What is “warmth” and how do you get it?

02-17-11: Learsfool
The only quibble I still have with this discussion is with the importance many of you are placing on minute changes of amplitude in overtones within a musician's timbre as being the major factor in a perception of "warmth." (I am not denying that this is a factor)...Harmonics do of course have to do with the "warmth" of the timbre, but the MUSICIAN has no control whatsoever over specific harmonics within the overall timbre. [emphasis added]

Hi Learsfool - I may be wrong, but I think the importance many folks have given to harmonics in the perception of warmth has less to do with how harmonics are produced by a MUSICIAN and more to do with how they are reproduced by the PLAYBACK SYSTEM. In other words...

Some systems merely PRESERVE whatever warmth exists on the recording. Other systems seem to ADD warmth, whether it exists on the recording or not. For these systems, a common belief is that the added warmth is often a consequence of added HARMONICS.

Do you not think that added harmonics, as you might get with a tube amp, are a significant contributor to the perception of warmth during playback?

Bryon

Hi Bryon - perhaps Al can weigh in on this and correct me if I am wrong, but I'm pretty sure harmonics couldn't truly be "added," (they are of course all already present in the timbre) though digital reverb would be an example of an "addition" that would certainly affect the perception of warmth. Overtones can be and are sometimes removed by digital processing.

Perhaps a better word to describe what you are speaking of would be "emphasizing." Atmasphere has posted quite informatively about these types of issues, talking of even or odd order harmonics being emphasized by the design in different types of amps, and whether or not the designer is thereby following the "rules" of human hearing. This is very similar to the way the acoustical environment affects the timbre, except the designer of a piece of audio equipment I suspect has alot more control over his end result than an acoustician does.

And yes, I would say that tube designs are certainly "warmer," speaking very generally, than solid state designs, therefore sounding more lifelike. I'm just saying that there is a whole lot more to do with that than amplitudes of individual harmonics within the overall timbre - again, these harmonics are inaudible to far more than 99.9% of us.

Learsfool, as I see it an audio system can introduce harmonics, enhance harmonics, or even reduce harmonics that may be present in the source material.

Harmonics can, and to some degree inevitably will, be introduced by the system in the form of distortion products.

They can be enhanced either by virtue of a frequency response emphasis that happens to occur at a frequency corresponding to some harmonic (multiple) of the fundamental frequency of a note, or by virtue of a frequency response dip that happens to occur at the fundamental frequency, or by virtue of distortion of the fundamental frequency of a note, the distortion products therefore occurring at the same frequencies as harmonics that may be present in the note.

They can be reduced by the converse of those frequency response effects, or by introduction of a distortion product that is out of phase with a harmonic that may be present in the note at the same frequency.

I'm just saying that there is a whole lot more to do with that than amplitudes of individual harmonics within the overall timbre - again, these harmonics are inaudible to far more than 99.9% of us.

I respectfully disagree. My understanding is that timbre and the relative amplitudes of individual harmonics are one and the same.

As I understand it, to cite an example, a violin playing a note whose fundamental frequency is say 1 kHz will produce very audible harmonics at 2 kHz, 3 kHz, and other higher multiples of 1 kHz. A flute playing a note whose fundamental frequency is also 1 kHz will produce very audible harmonics at those same multiples of 1 kHz. The reason that the note produced by the flute will sound different than the note produced by the violin is that the relative amplitudes of those harmonics will be in different proportions.

And, similarly, differences in timbre and tone between two different playings of the same note on the same instrument will be the result of differences in the relative amplitudes of those harmonics, as I understand it.

Best regards,
-- Al

Hi Al - thanks for weighing in on this. However, I think you are incorrect when you say "Harmonics can, and to some degree inevitably will, be INTRODUCED by the system in the form of distortion products." (My emphasis) Your own examples that follow are all examples of what I was speaking of in my previous post - what you call enhancing certain harmonics via distortions (and they are good examples). However, ALL natural harmonics are always present in the natural timbre, so you can't introduce a new overtone that wasn't there before, though you can distort (or even remove) it. This is what I was trying to say in my previous post. If this statement is indeed incorrect, please explain.

One other point - in your violin examples, yes, those overtones are of course part of what make differences in timbre. However, each individual one is indeed indistinguishable from the others to the ears of at least 99.9% of humans. It is not possible to tell which of those overtones are the ones that are different, in your example of two different playings of the same note on the same instrument. If I played the same note twice, at the same volume, on my horn, you would not be able to tell me which individual overtones were affected and how, and this is doing you the credit that you would be able to hear the difference in the timbre between the two at all - a great many audiophiles would not, especially if I tried to the best of my ability to make them exactly the same. And in the same case, it would have to be a VERY bad recording/system indeed that would distort them so much so that most people could hear it. These sorts of differences are MUCH more audible live and at very close range than they are on a recording.

Learsfool, I think you are absolutely correct. I could go on ad nauseum, but suffice it to say that I see only one term when talking about electric signals, 'harmonic distortion', with the word 'distortion' being a noun and 'harmonic' being an adjective modifying it. I am unaware of any naturally occuring 'harmonics' in an electrical signal. Only distortion of what ever type.

FWIW.