Gramophone Dreams

Intrigued by his small, full-range drive units, I used to chat with the late Ted Jordan at hi-fi shows when I lived in the UK. So, when Herb Reichert told me that he was reviewing an LS3/5a-sized standmount speaker that used the latest version of the Jordan driver, I volunteered to measure it.

The EJ Jordan Marlow is sold direct from the UK and costs £1960/pair (equivalent to $2650/pair at the time of writing) including free shipping to the US (footnote 1). HR enthused about the sound of the Marlow in his April 2020 Gramophone Dreams column: The EJ Jordan Marlow "offers a unique form of unmitigated clarity that will not be easily matched by conventional boxes with multiple drivers and crossovers," he wrote. "Its superfocused imaging and extraordinary transparency made recordings sound unusually direct and unsullied." He did caution that the Marlow's 4" drive-unit "cannot move enough air to sound big, strong, or real," adding that the Marlow "cannot deliver commodious deep bass or play loud." Nevertheless, he found that when he played deep-bass dub reggae and sub-bass ambient jazz, the Marlow went low enough to satisfy his needs.

Like the LS3/5a, the Marlow's grille fits over a recessed front baffle. I asked HR: Had he auditioned the Jordan speakers with or without the grilles? "I listened both ways, but mostly with the grilles on," he replied. His characterization of the Marlow's sound quality in Gramophone Dreams was with grilles. Accordingly, I left the grille on for these measurements.

I used DRA Labs' MLSSA system, an Earthworks microphone preamplifier, and a calibrated DPA 4006 microphone to measure the Marlow's farfield behavior, and an Earthworks QTC-40 mike for its nearfield and in-room responses.

EJ Jordan specifies the Marlow's sensitivity as 86dB/W/m; my B-weighted estimate was slightly higher, at 87dB(B)/2.83V/m. The impedance isn't specified. Measured with Dayton Audio's DATS V2 system, the magnitude (fig.1, solid trace) remains above 8 ohms in the treble and most of the bass with a minimum value of 6.13 ohms between 360Hz and 400Hz. The electrical phase angle (dashed trace) is occasionally high, which will increase the loudspeaker's demand for current from an amplifier. The EPDR (footnote 2) drops below 4 ohms between 50Hz and 65Hz, between 120Hz and 295Hz, and between 900Hz and 2.5kHz. The minimum EPDR is 3.06 ohms between 161Hz and 176Hz. The Marlow will work best with amplifiers that have no problems driving 4 ohms.

I investigated the enclosure's vibrational behavior with a plastic-tape accelerometer. The highest-level mode was at 176Hz, on the back panel. This mode was also present at lower levels on the top and side panels. There was also a strong mode at 1.2kHz on the side panel (fig.2), though this resonance is too high in both frequency and Q (Quality Factor) to have audible consequences.

The saddle centered on 41Hz in the impedance magnitude trace suggests that this is the tuning frequency of the port on the front baffle. The red trace in fig.3 shows the port's nearfield response. Its output peaks at the tuning frequency and the upper-frequency rolloff is clean. The blue trace below 300Hz shows the response of the woofer, again measured in the nearfield. It has the expected minimum-motion notch at the port tuning frequency, which is when the back pressure from the port resonance holds the cone still. The black trace below 300Hz in fig.3 shows the complex sum (amplitude and phase) of the nearfield woofer and port outputs. The usual boost in the upper bass, due to the nearfield measurement technique, which assumes that the drive units are mounted on a plane that extends to infinity in both dimensions, is absent. The EJ Jordan's woofer alignment is therefore overdamped, which suggests that the Marlow will benefit from the low-frequency reinforcement that result from a placement close to the wall behind the speaker.

The black trace above 300Hz in fig.3 shows the Marlow's farfield response averaged across a 30° horizontal window centered on the tweeter axis. The speaker's output rises by 5dB in the upper midrange, presumably because with no crossover, baffle-step compensation isn't possible (footnote 3). The output drops in the presence region before rising to a 7dB peak at 11kHz. I repeated the farfield measurement without the grille. The only difference was about 1dB greater output in the mid-treble region.

