Gradient Revolution loudspeaker John Atkinson March 1997
When listening critically to a product for review, you can spend so much time trying to detect faults that it only slowly dawns on you that not only is the product doing almost nothing wrong, it is actually performing superbly well. Such was the case when I auditioned the Gradient Revolution loudspeaker. Back in May 1995 (Vol.18 No.5), Dick Olsher raved about the performance of this unusual floorstanding loudspeaker from Finland. Combining an aperiodic mid/treble head-unit featuring a SEAS-sourced concentric HF/MF driver, crossing over below 200Hz to a dipole woofer using two 12" cone drive-units, the Revolution's design should make it relatively immune to the vagaries of room acoustics.
Indeed, DO was impressed by the Gradient, summing up his experience thusly: "The Gradient Revolution is certainly not a spectacular loudspeaker in audiophile terms. It lacks killer bass, and doesn't have the highly etched and overly detailed midrange to which many audiophiles are attracted. Neither is the Revolution's tonal balance distorted euphonically in a manner that would earn it instant converts, as so many 'different is better' designs tend to do. The presence region isn't elevated, and the lower mids aren't even lusciously fat; instead, the tonal balance is basically neutral.
"The Revolution suavely navigated the core of the midrange. Delicate harmonic detail was allowed to bubble to the surface of the musical tapestry, and low-level detail was nicely resolved. Male voice was naturally reproduced without emphasis or added color....The Revolution's greatest strengths are an organic wholeness, solid imaging, excellent microdynamic expression, and a convincing rhythmic drive born out of a pure and quick bass range."
However, Gradient took issue with some of my measurements that accompanied the original review, specifically the Revolution's apparently low sensitivity—a B-weighted 81.7dB/2.83V/m vs a specified 86dB/2.83V/m—and an impedance that dropped to just 3 ohms in the power region of the upper bass. While he had no explanation for the discrepancy in sensitivity, the Revolution's designer, Jorma Salmi, conjectured that a short-circuited capacitor in the midrange crossover filter might explain the impedance problem. Certainly some simulations he sent following the publication of the review duplicated my measured impedance very well (though it didn't explain why both of our original review samples suffered from the same problem).
The impedance of the new samples of the Revolution (fig.1) certainly looked a lot more healthy than that of the first (fig.2). The magnitude now remains above 6 ohms in the midrange, and drops to 5 ohms only in the midbass. The woofer's free-air resonance can now be seen to lie at a low 20Hz. However, the large difference between the maximum and minimum impedances will still make the Revolution a suboptimal choice for use with an amplifier having a high source impedance, such as a single-ended design.
Fig.1 Gradient Revolution, new sample, electrical impedance (solid) and phase (dashed) (2 ohms/vertical div.).
Fig.2 Gradient Revolution, original sample, electrical impedance (solid) and phase (dashed) (2 ohms/vertical div.).
My estimate of the new Revolution's voltage sensitivity is still significantly below specification, at an identical 81.7dB/2.83V/m (B-weighted). I rechecked my figures but came up with the same result. I must say that I was not surprised by this, as the Revolutions did indeed sound about as sensitive as my long-term reference BBC LS3/5A, and significantly quieter than my B&W Silver Signature or the Epos ES 12 for the same volume-control setting. However, I found the 30Wpc Pass Labs Aleph 3 amplifier would raise levels of 92dB in my room, loud enough for all but the most headbanging of musical moments.
The original sample's crossover problem did not seem to have affected its frequency response to any great extent. Fig.3 shows the response of one of the first review samples on the tweeter axis at a microphone distance of 50", spliced at 300Hz to the sum of the nearfield woofer and midrange responses. Fig.4 shows the same measurement for one of the new samples. Both graphs are dominated by the spike at 26kHz due to the metal-dome tweeter's "oil-can" resonance and some discontinuities at the top of the audible band, these due to the symmetrical environment presented to the tweeter dome by the midrange cone. Otherwise, the responses of both old and new samples are impressively flat.
Fig.3 Gradient Revolution, original sample, anechoic response on tweeter axis at 50", averaged across 30 degrees horizontal window and corrected for microphone response, with sum of nearfield woofer and midrange responses below 300Hz.
Fig.4 Gradient Revolution, new sample, anechoic response on tweeter axis at 50", averaged across 30 degrees horizontal window and corrected for microphone response, with sum of nearfield woofer and midrange responses below 300Hz.
This came as no surprise; in my own auditioning of the Revolutions, the first thing that had struck me (after their lowish sensitivity) was how neutrally balanced they were, coupled with a refreshing lack of midrange coloration. Voices, in particular, sounded free from resonant overhang and were superbly articulate. And, as DO found, the soundstaging was stable, well-defined, and spacious. I also agree with Dick about the Revolution's top octave, in that it is somewhat reticent. While the speaker is not mellow-balanced, the extreme highs do lack some subjective air.
The bass was also an area where I was surprised by what the Revolution had to offer. The box-free 12" moving-coil units offered an appealing combination of bass articulation and low-frequency extension, easily decoding Daniel Lanois's muddy, bass-heavy mix of Emmylou Harris's HDCD-encoded Wrecking Ball CD, for example (Asylum 61854-2). Fig.5 shows the Gradients' in-room response, spatially averaged to minimize the effect of the room (footnote 1). As well as the flatness—presumably due to its dipolar design and the effectiveness of its crossover, it meets astonishing ±1.3dB limits in-room, from the 32Hz 1/3-octave band to the 10kHz band—note the excellent extension, the speakers not starting to roll off until below 30Hz in my listening room! At the high end of this graph, the concentric tweeter's limited dispersion leads to a somewhat suppressed top octave, this audible if not particularly bothersome.
Fig.5 Gradient Revolution, spatially averaged, 1/3-octave response in JA's room.
My auditioning of the revised samples of the Revolution confirms Dick Olsher's excellent opinion of this speaker. Its modest appearance and size belie a sophisticated, superbly engineered, high-performance design. Highly recommended.—John Atkinson
Footnote 1: For this in-room spectral analysis I use an Audio Control Industrial SA-3050A spectrum analyzer with its own microphone and average six measurements at each of 10 separate microphone positions for left and right speakers individually. The 120 original spectra are then averaged to give a curve that, in my room, has proved to give a good correlation with a loudspeaker's perceived balance.—John Atkinson