JBL makes the important distinction between "sound producers" (musical instruments) and "sound reproducers" (commercial PA, home stereo) in electronics.
Colouration, distortion is fine and natural with musical instruments. THD isn't so important- the woods, varnishes used in a Stradivarius give its distinct sound. JBL electric guitar speakers can give high distortion and peaky freq. response.
But downstream, even going back to the 80's, THD is specmanship. A number for figure of merit.
I'm reminded of Sansui amplifiers that piled on the loop feedback for competitive/best THD numbers, but in use they went unstable, blew up and ended up in the repair shop en masse. Same for Douglas Self's Blameless early designs, they could not take any clipping.
And of course today you have people tweaking Spice sims and thinking they have perfection when that number is really low. But not a realistic result.
What I find important is the out of band stuff (pun intended?) and those paths.
Example: Maplin MOSFET amp poking around with a scope and I find RF on the O/P stage source. AM radio.
Of course- loudspeaker wiring and crossover inductors are lovely antennas, even a dipole, and guess what gets back into the power amp at its output. Same for headphone amps. But people don't really know.
Tube amps with RF on the filaments. It's not hard to peek with a scope there, hams know better but audio people are oblivious.
My point is filtering out RFI affects sound quality but there is no formal test for EMIRR.
On the subject of capacitors, the tone cap in Jimi Hendrix's guitars is the one you want. Nissea Green Chicklet (polyester film) cap. LOL.
See how a celebrity association with a part endorses it, meanwhile the science of paper-in-oil, ceramic, wax, polyester "sound" etc. is underground and the subject of religious wars.
Yes, one can bifurcate audio design into “production” and “reproduction”.
As Helmholtz discussed a century ago, musical notes from different instruments differ due to their overtone content (in the steady state), as well as their different transients (attack, decay, etc.). The harmonic overtones in a violin note are not “distortion”, but the natural result of a one-dimensional resonator (string, air column, etc.). Non-linear electronics are an inherent part of many electrical musical instruments, since pure sine waves from a good signal generator are boring. Guitar amplifiers are usually inappropriate to the reproduction from a CD of a string quartet.
In reproduction, THD is a useful
measurement of performance, along with IMD (intermodulation) that is more directly related to degradation of complex musical signals. THD is defined straightforwardly, but there are multiple ways to define the measurement protocol for IMD. EMI, hum, and other non-indigenous components added to the music can be annoying as well.
In RF design, one often uses 3rd-order IMD measurements, since the odd-order products are close to the original tones and muck them up in an unpleasant manner. Similarly in audio, the 2nd and 3rd order harmonics may be hidden by the harmonic components of the signal, but the IMD products are not harmonic.
This is why 3rd-order harmonic distortion is usually considered to be “worse” than 2nd-order, since the non-linearity that results in 3rd-order harmonics also results in 3rd-order IMD near the original tones.
I once posted a similar discussion to an audio forum, and someone agreed with me that 3rd-order distortion was bad, because he saw that the 3rd harmonic of A=440 Hz was 1320 Hz, which did not appear in a table of musical frequencies. The 2nd harmonic 880 Hz did, since it was an octave higher. He did not understand that the table was for the conventional “well-tempered” scale used on pianos, where the intervals are tweaked to sound equally wrong in any key, as an engineering approximation to allow practical keyboards.
Home
Help
Search
Login
Register
