What harmonic does to sound?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
As an amateur, I always wanted to built a good sounding power amp. From this site, I learn alot and it comes to this question. A good sound reproduction cannot be separated to its harmonic. While in theory, harmonics are distortion, which should be eliminated, I found that good sounding amps in the market are not completely doing that. Tube amps are the main example. It is said that tube amps produces big second harmonic so it sounds good to most of the people.
In some thread, I read that an audio power amp with excellent measurement figures do not sound so good. Also in some thread, it is said that audio power amp that is too "clean" produces sterile sound.
From the cct itself, like if we use differential at the front end, it is said to eliminate the even harmonic. Maybe this is why the JLH power amp do not use differential, only use 1 transistor for its front end. This can be seen more in tube amps.
What harmonic is really add pleasant sound? What is really the meaning of odd harmonic, even harmonic, second harmonnic etc? If differential is canceling even harmonic, is there a circuit that cancels odd harmonic? It seems even harmonic is the one that makes good sound.
 
There is no recipe on how to bake a good sounding amp. In a SS amp, even order harmonics are said to give a more pleasant, tube-like sound, but it's not really going to sound like a tube amp, 'tube sound' also involves some dynamic compression. The nature of the second & even harmonics make the sound have 'more body' and be less harsh. Accurate is not always good, but accurate does not have to mean that it sounds worse than a tube amp :) That's why there are people who won't put a tube amp in their system no matter what (this, BTW, is the solid-state forum :)

One can not just say that what measures good sounds bad. It's just that measurements cannot say how an amplifier will sound, and similar (in measurements) amps sound totally different. You may find amps that measure very good and have a very good sound, amps that measure good but sound like crap or any other combination.

I don't know if I read this in these forums but a wiser person said:

If it sounds bad and measures bad it's bad.
If it sounds bad and measures good then something's wrong with it.
If it sounds good and measures good everything's fine.
If it sounds good and measures bad then you've measured the wrong thing.

I'd also mention that the degree of subjectivity in audio is rather high, not to mention the very short-term memory when it comes to qualifying sound/quality.

Anyway, I'd say that you have to trust your ears to judge a bad amp from a good one :)
 
My personal rant follows - take it how you will.

The question also is what does it mean to "sound good". I have found the following link confirms some of my suspicions/pejudices:

http://www.enjoythemusic.com/magazine/viewpoint/0601/audiohell.htm


My problem with "good sounding" tube amps is that they really do sound "good" but that after while everything played starts to have sound the same in some sense and the experience becomes boring. I presume this to be the a sign of harmonics overlaying the artists creation.

By analogy, it is like wearing rose tinted glasses in to an art museum. The first couple of painting look REALLY PRETTY, but after while one notices the same coloration over everything from medeval alter pieces to Rembrandts to Otto Dix to Warhol and you realize something is very wrong. Not only should some of these look radically different from each other, but some are NOT supossed to be pretty at all! You might as well be viewing a collection Elvis-on-black-velvet portraits (or at least tourist-trap seascapes).

Thus for me "nice sounding" amps in the sense of colored with harmonics ultimately turn everything into elevator music.

Yeech.:gnasher: :headbash: :grumpy:
 
there are cogent arguments that each harmonic should
be weighted by n^2/n, over simply summing THD.

Class aB amplifiers produce liberal harmonics to at least n=20,
exacerbated by the fact these harmonics are effectively
differentiated by the effects of falling feedback in single pole
compensation.

It is entirely true that two amplifiers with the same THD can
sound different due to harmonic spectra, and that a amplifier
with a given THD can sound better than another with lower THD.

I am not of the school of thought that says some 2nd harmonic
distortion is good, I am of the school of thought that would
ignore 2nd harmonic distortion (and THD) if it meant higher
harmonics could be reduced.

I'll also point out that generally even harmonics (2,4,6,8 etc)
are reduced by symmetry whilst odd order (3,5,7 etc) cannot
be by topology.

I'm generally of the opinion that fully symmetrical amplifiers are
missing the point somewhat, and that fully symmetrical is not
desirable per se as a design goal if single ended is better.

And that for frequency dependant distortion, which applies to
an input stage, any distortion should be avoided if possible.

And as a final caveat, amplifiers distortion figures at high power
are not directly related to those at around 1W to 5W, which
most of the time is what counts.

edit : just a final note - tube amps produce low levels of high
harmonics and the spectra is not distorted so much because
of low levels of overall feedback, they do not sound better
than transistor amplifiers because second harmonic is higher.

:) sreten.
 
