Linn Komri loudspeaker Measurements

Sidebar 3: Measurements

Like the massive Snell XA Reference loudspeaker, also reviewed in this issue, the Linn Komri was too bulky to lift off the ground for the usual set of quasi-anechoic measurements I take with the DRA Labs MLSSA system. The presence of a floor reflection will therefore limit the farfield measurements' resolution in the midrange. However, I was able to lift the 176-lb Komri (but not the 210-lb Snell!) onto my Outline speaker turntable, so I could examine its off-axis behavior in detail.

I estimated the Komri's voltage sensitivity to be just under 87dB(B)/2.83V/m. This was both to specification and right in the center of the range I have encountered in my measurements. The Komri, however, is a relatively difficult load for an amplifier to drive. Fig.1 shows how its impedance magnitude and phase change with frequency. First, the magnitude drops below 3 ohms above 3.5kHz, reaching a minimum value of 1.7 ohms at 10.9kHz. This in itself will demand a lot of current from the amplifier with music that has appreciable high-treble content.

Fig.1 Linn Komri, electrical impedance (solid) and phase (dashed). (2 ohms/vertical div.)

But look also at the dotted trace in this graph. This is the electrical phase angle, and it reaches quite a large value in two regions. The first, in the middle bass and below, is due to the crossover between the Komri's passive upper-frequency and powered-woofer sections. In general this angle will be benign, as the passive drivers will experience very little drive signal in the affected band. But the phase angle still becomes fairly extreme in the crossover region, where the speaker will be handling large input voltages. At 107Hz, for example, a lowish impedance magnitude of 3.67 ohms coincides with a capacitive phase angle of 54 degrees. The situation is similar in the low treble, where there is a combination of 54 degrees phase angle and 3.6 ohms at 3.14kHz. Amplifiers incapable of easily swinging large output currents will get hot driving the Komri to reasonable levels.

The impedance traces, however, are free from the usual glitches that indicate cabinet resonance problems. Examining the output of a plastic-tape accelerometer fastened to the various enclosure surfaces revealed very few vibrations. Fig.2, for example, is a cumulative spectral-decay plot calculated from the accelerometer's output when fixed to the Komri's top panel. Two modes can be seen in the 360Hz region, but these are very low in level. They were even lower in level on the speaker's sides. The Komri largely behaves as a nonresonant box, which is commendable.

Fig.2 Linn Komri, cumulative spectral-decay plot of accelerometer output fastened to center of top panel. (MLS driving voltage to speaker, 7.55V; measurement bandwidth, 2kHz.)

Turning to the speaker's acoustic behavior, fig.3 is a composite graph showing the Komri's farfield response averaged across a 30 degrees horizontal window on its supertweeter axis, spliced to the complex sum of the lower-frequency units and superimposed on the individual nearfield responses of the powered woofers and the upper-bass driver. The powered section covers the bandpass between 25Hz and 90Hz, with rapid fourth-order rolloffs visible above and below that region. While small peaks at 200Hz and 450Hz can be seen in the woofers' out-of-band response, these are low enough in level that their effect on the speaker's sound quality should be minimal.

Fig.3 Linn Komri, anechoic response on supertweeter axis at 50", averaged across 30 degrees horizontal window and corrected for microphone response, with the nearfield active woofer and lower midrange responses and their complex sum plotted below 500Hz, 350Hz, and 300Hz, respectively.

What alarms me, however, is the fact that the passive upper-bass driver extends down too far in frequency to smoothly cross over to the active section. There is also a slight peak between 70Hz and 80Hz in the drive-unit's nearfield behavior before it rolls off very steeply at 24dB/octave. The result is a slight measured excess of energy in the midbass that might obscure low-frequency clarity. On the other hand, it might add a sense of bass "power."

Moving higher in frequency, the Komri's on-axis balance is smooth but not flat. The curve can be viewed in two ways, depending on what frequency regions the listener's ear seizes on as a reference and what kinds of music he or she prefers. Most probable, in my opinion, is that the equivalence of the acoustic levels in the midrange and mid- to high treble will make the region between, which is suppressed by 2-5dB, be perceived as recessed or distant. The result will be that the Komri's sound will lack immediacy. Alternatively, the listener could perceive the gentle slope down in the response from the midrange through the presence region as being the Komri's predominant character, with then the elevated mid-treble adding "sparkle" and clarity. Certainly that's what I heard with pink noise, the 6-12kHz octave sounding isolated against a smooth background when the speaker was in an open space.

However, when a loudspeaker is listened to in a real room as opposed to an anechoic chamber, its on-axis response is not the only factor that influences the perceived sound quality. The manner in which the speaker excites the room's reverberant field, which will depend on how it behaves off-axis, will also be a major contributor to the perceived balance. Fig.4 shows the Komri's lateral dispersion. It can be seen that the speaker has very wide dispersion in the same presence region where the on-axis response lacks energy, and narrower dispersion in the regions above and below.

Fig.4 Linn Komri, lateral response family at 50", from back to front: differences in response 90 degrees-5 degrees off-axis, reference response on supertweeter axis, differences in response 5 degrees-90 degrees off-axis.

In rooms of small or medium size, particularly rooms with fairly lively acoustics, this will result in a somewhat bright reverberant field that will balance the laid-back on-axis response. It's possible, therefore, that in all but very small rooms (where it might sound too bright) and very large, highly damped rooms (where it will sound too laid-back) the Komri will actually sound fairly neutral in the treble, despite its apparently less-than-optimal measured frequency-domain performance.

In the vertical plane (fig.5), a large suckout appears 10 degrees above the supertweeter, equivalent to an axis level with the top of the cabinet. But as long as the listener's ears are at a height between the center of the upper midrange unit and the bottom of the recessed upper woofer, the balance will not change too much. The supertweeter, by the way, is 34" from the bottom of the enclosure, which means that when the Komri is used with its bases, this drive-unit will be at almost exactly the 36" listening height indicated as average for seated listeners by work done for Stereophile in the early '90s by Stereophile Guide to Home Theater editor Thomas J. Norton.

Fig.5 Linn Komri, vertical response family at 50", from back to front: differences in response 20 degrees-5 degrees above tweeter axis, reference response, difference in response 5 degrees below tweeter axis.

In the time domain, the Komri's step response (fig.6) is complex. As best I can tell, what you can see in this graph are very short, positive-going steps from the supertweeter, tweeter, upper-midrange unit, and upper woofer. The powered woofers actually appear to be connected in inverted polarity, but their slow-risetime contribution cannot be seen in this graph. Note the various ripples between the 4.5ms and 5.5ms markers in this graph; I suspect these are due to reflections of the sound from the drive-unit "gantry." When you look at the Komri's waterfall plot (fig.7), there appears to be quite a lot of hash throughout the treble. However, I believe that this is due to these reflections, not to resonances per se.

Fig.6 Linn Komri, step response on supertweeter axis at 50" (5ms time window, 30kHz bandwidth).

Fig.7 Linn Komri, cumulative spectral-decay plot at 50" (0.15ms risetime).

On the face of things, some aspects of the Linn Komri's measured performance look somewhat disappointing, particularly when you take into consideration its price of $40,000/pair. But I suspect that the speaker's in-room behavior is actually better than you might think. What did bother me was its very demanding impedance characteristic, which mandates the use of beefy amplifiers.—John Atkinson

8787 Perimeter Park Boulevard
Jacksonville, FL 32216
(888) 671-5466