Professor David Keeports: The warm, rich sound of valveguitar amplifiers [2017]

[worker24x7]

Accidentally came across this paper, explaining that even harmonics (when frequency is multiplied by an even number) are much more pleasant to human ear than odd harmonics (when frequency is multiplied by an odd number).

Tube amps (as explained in details) tend to produce even harmonics (at certain amplification levels), while this is reversed (and therefore less pleasant) for the solid amps.

That's why, we want solid state amps (this is my conclusion) to be non-coloring, while we want tube amps (at least guitar amps, the subject of the article) -- to color sound with even harmonics.

I searched here before posting, to see if this paper was already discussed/analyzed (it was published in 2017). Could not find, therefore posting (apologies in advance if this is a duplicate).

--
David Keeports Department of Chemistry and Physics, Mills College, Oakland, CA
Phys. Educ. 52 (2017) 025010 (7pp)

But why are even harmonics generally preferable to odd harmonics? ... overdriving a guitar amplifier adds harmonics to each of the harmonics shown. Some music theory is in order here. Consider a Cmaj7 chord constructed from pure C, E, G, and B tones, which are the 1st, 3rd, 5th, and 7th notes of a C major diatonic scale. Additionally, assume that the diatonic scale is a so-called ‘just’ scale. This means that the frequencies of the 3rd and 5th notes of the scale are the 1st note’s frequency multiplied respectively by 5/4 and 3/2 [6]. ... For simplicity, assume that a tone’s strongest even harmonic is its 2nd harmonic and its strongest odd harmonic is its 3rd harmonic. Doubling the frequency of any tone produces a new tone one octave higher than the original tone. Tones an octave apart don’t clash with each other. Instead, such tones add warmth and pleasant complexity to a sound [my emphasis]

...

Now add 3rd harmonics to the notes of the original Cmaj7 chord. Tripling the frequency of any note produces a new note one octave and a5th higher than the original. For example, the 3rd harmonic of a C note is a G note more than an octave above the original C. Adding 3rd harmonics to each note of the original chord produces new G, B, D, and F# notes. While the first three new notes, each found in a C major diatonic scale, enrich the original sound, the F# that is not in this scale adds dissonance. Now imagine the effect of adding unwanted tones to each of the frequencies in the guitar chord in figure [6] !

In brief, second harmonics make a sound richer and fuller, but third harmonics can create dissonance. [my emphasis]

...

Tests conveniently performed in an introductory physics laboratory support the frequently made claim that overdriven valve amps preferentially add even harmonics to sounds while overdriven transistor amps preferentially add odd harmonics. [my emphasis] The analysis of the function of a triode amplifier valve predicts the effect of increasing the strength of the electric signal input from a guitar’s pickups upon output wave shape. Wave shape then reveals harmonic content.

The asymmetry that results from moderately overdriving a valve indicates the presence of even harmonics, while the symmetry resulting from very strongly overdriving a valve shows the presence of only odd harmonics.

At low input signal strength, a sine wave is linearly amplified without harmonic enrichment. As the strength of the input signal increases, a valve first adds even harmonics and then adds odd harmonics.

The art of amplifier design combines musical esthetics with physical principles.

Moderately overdriving a valve amplifier produces the warm, complex, and highly listenable sound that electric guitar players enthusiastically seek.

--
unfortunately, Professor Keeports has passed away in 2018

 

[mhardy6647]

oh, lordy...

So, just to be clear: the "even order harmonics" isn't a tube thing -- it's a single-ended thing. Push-pull amplifiers (be they operating at Class A, AB, or B) will have low levels of even-order distortion regardless of vacuum tubes vs. transistors as the active amplifier components. The two 'legs' of a push-pull amplifier 'cancel out' even order harmonics.

It is true (sometimes!) that typically the spectrum (i.e., relative levels of various harmonics) of harmonic distortion components is a bit different between, a push-pull solid state (power) amplifier vs. a single-ended (power) amplifier.

Mind you, I am a single-ended vacuum tube power amplifier kind of guy.
Direct-heated power triodes, at that!

Doesn't mean I'm stupid -- just a bit idiosyncratic.

 

[Barrelhouse Solly]

I've played a fair amount of amplified harmonica. What most players, me included, want is distortion that produces a particular timbre, similar to Little Walter or Big Walter. Harp players often substitute preamp tubes. Higher output tubes like the standard 12AX7 give you a different sound from the lower output tubes in the 12A family. Another feature harp and guitar players often want is separate preamp and power amp volume controls. The distortion from the two segments is different. You can get an amazing variety of sounds without using pedals by tube rolling and fooling around with the gain. No criticism of pedals. They can do wonderful things.

 

[mhardy6647]

I wouldn't admit this just anywhere, but -- I didn't previously know that there was a Big Walter!
I will say that the amplifier pictured above does reproduce harmonica very nicely.

 

[MattHooper]

So, just to be clear: the "even order harmonics" isn't a tube thing -- it's a single-ended thing. Push-pull amplifiers (be they operating at Class A, AB, or B) will have low levels of even-order distortion regardless of vacuum tubes vs. transistors as the active amplifier components. The two 'legs' of a push-pull amplifier 'cancel out' even order harmonics.

Out of curiosity (as I do not have the technical know how to determine this)… how does the following fit in with what you wrote?

I use Conrad Johnson premier 12 tube mono blocks. I’ve always felt that the way they sound different from a solid state amplifier in my system seemed to be a little more than only frequency response deviations. Like there’s a slight added thickening texture. Like on the edges of the sound… transients… high frequency sounds, have a slightly thickened texture.

