I have been having some friends over, at different times of course, and adjusting the xover on my test speakers.
In my latest experiment covering several months, I have been adjusting a 9k, 9,000 ohms, resistor by less than 0,1 ohms. The 9k ohms is in parallel with a 16 ohm fixed resistor which is in parallel with a full range driver. By doing such, I am altering both the tonal balance across the full range driver, and the relationship between the full range driver and woofer.
We found one can perceive such differences in tonal balance. (This is not the same as spl across the entire audio band.)
To be super cautious, let's use 0,1 ohm rather than hundredths, and let's leave out the full range driver impedance. We will just use the 16 ohm resistor parallel with 9k ohms.
9,000 in parallel with 16 ohms.
A. 15.97160603
9000.1 in parallel with 16 ohms.
B. 15.97160635
Divide B by A and we obtain 1.00000002 .
That is the resistance change just across the 16 ohm resistor when I changed the parallel 9,000 ohm resistor to 9,000.1 ohms.
And we have not included the parallel full range driver impedance (Z), which would make the ratio 1.00000002 substantially less.
The figure represents a tonal balance change of -120db to be on the cautious side. -130-140db is more like it.
I can perform the experiment using either a silver contact switch (10 milli ohms contact resistance) or a variable potentiometer across a very small fixed resistor, in series with a very very large resistor.
There are a couple of important points.
1. The perception is somewhat dependent upon the music played.
I have been using much more variety of music for testing.
2. One has to understand that weighting of harmonics is involved. Olsen's research has found that the higher the harmonic, the more sensitive the "ear" is to changes.
Changes to the 3rd harmonic is more easily perceived than changes to the 2nd harmonic. 10th more than 9th etc.
Cheers
Steve