the difference between averaged and single bursts, given good measurement conditions, is just a fraction of a decibel. Note that I didn’t encounter this issue when using the ForceField 3 in my system, probably because repeated deep bass bursts are very rare in normal listening material (hip-hop being the only possible exception I can think of).
All that said (phew!), output of the ForceField 3 is excellent for its size and price. In the low bass octave (40-63 Hz) it averages 116.4 dB, which is in the range of what I usually measure from similarly sized (but much prettier and more expensive) “lifestyle” subs like the B&W PV1D and the KEF R400b, to give two recently tested examples. I couldn’t get a measurement at 20 Hz, so per CEA-2010 practice I used the 25-Hz result minus 18 dB as the 20-Hz result to calculate the ultra-low bass (20-31.5 Hz) average — which, at 98.5 dB, is still very impressive for the ForceField 3’s price. — Brent Butterworth
Was there anything that I didn’t like about GoldenEar’s 3D Array Ultra High-Performance Soundbar System? When I switched back to my regular Triton Two-based rig, both imaging and soundstage width seemed better. And with two subs at work instead of one, bass also seemed fuller. But the difference wasn’t so great that I breathed a sigh of relief to get my big towers back; the soundbar-based system, to my ears, delivered at least 90% of the main rig’s performance. When you consider that most of that performance was coming from a soundbar, a product category normally consigned to, what did I call it, casual home theater use and background music listening, it’s clear that GoldenEar has pulled off something special here. And when others in your household see how little space the whole thing takes up, I’m sure they’ll find it special, too.