What is worst - even or odd order of distortion

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Even / Odd

Even order distortion tends to fatten the sound, odd order sharpens it.

Odd order distortion in small quantities can make things sound faster / punchier and more dynamic - artificially of course.

Of more importance is the harmonic extension, low order harmonics are relatively unobtrusive, extended order is much more easily detected by the ear.

e.g. A well set up cartridge generates realtively high levels of 2nd / 3rd order harmonics. When a cartridge mistracks the absolute distortion level changes much less than one expects, but the distortion harmonics extend to the top of the audio band, where the ear is more sensitive.

It's why CD players used to sound so poor, they generated low levels of extended, non-musically-related distortion at low signal levels. It made them sound bright, sharp and nasty.
You might find the following of interest:

<a href="http://hyperphysics.phy-astr.gsu.edu/hbase/music/pianof.html">The Piano</a>
"<i>Actually, the hammers are in sets for the keys which sound two or three strings. The striking point is about 1/7th of the way along the string to discriminate against the 7th harmonic resonance, which is significantly out of tune with the equal tempered scale. This "7th harmonic problem" can be illustrated by examining the musical intervals formed by a an exact <a href="http://hyperphysics.phy-astr.gsu.edu/hbase/music/harmon.html#c1">harmonic sequence</a>.</i>"

even/odd distortion

Quick Version:
Even order harmonic content is less offensive than odd order.

Longer (better) Version:

Harmonics are produced at multiples of the fundamental frequency.
Ex. If the frequency amplified(the fundamental) is 100Hz, then the 2nd harmonic is 200Hz, the 3rd is 300Hz, the 7th at 700Hz, etc...

Also, If the frequency is doubled, the same note in a musical scale is produced an octave higher.
Ex. The musical note "A" is produced at 440Hz, 880Hz, 1760Hz, 3520Hz, etc... (also at 220Hz, 110Hz, 55Hz, etc...)

By doing some simple math, you can see that the 2nd, 4th and 8th harmonics would sound "better" because they sound like the same notes 1, 2 and 3 octaves higher.

The other harmonics (3rd, 5th, 6th, 7th, 9th) would sound dissonant or "crappy-***"(as an engineer would put it) because they are related to the fundamental only in a mathematical sense, and not a musical one.

Tube amps produce mostly 2nd order harmonics. This accounts for the "fuller" (but less accurate) sound they produce.

But I don't want to spark a "tube vs SS" war...
I don't think so Dave...

Check this op amp comparison (Japanese). It tests a number of common op amps in a non-inverting gain of 10 circuit, and provides the distortion results and FFTs of the residuals.

Note that the devices that get the best distortion measurements, and have the fewest high order harmonics, have very high amounts of global feedback - and are screamin' fast devices.

IMHO, global feedback is not by itself inherently good or bad; it's just another tool in the engineer's kit. Global feedback has the desirable property that it can clean up after several stages at once, rather like the catalytic converter on a car.

It's not a panacea, though; applying global feedback to a non-linear amplifier only makes it appear linear. Where global feedback really runs into problems is when the amplifier can't slew fast enough or doesn't have enough gain at all frequencies across the band, which results in what was once called TIM but should really be called slewing-induced distortion; or conversely when there is too much loop gain and phase delay, causing oscillation.

Global feedback earned a black eye from early SS power amps and op amps, which didn't have enough slew rate, gain bandwidth product, or open-loop linearity. These designs measured OK on simple test signals, but fell apart when you pushed them with real-world audio signals and loads. The result, in hindsight, was predictable, as was the backlash against global feedback that followed.

Note that most of the designs that use small amounts of global feedback are using comparatively large amounts of local feedback, in the form of emitter/cathode/source degeneration or inter-stage feedback loops.

At least that's the way it appears to me from 30 years or so of armchair audio engineering. I reserve the right to change my story in the face of hard evidence.
the most enlightening article

Since this seems to be the perfect place to comment on this, here is a link I found just recently that is really fantastic. If you have ever wondered why SS amps with great THD numbers sound like crap and tubes can sound so good, this is the article for you.