I use an Outline computer-controlled turntable to examine the off-axis behavior, rotating the loudspeaker under test in 5° increments. Measured in this way, the EJ Jordan's horizontal dispersion is shown in fig.4. (The traces are normalized to the response level with the center of the drive unit, which thus appears as a straight line.) The radiation pattern is well controlled in the midrange, but the small suckout at 2.75kHz in the on-axis response fills in to the speaker's sides. The Marlow's output becomes very directional above 4kHz. This will tend to compensate for the on-axis peak at 11kHz, but without toe-in to the listener's position, the sound will be too dull. The vertical radiation pattern over a ±45° range, again normalized to the response on the drive-unit axis, is shown in fig.5. As expected (because there is only one drive-unit and so no interference to affect the vertical response), it is similar to the speaker's behavior in the horizontal plane.

The red trace in fig.6 shows the Marlows' spatially averaged response in my own room. For reference, the blue trace shows the spatially averaged response of the Vivid Kaya S12 that HR reviewed, also in the April 2022 issue, and the green trace that of the GoldenEar BRX, which I reviewed in May 2020. Like the EJ Jordan speaker, the Vivid and GoldenEar are standmounts, and all three pairs were placed on 29" Sanus stands in the same positions of my room to generate the traces in fig.6.

All three speakers would have benefited from placement closer to the room boundaries, as their low frequencies shelve down below 150Hz, though to slightly different extents. The spatial averaging tends to average out the peaks and dips below 400Hz that are due to the room's resonant modes. Above 400Hz, the Vivids and GoldenEars have very similar in-room behavior, with the slight downward slope in the treble due to the increased absorption of the room's furnishings and the tweeter's increasing directivity at high frequencies. (One thing you don't want to see in this type of measurement is flat treble output, as this would sound too bright.)

The outlier in fig.6 is the spatially averaged response of the Marlow. The boosted upper midrange seen in the farfield response in fig.3 is present here also. With a noise signal, I could hear this as an "aww" coloration. I could also hear the small peak above 10kHz. With music, the boosted upper mids may well have contributed to the added sense of recorded detail that HR reported hearing. But whether the Marlow is perceived as having too much upper midrange or suppressed lower midrange and treble will depend on the recordings being played.

In the time domain, the Marlow's step response on the tweeter axis (fig.7) indicates that the full-range drive-unit is connected in positive acoustic polarity. Other than a slight discontinuity just after the peak, the step response is time-coincident, which will contribute to the stable, accurate stereo imaging noted by HR. Some oscillations with a period just below 1ms can be seen in the decay of the step response. The Jordan's cumulative spectral-decay plot (fig.8) has an associated ridge of resonant energy at 1.15kHz, the frequency of the notch in the on-axis response, and there is a faster decaying ridge at 11.2kHz.

So, what to make of the EJ Jordan Marlow's measured performance? (My measurements were similar to those provided by the manufacturer.) Of necessity, all loudspeaker designers have to make trade-offs between different aspects of performance In the Marlow, the use of a single drive-unit allows superb performance in the time domain. The trade-off is the frequency-domain behavior.

I am reminded of something the late Art Dudley suggested in 2004. We were discussing Stereophile's Recommended Components, and I mentioned a loudspeaker he was reviewing that had, shall we say, "idiosyncratic" measured behavior.

I asked Art what he thought I should say in my concluding remarks. "Write that this loudspeaker's sound character will be for special tastes only," he replied. He said he wished we had a symbol to denote that a product was "for special tastes" or "for special systems." "[This] would be great for speakers that have imperfect amplitude response but superb dynamics and sensitivity," he said. "This would be the hardware equivalent of a 'conditional rave' for music aimed at special tastes—Captain Beefheart, say. The Spotlight Kid is without a doubt a Class A album, but it sure isn't for everybody!"

The measured performance of the EJ Jordan Marlow suggests that this, too, will be a loudspeaker for special tastes or special systems.—John Atkinson

Footnote 2: EPDR is the resistive load that gives rise to the same peak dissipation in an amplifier's output devices as the loudspeaker. See "Audio Power Amplifiers for Loudspeaker Loads," JAES, Vol.42 No.9, September 1994, and stereophile.com/reference/707heavy/index.html.

Footnote 3: This adjusts the loudspeaker's output to compensate for the fact that its radiation pattern changes from omnidirectional at low frequencies to just the forward direction above a frequency related to the size of the front baffle. See youtube.com/watch?v=MlFsBgS_rAg.