If they impart no sound character, why have some people commented about an improvement in exchanging components of the same values?
Confirmation bias is rampant in HiFi.
He also used cryoed Caddock MP resistors and Hovland mylar capacitors
Replace the Mylars with PolyPropylene and your problem may likely go away. It still astounds that builders compromise a design for a few pennies. Or is it ignorance?
Additionally, there is space, so you may want to try a felt donut around the tweeter to reduce off axis energy and also ameliorate diffraction effects at the box edges.
The amplifier / cable combo [don't be shy tell us what you have] may not
like a certain combination of load and drive request.
There may be a resonance due to cone breakdown.
If the tweeters had ferrofluid, it maybe toast.
If there are any electrolytics in the XO, replace them with PP.
Do you have any room treatment? Or is it a far too typical plaster box?
Finally, some recordings are not playable on systems due to Monte Carlo accumulation of faults in recording, playback electronics, cables, speakers and room.