The link is here . It is an article by Lynn Olson, who, among other things, is the designer of the Ariel TL speakers.

Some may argue that it more appropriate in the Tube section, but in this case I disagree :cool: . Generally the tube fans already know that THD is a small part of the equation (preaching to the choir). As an SS fan, this article was an amazing revelation.

It really exposes the fact that attempting to judge the performance of a 20kHz spectrum with 100+ db of dynamic range (typical audio CD) with a single numerical percentage is a waste of time. Most revealing :D .
fashinating article, one of the best I read.
As a chemist I can tell you that the MS parallel, though colorful and apparently insightful is just a crock of ****. The able chemist could do a pretty good job in separating the ingredients, quantify the amounts of fats, proteins, decomposed proteins, aromatic essential oils and spot without the shadow of a doubt which was fast food which was ***** food.
These bad paralles are the main source of misunderstanding, and where the subjectivists and objectivist views clash.
The problem is that a better sounding audio amplifier is not going to cure cancer, nobody really cares about spending a few hundered million dollars to quantify the sound of a piece of audio equipment. Audio measurements are left at the mercy of the ingenuity and creativity of EEs (a contraddiction in terms) with enough money to afford the appropriate and very costly instrumentation.
The instrumentation is the other side of the equation, in order to measure something you need a detector. How lucky, we are trying to detect audio frequency! Any moron has been given a pair of fine detectors for free.
This means a fauna of self tought-motivated individuals with little or no knowledge of science, with no instrumentation, messing around with a soldering iron and a few bucks worth of components trying to convince other people that all that really matters is what their ears can hear. Subjectivism was born.
I can only imagine the frustration of the open mined objectivist, knowing full well that the measurement should be able to distingush between the good the bad and the ugly, and yet not able to do so. Many of such obectivists moved to the other side and joined the clueless club, many rejected totally the concept and went into audio seclusion, many still walk the razor blade.
Everything is Ok until we get to this kind of statement:
If you have ever wondered why SS amps with great THD numbers sound like crap and tubes can sound so good,

I'm guessing what was really meant is something like:
If you ever wondered why some tube amps with relatively poor numbers can sound better to you than some solid state amps with great THD numbers.

Sorry. :(
mlloyd1 (who's obviously feeling pretty cranky this morning)
Easy, mlloyd1 ! Too much grump can be hazardous to your health :D .

I'm guessing what was really meant is something like:
If you ever wondered why some tube amps with relatively poor numbers can sound better to you than some solid state amps with great THD numbers.

That is indeed what I meant. One thing I don't think I mentioned is that I am an SS fan, so I have no interest in making statements to the contrary :p .

grataku : I have no idea where you are coming from with the "MS parallel" thing :p . Are you referring to something in the article I posted, or another post?
Olson makes some good points in that article. That article and the responses to it, BTW, made it clear to me that I am probably not on the same wavelength as many of you. Oh well.

I would think that with massive amounts of computing horsepower available cheaply, and with the quality of A/D conversion in "prosumer" digital recording gear rapidly improving, spectrum analysis of distortion products should be done far more widely in the coming years. I hope this foreshadows a re-evaluation of how to correlate measurement with perception in audio.
"Feedback is like a credit card. Used conservatively, it's very helpful."
"Feedback is medicine, overdose becomes poison..."

I know exactly what you mean from my own experience. But how do you explain this? Why should high feedback ruin the sound? One of Naim's contributions to audio amps seems to me to be the way they have cracked this problem - their BJT amps use about as much feedback as they possibly can and do not sound terrible (argumentative ;) ). What's their secret?
Hello all,

1000 thanks for enlightening the subject, learned a lot!

Hello traderbaum

sorry for coming with such a crude commonplace: See a global feedback loop as a regulator's loop. Regulators always are late, an error signal is needed to get the regulating structure into action. It just is the question, how late and does it hurt? Can we live with delay, overshoot, phase mess?

To my ears zero loop feedback amplifiers sounded *speed-of-light* fast and utterly relaxed so far. Talking about tube amps.
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.