Why? The phenomena is simple in theory but complex to explain fully. I'll make
an attempt here to explain. We are talking about the difference between vacuum
tubes and solid-state devices, e.g. bipolar transistors and semiconductor IC's.
Note that I did not include FET devices, which are a form of solid-state device - I'll
cover these later. The answer is rooted in the harmonic content that the device
can pass or generate. What's known is the human ear finds even harmonics pleasing
and odd harmonics annoying or grating, or in the least unmusical. First a word about
harmonics. Hamonics, for those who are not versed in wave theory, are simply frequencies
which are related to the fundamental freqency being reproduced. As an example, a 100Hz wave
may generate harmonics out to the nth order. How strong and how many of these harmonics
are present will determine the harmonic content, or harmonic distortion (THD) as it is
haracterized (when deemed undesirable). An example: a square wave has
100% harmonic distortion, the fundamental plus all its harmonics are present in equal amounts.
In a square wave (e.g. digital) harmonics are very desirable. In fact the more the better.
In the analogue world and musically speaking, the inverse is true.

What are even and odd harmonics? It just refers to their respective mathmatical realtionship
to the fundamental. Using the same 100Hz, the 2nd, 4th & 6th harmonics (even) would be
200, 400 and 600Hz, the 3rd, 5th and 7th harmonics (odd) would be 300, 500 and 700Hz, etc.
In low distortion equipment (i.e. audio gear) these harmonics will decrease exponentially
as the order goes up and the faster the better, resulting in lower THD

Tubes: there are a couple of important characteristics that separate tubes and bipolar
transistors. First, by nature vacuum tubes are not very fast elements. That's a good thing
for music lovers. Being relatively slow, the tube cannot generate huge amounts of harmonics
even if it wanted to so they act as a kind of intrinsic filter to our benefit. Second, tubes favour even
harmonics and are less inclined toward odd ones. In fact tubes a rich in second order harmonics
and by the time the 3rd shows up, it is greatly reduced by the tubes natural characteristic.
This is what is attributed to a tubes "warmth" . And lastly tubes are voltage based devices,
they amplify voltage. That I'll save for another time.

Bipolar transistors and IC's are very fast, making them ideal digital devices. They operate in
two modes: switch mode, e.g. ON or OFF for making square waves or pulses for the digital world,
and linear mode, meaning what comes out is proportional to the signal that is put in and is how
they are used in amplifiers (power and preamps). They can and will generate harmonics waaaay out there.
They are also rich in odd harmonics which results in the unpleasing or harsh, edgy quality often
attributed to them. These devices are, in linear mode that is, as they are used in amplifiers
current amplifiers. Side note; The first IC's and transistor based opamps that came into being
around 1969/1970 sounded awful - very thin and "card boardy". I remember the first recording
console delivered to the studio a very famous one back then] where I was working at the time and
had these opamps in it. It sounded really bad. This was at a point in time where the world was
changing from discreet transistors to opamps (amplifiers on a chip) in the professional audio
equipment world. Also one inch 8 track tape recorders (not the consumer cassette machines) were a big deal too. Anyway I digress.

Finally, earlier I mentioned FET's or field effect transistors. Like tubes they are voltage amplifiers and generate predominately even harmonics, making them a very good substitute for the vacuum tube with the same pleasing mucical sound. That is why you see some amp manufactures making a big deal of MOS-FET and other similar devices in their products. Hope that helps and makes some sense.

It could be just rounded down to what I've always believed - solid state devices switch faster than valves therefore making the output harsher. That is my theory anyway.

Workwise, it has been a busy year for me, leaving me little or no time to build and enjoy the Currawong amplifier. That may get done over Xmas though and I am looking forward to it.

My ears like the 2nd harmonic around 45dB down. Much higher you risk getting noticeable intermodulation products, sounds like mud. I've used cathode followers or preamp audio voltage amps in CD players to get warmer sound from CDs.

See http://pw2.netcom.com/~wa2ise/radios/tubedac.htm

So this sort of intertwines with with another post I had about why we would rather listen to our old valve radios its a mixture of the technology and our childhood memories.

Hi, Yes I’ve heard that one before, so you have to go make a four channel solid state and valve amplifier for full stereo.
I only buy that in the context of musical instruments (the quality of a piano playing beneath the static of AM radio).

Riddle me this:
Most people here would say an AM valve radio sounds better than a new one,
but the radio station could have valve final transmitters at best, and an automated line up of digital audio as the source.
Most people at my radio club would agree that nothing beats valve audio over the air for voice.
A transmitter with valve audio also has valve final amplifiers,
but it matters not if the receiver is a valve radio or solid state receiver.

So it doesn’t matter where you throw the valves in the system to somewhat fix the audio.
You can even play an iPhone through a Leak and the audio is fixed.

My theory is this: "Like tubes they are voltage amplifiers” and not what is said after it.
The proper application of high voltage. There are high voltage FET transmitters that have a similar quality about them.

edit.. here’s one, it’s a solid state receiver, and you can play it on a computer through Youtube, but I’d still know there’s a valve amplifier on the end of a solid state transmitter, or a valve transmitter at the transmitting end.

https://www.youtube.com/watch?v=xqe1Y7SQX2k

How much feedback there is in the amplifiers also makes a difference. Lots of feedback will tend to remove the warmth of the tubes. And, as "warmth" is mostly due to a small amount of 2nd harmonic, push pull amps tend to cancel that out. Thus single ended class A tube amps with minimal feedback tend to sound warmer. But be careful of too much 2nd H (about more than 45dB down), else you can get too much intermodulation distortion products, which sound kinda like "mud".

I’ve been playing with Fourier series programming, and it occurred to me a Fourier Transform library
could measure the presence/amplitude of odd/even harmonics of fundamental tones.
But how could it ever be measured?
How did you get an audio signal into a software FFT library without solid state pre-amplification
from the computing platform’s own audio input hardware?
If you were looking at a audio from a loudspeaker on a scope or other analogue instrument,
then what was the pre-ampfification between the microphone and the rest of the device?