But I don’t know what causes it, and I always wondered whether it was possibly second order harmonics… or some other distortions ,

I see in the Premier 12 Stereophile measurements. This is stated:

The 1kHz THD+noise waveform at 2W out into 4 ohms, shown in fig.4, is heavily second-harmonic. The same is true at other loads (not shown).

And

The distortion spectrum resulting from a 50Hz input at 93W into 4 ohms (2/3 rated power) is shown in fig.5. The largest artifact is the second harmonic, at -38dB (about 1.2%)

I also use a Conrad Johnson premier 16LS2 tube pre-amplifier. The stereophile measurements for the Premier 16LS preamplifier mentions:

Above 1V, the reading is dominated by true distortion. The oscilloscope display indicated that this was almost pure second harmonic, resulting from the parallel triodes' slightly "bent" transfer function.

It is true (sometimes!) that typically the spectrum (i.e., relative levels of various harmonics) of harmonic distortion components is a bit different between, a push-pull solid state (power) amplifier vs. a single-ended (power) amplifier.

Mind you, I am a single-ended vacuum tube power amplifier kind of guy.
Direct-heated power triodes, at that!

I haven’t listened to a lot of single-ended amplifiers, but my brief experience leads me to empathize with your preference.

I heard a pair of Totem loudspeakers at my friend’s place transformed from “meh… sounds pretty good” to “ oh my God this is heaven I could listen all day” when powered by a Fezz Audio 300B single ended design.

 

[voodooless]

You know this is about guitar amps, right? And a guitar is an instrument, and the amp is an extension to it. Whatever the amp does, it’s up to the performers to determine if it is the right thing or not.

This has zero to do with HiFi.

 

[mhardy6647]

You know this is about guitar amps, right? And a guitar is an instrument, and the amp is an extension to it. Whatever the amp does, it’s up to the performers to determine if it is the right thing or not.

This has zero to do with HiFi.

Well, the amplifier topologies and active circuit components employed cannot generally tell if the signal being fed through them is live or Memorex.

 

[mhardy6647]

Out of curiosity (as I do not have the technical know how to determine this)… how does the following fit in with what you wrote?

I use Conrad Johnson premier 12 tube mono blocks. I’ve always felt that the way they sound different from a solid state amplifier in my system seemed to be a little more than only frequency response deviations. Like there’s a slight added thickening texture. Like on the edges of the sound… transients… high frequency sounds, have a slightly thickened texture.

But I don’t know what causes it, and I always wondered whether it was possibly second order harmonics… or some other distortions ,

I mean... that suggests the conrad-johnson amplifier was deliberately designed or adjusted to operate incorrectly (i.e., as a creator of waveforms vs. a reproducer of waveforms & vis-a-vis @voodooless's earlier MI vs. hifi amplifier comment).

This being said, 1) I would not expect such of c-j and 2) the c-j amps of old were notoriously not tube-y sounding!

some random googling:

The Valve Wizard -Push-Pull

Distortion and Feedback

Distortion and feedback - is feedback as bad as some people claim?

sound-au.com

 

[Redacted]

This is why I like my valve guitar amp.
Bear in mind when running clean on a guitar amp, the even order harmonics are at a very low level - the whole point is to drive the thing to break-up or crunch, and there is the full gamut of even and odd harmonics if you do.

Modern players are all going to "FRFR" (full range flat response & clean) amps, all class D, and using modellers (inc. other effects pedal functions) to dial-in valve amp and speaker sounds. Many guitarists then go on to discover that it then becomes all about the speakers! So full circle back to basic ASR stuff!

 

[EERecordist]

The paper author is not a reliable source on the topic. The paper references are not deep. A physics appointment at Mills is not top academic appointment.

There are many papers on the perception of distortion and software adding controlled distortion to try your own hearing on your own material.

 

[voodooless]

Well, the amplifier topologies and active circuit components employed cannot generally tell if the signal being fed through them is live or Memorex.

The amp doesn’t have to know, though…

 

[Redacted]

The amp doesn’t have to know, though…

What? You mean to say an amp cannot lookup an online database to learn about all possible signals in advance? In this day and age?

 

[mhardy6647]

The amp doesn’t have to know, though…

I think we said the same thing.


AC is AC -- even if the AC is

 

[MattHooper]

Thanks for the reply.

This being said, 1) I would not expect such of c-j and 2) the c-j amps of old were notoriously not tube-y sounding!

Yes, the word is that CJ have tube amps became more neutral and less lush and tubey over time. At one point, I was intrigued in picking up one of their newer amplifiers but ultimately, I decided for me the main point of a tube amplifier sonically is to get that extra tubey quality. So I’m happy sticking with their old amplifiers - I don’t want solid state sound from a tube amplifier.

If I want solid state, and occasionally it’s fun just for a change, I’ll throw an actual solid state amplifier in the system.

 

[AJM1981]

Get the impression that the solid state vs tubes is a bit a ying/yang thing. One can reason about what the microphone picks up should be key and so should specs be leading. On the other hand is the perception of what is less harsh to someones ears not wrong either.

Also this whole balance is in debate with the type of music one listens to and what gives the illusion of more realism. I consider warm sounding gear excellent for Jazz, classic rock singer songwriter etc. I have not heard a better display of Hendrix his recordings than with a warmer amplifier I own. One that will miss crisp details (not that present either) , but the illusion of having a stage is key. When I listened to the same recording with a Nad amplifier I heard people in the crowd say and shout things really specifically that I did not hear before. But it took me off of what was going on behind. Maybe that is what the saying that it is more about the music is about.

Though this is a double edged sword. When listening to classical music with loads of strings, minimalistic pieces with lots of things going on in interaction or new wave pop, edm, metal etc I prefer the colder sounding "faster" solid state amps and the same saying applies.

There is not one single amp that does everything as it is best experienced (in my opinion that is).