Popping in here... If you are interested in hearing how different harmonics *sound*, you could try this little singing synthesiser. It has a lot of things that are not relevant for this topic, but just set flutter and vibrato amplitude to 0 and uncheck all formants. Now you can combine partials by dragging on the partial graph.

http://www.speech.kth.se/music/downloads/smptool/madde.exe

Not that it is the same as distortion, but it is fun...
 
In some thread, I read that an audio power amp with excellent measurement figures do not sound so good. Also in some thread, it is said that audio power amp that is too "clean" produces sterile sound.
Disagree. There is a bad premise in these all-too-common statements. It is that the measurements are comprehensive. Which they aren't.

I want to hear the instruments and their harmonics as recorded. I am confident that they are perfectly fertile. So an amp that changes nothing about them is the one for me. I don't want any modification or extra harmonics thrown in, nor should it be necessary to do so. These "sterile" amps are modifying the original music content: it's not that they are too clean, rather it is that they are too dirty.
 
sam9

"And as a final caveat, amplifiers distortion figures at high power
are not directly related to those at around 1W to 5W, which
most of the time is what counts."

To expand on the above, single point distortion figures are pretty useless. Even a boom box gets good figures at 1W at 1kHz! If I were king, distortion data would alway be presented as a 20-20kHz plot at 1W, at 3db below clipping, at rated power and at 50% of rated power. Or at least something along those lines. I would also dictate that graphical presentation have some standards about scaling the axes. Oh yes, do all of the above for both nominal 8 ohm and 4 ohm loads. The chance of this happening is virtually zero, of course.

The reasons I would like the above scheme are:
A- some designs may be ok but unstecpacular at 1kHz but have much better results at high freqencies

B- different designs figures change at different rates as rated power or clipping is approached

C- "Rated power" is ambigous when one maker calculates is from the rail voltage while another calculates at peak voltage before near-clipping distortion sets in.

D- Ones particular circumstances (speakers, room size, favorite source material, and even age based hearing loss) could be better addressed with more complete information which would allow one to make the trade-offs ther best suit ones needs.
 
Hi,

My 2 cents:

http://www.zainea.com/multidimensionalaudio.htm

Note that this was written 35 years ago. I love this one:

"Before the days of so-called Hi-fi systems there were no audio measurements, there were only natural sounds with a perfect signal to noise ratio, unlimited power handling capability, no distortion of any kind, but the number of people with a possibility of ever listening to music was rather limited. Then came Hi-Fi. The invention of the phonograph record ruined high fidelity, but made music universally available."

:)
 
lumanauw said:

What harmonic is really add pleasant sound? What is really the meaning of odd harmonic, even harmonic, second harmonnic etc? If differential is canceling even harmonic, is there a circuit that cancels odd harmonic? It seems even harmonic is the one that makes good sound.

There is already quite enough posts and information about the technical stuff which is very valid. If you listen to music instead of pure sine waves, here's a slight twist.

Music Theory 101...

Music is made up of notes. This includes voice; Acapella singers generally try to vocalize notes. For simplicity lets just take a simple instrumental piece, piano or classical guitar. It would be boring if music was played using single notes at a time. This begs the question: what also sounds "good" with a particular note? Well music theorists have answered this. "good" in music is a term called: consonant. The most consonant note to go with a note is obviously the same note. This is known as octaves of that note. Octaves in frequency terms are doubling or halving of the frequency. This means harmonics, specifically the 2nd, 4th, 8th, 16th, etc... notice all even. Well that's good, but what else sounds good with a note. Ask any trained musician and you'll find out that the next most consonant note is something called a 5th. 5ths in frequency terms are 1.5x a fundamental. Wait a minute, that means it's the 3rd harmonic. It's odd. Are there other odd harmonics that are consonant? Actually yes, for example, the 5th harmonic is known as a major 3rd. It would be the next most consonant sound.

There is a pattern. Low order harmonics are more consonant, they become more dissonant as they get to higher order.

Now, to answer your question: Lets take a standard chord from music. The notes in that chord are these: A fundamental, an octave, a fifth, and a major 3rd. So you see, if this were what was played, the 2nd, 3rd, 4th, and 5th harmonics would all be enhancing the sound. Actually since the 6th harmonic is actually a minor third, I would even argue that the 5th harmonic would be more "musical" than the 6th although the opposite would be true for a minor chord.
--
Danny
 
To me harmonics are not good, but hundreds or thousands of times as much second harmonic distortion can be tolerated as being musically acceptable as high order harmonic distortion which is more characteristic of solid state multistage amplifiers.

