• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Do we crave distortion?

Better say canceled then?
I think I caught one or two references in TI and AKM papers about it but ESS lets you mess with it (adding or whatever) .

I'll dig further and report!

Here's an example created in DISTORT showing the transfer function curve producing some large-level harmonics. Imagine that this is what some device (say a bad tube amp) measures like:

1701277537597.png



Now, imagine that the inverse of that transfer function (flipped around x=y line) is applied to the incoming signal. That would completely negate the effect of the non-linear transfer function of the device, and produce this, pure original signal:
1701277621768.png


This is the effect of the non-linear THD compensation -- it straightens out the transfer function to make it as close as possible to x=y line. The correction is not as straightforward with real devices, since most devices don't have a static transfer function. Some have dependencies on frequency, some on level, and others on both. These become much more complicated to compensate properly.
 
Curious. When you say "bad tube amp" are you referring to a SET distortion profile?
(Often more expensive for sure but unsure why.)
 
Curious. When you say "bad tube amp" are you referring to a SET distortion profile?
(Often more expensive for sure but unsure why.)
It wasn't any specific amp, just a very large non-linearity causing a very high-level of THD/IMD. Certainly some part of the audiophile industry appears to think that there's something good about such a high-level of distortion, that, as @fpitas pointed out, is often very expensive.

Examples:
index.php



index.php



index.php


index.php


index.php
 
Thanks.

Currently using my VTL-ST85 pentode push pull (with moderate distortion).

It's proper cold here in Edinburgh. My living room has high ceilings and is a nightmare to heat. Consequently I am wearing down and rarely heat this room. Due to the recent energy price hikes, the VTL and Starbucks below me are my sole heat sources. All good.
VTL = 140W average.
 
Very simple proof: just listen to a pure sinewave compared to a piano note (of the same frequency). "Surprise": ~everyone prefers the piano note and hates the sinewave.
One of the major extras that make the piano note sound much more enjoyable is, suprise again, Harmonic Distortion: i.e. the piano note contains 10+ HDs which contribute to its pleasant timbre (and make it sound different from the same note played on guitar, trumpet etc...). A nice video explanation.
That is not distortion - or is a silly definition of distortion:

Distortion is a change to the waveform. Think of *any* acoustic instrument, or the human voice.

The waveform comes from those instruments fully formed with the fundamental, and all relevant harmonics to create the tone and timbre of that instrument. The fundamental has never existed standalone in order to be distorted. The instrument *cannot* create a sine wave, and it certainly doesn't create a sine wave and then distort it.

Distortion then is anything done to that original waveform that changes it. In terms of reproduction gear (which we mainly discuss here) that comes from uncontrolled non linearities in the process. It results in a reduction in the accuracy of the reprocution of an instrument's (or voice's) timbre.
 
Last edited:
Presumably this means that changing EQ is also adding distortion then. Or no?
 
Presumably this means that changing EQ is adding distortion then.
Not in the conventional sense, at least. Distortion means creating new frequencies.
 
Just changing the amplitude of frequencies surely. Or no.
 
Just changing the amplitude of frequencies surely. Or no.
No, not unless new frequencies appear. That's a sign of non-linearity. Besides I'd argue that EQ done by a sane person flattens the at-ear response.
 
Just changing the amplitude of frequencies surely. Or no.

When we talk about distortion we conventionally mean the non linear processes that create additional harmonic or non harmonic freqeuncies.

EQ changes the frequency response - and only changes existing frequencies - and is a linear process. We talk about that as a FR change rather than a distortion (though the wavform in the time domain is actually changed).

The two types of change sound very different (when audible), and the mechanisims to create them are very different. Futher - EQ is generally a desired change (it is pretty trivial to get flat frequency response in electronics these days) so it makes sense to treat them separately.
 
Thanks. Indeed. I guess distortion is actually a fairly vague term. Like energy.

Anyhoo. I definitely prefer a flat response via MathAudio RoomEQ for either my low-fi valve amp or KH310s. Got several curves/presets for each and easy to switch.
 
Last edited:
That is not distortion - or is a silly definition of distortion:

Distortion is a change to the waveform. Think of *any* acoustic instrument, or the human voice.

The waveform comes from those instruments fully formed with the fundamental, and all relevant harmonics to create the tone and timbre of that instrument.

