[nick_charles]

Jim leSurf is a retired professor of Physics & Electronics  at St Andrews University in this page he talks about Super Tweeters  LINK - and his general audio pages are always fascinating !

 

[watchnerd]

So basically he's positing that a super tweeter may be 'audible; if the crossover mucks with frequencies in the audible range.
 
Okay, sure --- but also a bit 'well, duh....'

 

[spruce music]

  So basically he's positing that a super tweeter may be 'audible; if the crossover mucks with frequencies in the audible range.
 
Okay, sure --- but also a bit 'well, duh....'

Maybe, but how many people have extolled with supertweets that can be switched off that they can hear a difference and never think the reason is still in the audible range?  In my experience about 100% admittedly with only a few people having such a speaker.
 
Jim LeSurf writes pretty solid articles and many are quite interesting. 

 

[gregorio]

 
how many people have extolled with supertweets that can be switched off that they can hear a difference and never think the reason is still in the audible range?  In my experience about 100% admittedly with only a few people having such a speaker.

 
In my experience, it would be closer to 0% but admittedly my experience is mainly audio pros and I don't know any audiophiles with super-tweeters.
 
Let's not forget, that even with super-tweeter crossover distortion, it might still represent an "improvement" under certain conditions. If, for example, we are playing high sample rate files with significant ultra-sonic content, it's entirely plausible that the distortion of a super-tweeter's crossover is less than the IMD of a standard speaker system trying to reproduce that same ultra-sonic content. Ergo, the super-tweeter system could measure (and sound) better.
 
Obviously the best solution is to avoid both super-tweeter crossover distortion and IMD in the first place, by not having a super-tweeter in the chain and by band-limiting the signal to near the limits of a standard speaker system (20kHz ish). Which is why I recommend no higher than 44.1 or 48kHz sample rates as a consumer distribution format.
 
G

 

[charleski]

Speakers have problems with phase-alignment that have most impact at high frequencies. A 1kHz tone has a wavelength of ~34cm, a 15kHz tone has a wavelength of  ~2cm. The shorter the wavelength, the greater any impact of a small physical mis-alignment. Most speakers avoid this by using a single radiator to handle the entire high-frequency spectrum, and he's pointing out that introducing an extra radiator that covers this area can result in serious problems (unless the design specifically addresses this issue). Plopping a supertweeter on top of your existing cabinets is practically guaranteed to produce audible changes simply by de/constructive interference.

 

[watchnerd]

Sometimes I'm so glad I use electrostats.