There is also intertransient performance which is not related to harmonic distortion in any direct sense. You can put a Z5U ceramic capacitor in an audio circuit as a coupling cap with measured audio distortion of 0.01% or less and it will greatly deteriorate the sound with much detail, soundstaging and musicality lost partly because it doesn't store and release musical (transient) information linearly (at the proper time) due to its high K nature. However, in a HD test, this behavior is swamped by the constant signal and barely shows up at all.
 
Re: Re: What harmonic does to sound?

While I do find this type of reasoning very interesting, there are
a few things that seem worth pointing out or that evenseems plain
wrong (disclaimer, it is late, so maybe my brain is playing tricks).

azira said:
...
Octaves in frequency terms are doubling or halving of the frequency. This means harmonics, specifically the 2nd, 4th, 8th, 16th, etc... notice all even.

Don't forget however that we are likely to also get even harmonics
that do not fit into the pattern, the 6th, the 10th etc.


5ths in frequency terms are 1.5x a fundamental. Wait a minute, that means it's the 3rd harmonic. It's odd.

No, the 3rd harmonic is 3x the fundamental. you are not likely
to get 1.5x the fundamental out of an amplifier except possibly
as IM. BTW, I am not so sure the the 5th is exacly 1.5x the fundamental either, but i am too tired to calculate it right now.


There is a pattern. Low order harmonics are more consonant, they become more dissonant as they get to higher order.


Somewhat oversimplified, but yes there seems to be some truth
to this. Still, no harmonics att all would reasonably better than
additional consonant harmonics, especiall if the relative amplitude
of them deviates from the spectrum of real instruments.
 
Re: Re: Re: What harmonic does to sound?

Christer said:

Don't forget however that we are likely to also get even harmonics
that do not fit into the pattern, the 6th, the 10th etc.

Those harmonics are not octaves, they are other notes.

Christer said:



No, the 3rd harmonic is 3x the fundamental. you are not likely
to get 1.5x the fundamental out of an amplifier except possibly
as IM. BTW, I am not so sure the the 5th is exacly 1.5x the fundamental either, but i am too tired to calculate it right now.


You are correct, the 3rd is 3x the fundamental. However, it is the same note as a 5th of the first octave which would be 1.5x (ie 3/2).

Christer said:

Still, no harmonics att all would reasonably better than
additional consonant harmonics, especiall if the relative amplitude
of them deviates from the spectrum of real instruments.

I agree 100%. I do not nearly intend to mean that harmonics are good. I only intend to challenge the even vs odd harmonic prejudice. IMO harmonic distortion issues should be seperated by high vs low order not even vs odd.
 
There's an interesting example of what harmonic does to human perception. An electric guitar player (mostly pop-rock), who do not know anything about electronics, likes to hear "intended distorted sound" from his guitar. There are lot of this "distortion generator" cct for guitars. I certainly likes the sound of Santana's guitar and it must be distorted purposely.
Is it our hearing more pleasant with a certain kind of distortion?
 
Re: Re: Re: Re: What harmonic does to sound?

azira said:

Those harmonics are not octaves, they are other notes.

Agree. Maybe I read too much into that sentence I quoted. So then
the 6th harmonic would be two octaves plus a 5th, right? That
means it should have a similar character as the 3rd harmonic
which is odd. This seems to support the argument that an
even/odd distinction may not be all that relevant.

BTW, I just made the calculation of the 5th, and it seems it is
only about 0.1% off from x1.5, and the same goes for the ocatve
plus the fifth which also about 0.1% off from the 3rd harmonic.
I wonder how sensitive the ear is to such minor deviations?
 
One thing we must always keep in mind: The very same mechanisms (i.e. nonlinearities) that are responsible for causing harmonic distortion do also generate IMD which is much less natural than harmonic distortion.

An instrument by itself may very well generate "unmusical" overtones (i.e. 7th, 9th etc). So that alone cannot be a large problem as long as it isn't too much.

Some overtones may not even be INTEGER multiples of the fundamental (i.e. they are a little off). If you now add integer ones then this might possibly cause problems as well. Together with the intermodulation products this my lead to a "carpet" of signals that were not present originally.

Regards

Charles
 
"I certainly likes the sound of Santana's guitar and it must be distorted purposely."

Yes that is true but thats the recorded guitar with the distortion purposfully added to improve or to get the sound the musician intended. So surely the same holds true, you want to produce Santana's added distortion as cleanly as possible, clean distortion ay eurgh!
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.