That is just an oppinion/point-of-view. It's somewhat fair, but I do not share it much.

The math does not share it either: a harmonic signal added by an instrument or by any other means is exactly the same thing. There is zero math-difference between 'artificial' amp-HDs and 'natural' instrument-HDs, so it's mathematically improper to call/treat them differently.

Musical science/theory does not share it either: a harmonic is defined as a multiple of the original frequency. That's all, it doesn't matter who/where/how generated it.

The human ears do not share it either: whatever way a harmonic signal was added, it has the exact same effects on your ears/brain/perception.
Simple proof: a modern/good synth generates all the piano HDs 'artificially' and sounds exactly the same as the acoustic piano (which generates the exact same harmonics 'naturally').

The fundamental has never existed standalone in order to be distorted.
The instrument *cannot* create a sine wave, and it certainly doesn't create a sine wave and then distort it.
That may be true for acoustical instruments (btw, do you have a citation/paper?)
But an electronic synth does exactly that: it does "create a sine wave and then distort it". And it sounds the same as the 'real' piano.
Even if true (for acustical instruments), your argument makes zero difference for the human ears/brains (also none for the physical air-waves).

Distortion then is anything done to that original waveform that changes it. In terms of reproduction gear (which we mainly discuss here) that comes from uncontrolled non linearities in the process. It results in a reduction in the accuracy of the reprocution of an instrument's (or voice's) timbre.
You call the HD "reduction in the accuracy" and you have a fair (engineer's) point of view.
Some people may call it "more musical" and they have a fair (listener/musician's) point of view.

Maybe you should stop calling other people's point-of-views "silly".
Especially when theirs seem to be well supported by math, musical science and human physiology. And yours is supported by ... well, not sure what besides the EE-book definition of 'distortion'.
 
Distortion is a waveform that has undergone changes.

An instrument simply has attack, sustain and decay and different harmonics during all of them.

These are definitely not the same thing. In both cases harmonics are involved.

An electronic instrument does not distort the fundamentals. In case of an electronic organ (or synth) harmonics are added. Of course one can also use non linear devices on purpose but that is part of the creation process in order to create a certain sound.

After that the result should be reproduced faithfully adding more effects (regardless what they are, linear or non-linear) is adding distortion.
Room correction, of course, is also adding (linear) distortion but with the goal being less (linear) distorted reproduction of the recording.
 
The math does not share it either: a harmonic signal added by an instrument or by any other means is exactly the same thing. There is zero math-difference between 'artificial' amp-HDs and 'natural' instrument-HDs, so it's mathematically improper to call/treat them differently.
There is quite a bit of math difference between a musical instrument producing harmonics as part of note generation and an audio device doing the same as distortion. One difference, and the one that can really matter, is that the audio device distortion adds not just extra harmonics but also IMD. IMD is very much in-harmonic with the main tone, and may be much more audibly objectionable at higher levels than HD. Of course, IMD also very much depends on what other notes are being played at the same time and will vary significantly from second to second.

And yours is supported by ... well, not sure what besides the EE-book definition of 'distortion'.
It is the definition of distortion that's used in audio. Redefining standard terms to the "lay understanding" of what they might mean leads to a serious break-up in communications. Unless you start defining each term you use in each of your posts, nobody would ever know what you mean.
 
The math does not share it either: a harmonic signal added by an instrument or by any other means is exactly the same thing. There is zero math-difference between 'artificial' amp-HDs and 'natural' instrument-HDs, so it's mathematically improper to call/treat them differently.
That is not correct, either. The harmonics from an instrument have a definite phase relationship with their fundamental. A distorting amp might or might not create that relationship.
 
The math does not share it either: a harmonic signal added by an instrument or by any other means is exactly the same thing. There is zero math-difference between 'artificial' amp-HDs and 'natural' instrument-HDs, so it's mathematically improper to call/treat them differently.
Structural vibrations are only "harmonic" in 1-D, which is merely an approximation for strings (and woodwinds too). For strings, the key assumptions are infinitely thin and infinitesimal amplitudes.

Vibrations from non-1-D structures are inharmonic, e.g. that of a drum. The math for structural vibration is completely different from the nonlinearities in electronics.
 
Back
Top Bottom