The phase errors talked about in the article were not due to differing cable lengths. Even with absolutely identical cables, the loading presented by the two drivers is very different. Since a cable is a reactive component, this leads to a fairly complex system.
The model used in the article simulates the cable as an 8uH + 50mR impedance, which does not seem an entirely unreasonable simplification. They simulate the woofer as 500uH + 6R and the tweeter with a rather more complicated circuit. In my scenario, the passive high and low pass filter components are absent but the reactive components of the cable and drivers are still present.
He was seeing up to 3 degrees of error introduced in the crossover region, purely because of the bi-wired cabling. I won't have the passive crossover, but the characteristics of the two amplifiers will not be identical so I guess the phase errors will still be of this magnitude.
My drivers are vertically aligned with a central separation of 250mm. With the drivers in perfect sync there would be zero phase error with ear at the central height. At 2.5Khz, 1 degree is equivalent to 340ms-¹ / 2500s-¹ / 360° =0.38mm. At my usual listening position 2.5 meters away, that represents a vertical distance of about 10mm.
Therefore, I think you're probably right. I do sometimes move my head by more than 1cm from the optimum position.
I guess if I really want to know, then I should simply wire up the speakers to a sweep oscillator and analyse the result with a dual channel scope. I suspect the mechanical response of the drivers is actually the larger issue, so testing with a mic in addition to the voltage at the speakers would be interesting.
I might even listen to it and see if it sounds better, but that means building it first.
*blows dust off 'scope and soldering iron*
_________________jonbwfc's law: "In any forum thread someone will, no matter what the subject, mention Firefly."
When you're feeling too silly for x404,
youRwired.net