Is all distortion bad?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Discussions like these always seem to include references to various musical instruments. As already mentioned, I think it's important to emphasize the distinction between producing music and reproducing it.

That said, I prefer low distortion in a music reproduction system. Euphonic colorations of various types can be pleasing for awhile but I tire of them rather quickly, more so as I grow older for some reason. Even things like distorted guitars and the like sound more satisfying to me, the more accurately they're produced. In fact, I've come to think these sorts of chaotic, dense sounds can provide some of the most informative tests of reproduction accuracy.

I'm also learning more about mistaking linear distortions for the nonlinear kind. For years I thought my tweeters were producing some audible distortion in the form of an occasional "scratchy" sound, which after careful measurement turned out to be just a small-but-audible response bump at about 10 Khz.
 
Disabled Account
Joined 2010
<snip> Most music will have multiple sounds playing at the same time. Pure HARMONIC distortion does NOT occur by itself. It usually has an ugly friend called IMD. IMD is the mixing products created when two tones meet a nonlinear device. ALL real devices are nonlinear, so some IMD will be generated in them. With few exceptions, most IMD created tones are NOT musically pleasing.

A musical CHORD is two or more musical notes played together at the same time. Most guitar players are familiar with Power Chords. These are chords that have their notes (usually only 2 or 3) chosen such that the IMD products generated are less objectionable than most multi note chords and therefore can be played through a highly distorted amp.
yessir, totally agreed which each word!:cool:
 
Yes, "all" distortion is bad. If you subtract distortion from the music signal and listen to distortion component only, you will find that it sounds just horrible. This applies to low-order distortion as well, regardless it is less audible due to masking.
It's too simplistic to say that an extracted component of the sound will be perceived in the same way when in the combination it was extracted from.

We also seem to all agree that added dither noise (a distortion which can be configured in different ways) is a useful distortion to add to the signal?
 
There have been threads of this nature come and go over the 13 years that I have been a member here. I can guarantee that I (or most anyone) can make two or three different files, each with 1% distortion, and just about anyone can hear the difference between them. I had these files on my computer, but can't find them now. They are easy to make given some software, a pair of identical clean audio oscillators, or a music synth having at least two sine wave oscillators.

When presented with those files and a few others, with one being a pure 261.6 Hz (middle C) tone, then asked to choose the "best" sounding, the pure tone is rarely chosen.
I would suggest that is because our auditory processing does not evaluate sounds based on purity of tone - it builds sonic models which have been learned through exposure to real world sound & pure tones are seldom encountered in the real world - hence our evaluation model is very weak for this particular sonic fingerprint

The files:

261.6 Hz pure sine wave tone with low distortion.

261.6 Hz AND 523.2 Hz low distortion sine waves mixed in varying proportions, with one being mixed to result in exactly 1% THD. I also made 0.5%, 2% and 3% files.

261.6 Hz AND 784.8 Hz at 0.5%, 1%, 2% and 3% THD.

261.6 Hz, 523.2 Hz and 784.8 Hz mixed as follows. set a level of 261.2 Hz, bring in some 523.2 Hz to result in about 0.4% THD, then add some 784.8 Hz top bring the THD up to 0.5% total. Repeat this to make 0.8 H2 with 1.0% total.

These are all "pleasing tones" with the 1 to 2% H2 only and the lower levels of H2 H3 mix usually winning over a pure sine wave for "sounding best" It's better to do this with a music synthesizer that imparts some dynamics such as a medium attack and a slow decay to the tones, such that they sound somewhat like a musical instrument and not just a tone.
See above to explain this

Now make some files that are 261.6 Hz mixed with 1308 Hz or 1831.2 and higher harmonics, or worse, several mixed to result in exactly 1% THD. virtually nobody will pick these files as "best" as they generally sound dissonant......because they are.

Pure H2 is also a "C" note, just played one octave above the fundamental. It will sound good when added with the fundamental in almost any ratio.

The third harmonic falls very close to a "perfect fifth" that is one octave above the fundamental. It is also one element of a triad chord formation, and also sounds pleasing in low to medium amounts.

The fifth harmonic falls far enough from a musical third to create beat notes and begin to sound dissonant if strong enough.

The seventh harmonic is not close to any musically pleasing interval to be liked.

So SOME HARMONIC distortions can be tolerated, and may be preferred over pure tones to some if not most ears.......But, we don't listen to single tones, we listen to music.

