To bi-wire or not to bi-wire?

I would like to add 2 other reasons. One is skin effect that starts at gage 20 at 20kHz. Many woofer cables use gage 7 to preserve low impedance (damping factor). Second reason are the Eddy currents, generated by speakers, getting from speaker to speaker (tweeter to midrange or midrange to tweeter) in spite of crossover (far from perfect) and amplifier's output impedance separated by (inductive 0.5uH/ft) impedance of the cable (equal to about 0.5 Ohm at 20kHz = 1/10 divider with speakers source impedance). Effects are very small but remember incredible range of our hearing instrument. Bi-wiring creates divider of the mentioned 0.5 Ohm cable impedance and very low amplifiers output impedance. Amp's output impedance of 0.05 Ohm at 20kHz will create 1:100 divider with speaker impedance before it gets to the other speaker - 10 fold improvement.

My speakers are bi-wired with a shotgun cable and sound slightly more "airy" in comparison to non-biwired (shorted terminal). Is it worth to pay for that about double (of expensive speaker cable)? I don't know - probably not, but I treat cables as non-perishable items and over-invest a bit.

The fact that according to many users bi-wiring improves some speakers but has no effect on the others might be related to design (steepness and configuration) of the crossover.

actually our "hearing instrument" does not have an incredible range and is not all that accurate. A 3 db spl sound change is barely perceptible to the average person. While the audio range is defined as 20 to 20k most people cannot hear frequencies near 20 k and most people over 50 cannot hear 15k. Many years ago I repaired tv's as a part time job. I was in my 20's and worked with two guys in their 50's, I could hear the horizontal oscillator vibration on some sets - crt's ( more likely something derivitive of it from mechanical vibration - but regardless - a high frequency) which is 15,750. The other guys in the shop could not. Their hearing otherwise appeared normal, i.e. normal conversation etc. This is not merely anecdotal, , high frequency hearing loss with age, presbycusis, is very common and audiology testing does not go beyond 8k - the frequency losses are evident at 4 and 8 k with mild cases showing 30 db attentuation from normal. That is my point on a lot of my posts - the instruments we have available to measure everything that has to do with hearing (different than something to measure our tastes in what we hear or what the sounds mean to us - which is where the art comes in) are orders of magnitude more sensitive, accurate, and resolving, than our ears. So the values measured with such instruments should be the base from which to evaluate the effects of many of these quasi technological solutions.

Musicnoise...Your statements about lack of HF hearing are correct, but the conclusion, that HF response of an audio system is unimportant is not correct. Hearing tests for frequency response are done using sine waves. I can't hear a 14KHz sine wave, but I can detect when music is limited to 14KHz. (14 KHz was a while ago and it is probably worse now). In seeking an explanation I have come to believe that the ear senses, not only pressure change, but also the rate of change. This would correspond to steepness of the sound wavefront. A 14KHz tone that is not a sine wave can have a wavefront steepness corresponding to a 20 KHz sine wave. I haven't tried it lately, but supertweeters with response to 40 KHz and higher are audible to some people.

Musicnoise - with normal hearing, one can hear intensities from 0 dB to 140 dB. This corresponds to power ratio (defined as ratio of the highest audible intensity to the lowest audible intensity) equal to 100,000,000,000,000 - I would call it incredible range. Often structures cannot sustain sound levels and vibrations at 140dB SPL that we can.