Once a credible and reliable free-field anechoic or quasi-anechoic frequency response is available, it is loaded into the simulator and a theoretical crossover model is developed in the simulator. The next step is to physically build the crossover and then set-up to measure the frequency response of the speaker system with that crossover in-room. This is a key step: in our opinion it is absolutely essential that the speaker's behaviour in the typical customer's listening environment is understood; this is more relevant than the anechoic response. But, from a perspecive of the designer's personal pride, and in keeping with Harbeth's BBC traditions, every effort is made to achieve a good anechoic frequency response and simultaneously a good in-room response. Where these must differ for diffraction or whatever reason it is listening experience which must be maximised.
This part of the design process - in mono with a single speaker - is slow and challenging but if mistakes are made here, the entire design process onwards (including all the listening tests in stereo) will be compromised and may be a total waste of time and effort. So if this iterative basic-mono crossover design takes man months (it always does) that process cannot - dare not - be hurried. Nor is it team work: Harbeth's designer works entirely alone.
So, perhaps for the first time, a speaker designer is exposed in curia. The designer takes no pride in the tortuous journey; it is arriving at the destination of the design process with a new acclaimed product to market which is the reward for this dedicated, concentrated attention to detail over hundreds of hours.
This video was edited down from an hour session during which time the camera was running continuously. As there was (inevitably) much repetition of the process which follows a loop of adjust - measure- check simulation -adjust etc. this 9 minute clip shows one cycle only. We hope that it helps de-mysify the design process.