KEF Reference Series Model Four loudspeaker Measurements
JA measured the KEF Model Four using the DRA Labs MLSSA system and a calibrated B&K 4006 microphone and provided me with the results after I had completed my listening tests.
The KEF's calculated sensitivity measured a very respectable 90dB/W/m. Though this is slightly below specification, the fact that this figure is B-weighted might explain the difference. Its impedance characteristic is shown in fig.1. With a minimum impedance of 3.2 ohms at about 70Hz, and a notable increase in the phase angle below 20Hz, this is a relatively demanding load. The two dips at 28Hz and 70Hz define the twin woofers' bandpass tuning.
Fig.1 KEF Reference 4, electrical impedance (solid) and phase (dashed) (2 ohms/vertical div.).
Fig.2 shows the responses of the internal, bandpass-loaded drivers as measured nearfield at the port, and the summed on-axis response of the exterior drivers on the tweeter axis (43" from the floor). The bandpass port output shows well-suppressed modes (the top-end rolloff of the port output is very smooth). Its LF response rolloff appears to be greater than expected, presumably due to an additional series capacitor in the feed to the bandpass woofer, probably to minimize LF overload. The strong low bass response, -6dB at 28Hz, is clearly visible.
Fig.2 KEF Reference 4, acoustic crossover on tweeter axis at 50", corrected for microphone response, with nearfield port response and complex sum of the midrange unit responses plotted below 500Hz.
The overall response—complex sum of the nearfield bandpass and midrange responses combined with HF response averaged across a 30 degrees lateral window—is shown in fig.3. The slight rise in the lower-midrange/upper-bass might explain the problems I had taming the loudspeaker's warmth, though as loudspeakers go this rise is not great. There is a rise between 1 and 3kHz which might explain the slightly forward quality I noticed, but this is small and well-controlled. In fact, the response in fig.3 is, overall, a superb result.
Fig.3 KEF Reference 4, anechoic response on tweeter axis at 50", averaged across 30 degrees horizontal window and corrected for microphone response, with complex sum of nearfield midrange and port responses plotted below 300Hz.
The horizontal response family of a KEF Model Four, with any on-axis response deviations subtracted out so that the reference response appears flat, is shown in fig.4. The off-axis response rolls off quite smoothly, with the off-axis response helping to fill in the averaged response for some small on-axis irregularities. However, the upper treble does roll off quite rapidly to the sides. This could well account for the slight lack of air and closed-in sound noted in the review, which even in the end I could never entirely eliminate. The vertical response plot (fig.5) is well-maintained off-axis, dropping significantly only when the listener's ears are level with the top of the cabinet (about 49") or the lower-midrange driver (about 29")—both unlikely listening positions.
Fig.4 KEF Reference 4, horizontal response family at 50", normalized to response on tweeter axis, from back to front: differences in response 90 degrees-5 degrees off-axis; reference response; differences in response 5 degrees-90 degrees off-axis.
Fig.5 KEF Reference 4, vertical response family at 50", normalized to response on tweeter axis, from back to front: differences in response 20 degrees-5 degrees above tweeter axis; reference response; differences in response 5 degrees-10 degrees below tweeter axis.