Most music will have multiple sounds playing at the same time. Pure HARMONIC distortion does NOT occur by itself. It usually has an ugly friend called IMD. IMD is the mixing products created when two tones meet a nonlinear device. ALL real devices are nonlinear, so some IMD will be generated in them. With few exceptions, most IMD created tones are NOT musically pleasing.

A musical CHORD is two or more musical notes played together at the same time. Most guitar players are familiar with Power Chords. These are chords that have their notes (usually only 2 or 3) chosen such that the IMD products generated are less objectionable than most multi note chords and therefore can be played through a highly distorted amp.

Yes, I agree that we are much more sensitive to IMD in our replay systems (not so in musical instruments, our ears, etc) - why?
 
Administrator
Joined 2004
Paid Member
Sorry it's a very poor analogy because it's tuned to the equal temperament scale, which is mathematical and not "musical"
That's very basic, there is so much more to it. I've had the privilege to work with some world class piano tuners and learned a little about how complex it all is. There is way more to it than just tuning to a tempered scale.
 

PRR

Member
Joined 2003
Paid Member
The equal-temper scale is on the wall in most of our music. It is equally objectionable in *every* key.

But that's not the real issue in Piano tuning. You don't need a temper to hear it. Just two notes "an octave apart". Play A=220. It has overtones. Now play A=440. The fist overtone of 220 is NOT 440, but some higher. The overtones of 220 beat with 440, when they should blend. This is because in the 1800s Steinway and others put fatter and fatter strings on for more power. Strings so stiff the end-effects raise the overtones. (Longer scale-length is the right way to go but soon runs into other troubles.)

"Stretch tuning" spreads the errors. Higher notes are some higher than nominal to blend with lower notes. If the piano maker picked happy strings and lengths, and the tuner understands the process, the sunjective tuning will be "excellent" even though nothing hits the nominal frequency.

Note that most string and wind instruments have off-harmonic overtones. This is in contrast to what you guys are discussing: harmonic distortion is EXACT integer ratios.
 
Last edited:
But that's not the real issue in Piano tuning. You don't need a temper to hear it. Just two notes "an octave apart". Play A=220. It has overtones. Now play A=440. The fist overtone of 220 is NOT 440, but some higher.
If you mean the strongest harmonic of A=220Hz is not 440Hz then you may or may not be right. To say that the second harmonic at 440Hz doesn't exist for 220Hz is not right. Every harmonic will be present, just with differing amplitudes. It is my understanding that for a piano, the second and third harmonics are the strongest, followed by at least another ten harmonics at audible levels. It is the amplitude of the harmonics relative to the fundamental and the attack/decay of each of the tones which gives an instrument its 'timbre'. Due to soundboard resonances and so forth, the timbre will change for higher and lower frequencies.

Btw, the harmonics of two notes will only 'beat' together if the instrument is mistuned.
 
Last edited:
Sorry, but PRR is correct - many natural instruments have overtones that are not exact integer multiples of the fundamental pitches produced. We're talking frequency here, not amplitude. Then there's the mighty Hammond organ, wherein the harmonic frequencies are determined not by string gauges or hole locations but gear ratios (among other things)!

It's kind of amazing that these things sound so good when we put them all together.
 
Last edited:
Yes, "all" distortion is bad....
Agreed, including conventional speaker driver distortion with positive phase second harmonic that blur spatial cues and compress soundstage.
... In my experience, the THD needs to be inaudible (preferably below 0.01%) within the entire audio band.
Is there a system that we can have to enjoy this level of distortion from a whole reproduction chain at ~90dB? Kindly share anything available approaching this state of the art. I found electrostatic driver still generate close to an order of magnitude higher distortion.
Adding second harmonic distortion, even if negative phase, is just wrong....
There is a limited way that an amplifier with a certain amount and phase of second harmonic distortion can reduce total chain distortion by cancellation of speaker driver induced distortion in:
1. Single amplifier to a single driver system or
2. Multiple amplifiers to a multi driver system using line level crossover
Please note that nothing is implied about the amplifier with cancellation property being tube, solid state, discrete, a certain topology, politically correct, conforming to a certain school of doing things or otherwise, and much work still need to be done on the touchy subject.

As a listener, I simply would like to have clean audio reproduction with the least noise and distortion of any kind possible at my listening position, I am not particularly choosy about the means to achieve the objective. :)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.