KEF Reference 104/2 Floor standing speakers

KEF were, as far as I recollect, the first British company to come up with three-way loudspeakers in which separate detached enclosures were provided for each unit, although the concept had been seen before in Europe, particularly in France. KEF's Reference 105, with its swiveling head embodying the mid-range and tweeter units, set a pattern which was soon copied by others and is still popular in spite of the rather ungainly appearance acceptable in only the largest rooms. The 104/2 retains the separate enclosure theme but, in this case, it takes the form of a shaped box resiliently fixed to the upper front of the main cabinet. This section contains two 110mm Bextrene cone units with a 33mm fabric dome tweeter fitted in a flared recess between them. These are familiar KEF drive units, but this ingenious construction avoids the use of the normal loudspeaker chassis by mounting the cones on the front of the box and their magnets on the rear. The completed assembly, filled with a quantity of "new high-density polymer" damping material, has the makings of a respectable small loudspeaker on its own account; in this application it handles all signals above 150Hz, which is a very large proportion of the total energy in any programme signal - KEF estimate 85%.

The low-frequency cabinet of the 104/2 is even more unusual for there are no loudspeaker units to be seen! Based on an idea which I believe originated with Peter Baxandall and was developed into a workable form by KEF's Laurie Fincham, it is called in the present design "Coupled cavity bass loading". At each end of the 900mm tall cabinet is a sealed section occupying something over a quarter of the internal volume. Each has a 200mm papercone bass unit fitted in its inner panel; the driver in the lower box is mounted normally, while the other is inverted, i.e. hanging from the upper box-a distortion cancelling ploy devised by Hugh Brittain for his 'Periphonic' system of the 1960s. The two units are rigidly bonded together by a metal rod passing through their magnet assemblies. The remaining cabinet volume between the units is fitted with a smoothly contoured duct of 110mm diameter exiting just below the MF/HF box, and all the low frequencies are pumped through this. There are several ways of looking at this arrangement (e.g. as two infinite baffle enclosures coupled to a resonator) but they all lead to the trading of increased efficiency for a limitation in bandwidth (we never get something for nothing). In this case 150Hz is the upper requirement and a chosen bandwidth of about 1.5 octaves sets the lower roll-off around 55Hz.

Sensitivity a priority

It seems clear that the demands in overseas markets for loud and sensitive loudspeakers have been given high priority in this design, and the KEF team have risen to the challenge-whereas most competitors have accepted that quantity and quality are uneasy comrades and been content to let efficiency slide in their upmarket ranges. To this end, KEF have allowed themselves a little license by making this a 4-ohm loudspeaker; so, if used with the usual 8ohm amplifier, it will sound louder because it is drawing more power. In justification KEF claim that many, if not most, nominally 8-ohm loudspeakers have lower impedances in one or more areas of their frequency range and, because of this, "every good amplifier is overdesigned to cope with a wide range of loudspeaker loads". To take advantage of this possibly optimistic presumption, the 104/2 has "Conjugate load matching" which is claimed to make it look to the amplifier like a 4-ohm resistor at all frequencies, thus avoiding the so called 'wattless current' demands of reactive loads. They use this concept to justify the headline, "How to double the power of your amplifier without having to buy a new one". I do not consider this to be an altogether realistic proposition, and users who pursue it will certainly find that the operating temperatures of their amplifiers will rise accordingly, for there is no getting away from the fact that real watts have to come from somewhere. I would also direct attention to the 92dB sensitivity figure in their specification which is referred to an input signal of 2.83 volts. Into 4 ohms this corresponds to 2 watts, so the sensitivity for the more usually quoted dB/watt rating is actually 89dB. This is still respectably high, but this way of specifying sensitivity has already misled a number of writers into believing it to be better than it really is.

How it performed

Connecting a pair of KEF 104/2s in place of the Quad ESL63s I had been using required a swift dash for the volume control. Not only were the KEF loudspeakers much louder, but the forward quality of their sound dramatized the effect. I would say that this is a loudspeaker which always sounds louder than it really is; having said that, I must report that it can be made to go very loud indeed without any sign of distress (except perhaps to the listener's hearing if turned up too far). Sampling broadcast speech at realistic levels, both male and female, showed that this is not a totally accurate loudspeaker, for one could never get the volume 'right'-although this was easy to do on reverting to the ESLs. The 104/2, as I have noticed in others in the KEF range, possesses a degree of over-articulation which exaggerates the consonant components of certain voices; for an example, listen to BBC Radio 2's Jimmy Young. This does not seem to be a tweeter problem, although there is a cert a in sibilant emphasis from it, but rather a characteristic of the 110mm Bextrene mid-range cones. In spite of the care devoted to the avoidance of coloration, there is also an unnatural veiling and hardening imparted to female voices. The former effect, an over-emphasis of the formant tones, can easily be assessed on that lovely Kirite Kanawa CD of the Songs of the Auvergne and the latter imparts an uncomfortable readiness to the liquid voices of the Ricciarelli/Terrani ductto in Rossini's Stabat mater. On much music, however, the overall sound can be attractive and revealing; e.g. the Philip Jones Brass Ensemble and the Mozart Horn Concertos, but this was not maintained over all types of programme. Extreme bass from large pipe organ or bass drum failed lamentably, often sounding an octave higher than it should; this was not due to doubling but to the altered ratio of fundamental and harmonic content. It was intriguing to revert to the ESLs, so often accused of poor bass, and hear the balance restored. A further difficulty was experienced in attempting to follow the cello line in chamber music; a recent "Sounds in Retrospect" choice, the CD of the Schubert String Quintet in C major was not given justice here, and it may be that this indicates some discontinuity in the respective contributions of the two boxes. Similar vagaries were noted in the stereo picture; some sounds had pinpoint positional accuracy; others, without apparent reason, flew to one or other of the loudspeakers, while others still were vague to the extent that, on one occasion, a knowledgeable but puzzled visitor doubted whether I had the loudspeakers correctly phased and insisted on reversing one.

It was on turning to the sounds of rock, pop and jazz that these 104/2s came into their own; the concentration of low-frequency activity into the area encompassed by electric bass guitar was dramatic and the glowing edge imparted to brass and saxophone quite stunning. Percussion percussed as never before, and the highly equalized voices of singers holding microphones against their teeth really bit the air. Everyone with the slightest interest in this area of music who has heard these loudspeakers has been considerably impressed. One, searching for high praise, compared them to the much more expensive Linn Isobarik (also using KEF drivers) although I am not sure what they will make of that remark in Tovil.

This loudspeaker has already proved to be extremely popular with those listeners I have mentioned, whose choice of musical experience falls within that area where subjective impression outweighs accuracy; an area where, in truth, reality does no t exist. Its undoubted ability to go very loud indeed pleases them and should also find application in the more sophisticated discotheques now abounding. Overall, though, I come back to the wider performance criteria demanded of a 'Reference' loudspeaker, for this is where KEF themselves have placed it; sadly, in this case, I find myself asking, with reference to what?

KEF Reference 104/2 Floor standing speakers photo