KEF’s Reference series is a stalwart and a fixture in our audio business. There has been a Reference model in the range since the Model 104 of 1973. Despite decades of leading-edge development, the Reference Series stays true to the original goals of that first model 42 years ago: using sophisticated analysis to define and control the loudspeaker environment, and building speakers with state-of-the-art production and quality control techniques. The means whereby these elements come together have changed radically over the decades, but the forces that created that Model 104 are the same that drive today’s new Reference 1.
Let’s unpack that first paragraph a little, because it’s more than just a throwaway opening gambit. Using computer design today is not such a big thing when practically everyone in the developed world over about 2 and younger than about 92 has at least one computer to their name. But just 42 years ago, using computers in the development of loudspeakers was NASA-grade engineering – four years later when I was one of the first teenagers at my school to study ‘computer science’, we were submitting our programs on Teletype and even Hollerith cards to the only computer in the borough. This was the horse and buggy era of computing, and yet KEF was already modelling its loudspeakers on computers, and this dedication to the application of science to technology has run like a red thread through the company and its products, but most pointedly through the Reference models.
This dedication to science-based audio was what sparked the company’s ‘total system design’ philosophy at the start of the 1980s, which saw the Model 103.2 incorporating drivers, cabinet, and crossover network as a complete project to be developed together, years ahead of its rivals. It was the impetus behind the Eureka/Archimedes project, which attempted to liberate the loudspeaker from the tyranny of the room it sits in (and which resulted in the Uni-Q drive unit, used in the Reference range in 1989’s Model 105/3). It was this uncompromising objectivity that developed technologies such as conjugate load matching for designs like the Model 103/4 of 1992, and ultrasonic improvements to the Uni-Q in the Reference Model 201 et al of 2001. Two things come out of this potted history; the Reference models all hark back to the laboratory-maintained scientific and manufacturing reference points, and that in more than 40 years of continuous reference points, KEF doesn’t feel the need to change that often.