Most of these people like to talk about science, measurements and distortions - having zero engineering and technical background.
Here is the great article about tubes vs transistors:
https://spectrum.ieee.org/consumer-electronics/audiovideo/the-cool-sound-of-tubes
1. Transistor as amplification elements has much higher level of distortions compared to triode.
In addition to that, transistors have temperature dependence distortions that tubes don't have. This kind of distortions have inertia that make them even worse for human brain perception.
2. "Good characteristics" of transistor amplifiers achieved by using feedback everywhere, that causes to replacing of low order harmonic distortions to high order harmonic distortions. Human brain recognises of a smallest amount of high order harmonic distortions because they are unnatural.
3. The single advantage of transistor amplifiers is lack of output transformers.
4. Our brains perceive distortion in a completely different way than instruments. There is a whole science of psychoacoustics that deals with this issue.
5. High power amplifiers are useless, cause compression and distortion of speakers.
Here is an example:
https://www.linkedin.com/pulse/power-compression-vs-thermal-distortion-loudspeaker-alexander-wilson/
"
Taking an example of a voice coil using 10 meters of 0.2mm Copper with a typical resistance of 5.5 Ohm at 20 degrees C.
Apply power (music signal) & the voice coil jumps to 120°C and the resistance rises to approx. 7.5 Ohm….Now the resulting current flow through the voice coil drops by approx. 35% ….So does the SPL….This is a 35% distortion!
Orders of magnitude worse than the usual suspects of THD or IMD which loudspeaker manufacturers like to quote.
With passive crossovers in multi way systems thermal power compression will shift XO frequencies when the voice coil resistance changes – resulting in summation errors between the two drivers involved, depending on program and listening level.
The same holds true for notching out resonances or compensating impedance….The more complex the passive crossover the worse the TD effect becomes.
Why 120 degrees C?
This is a safe or medium range voice coil temperature, many voice coils hit double or triple this temperature on a regular basis.
The energy involved is as follows:
The weight of our VC wire is slightly below 3g if we calculate length multiplied by cross sectional area multiplied by specific mass of copper.
We require approx. 1 Watt for 1 sec for each Kelvin temperature increase (no cooling assumed here), looking at the thermal material constants for copper
To heat up this < 3g of copper wire our voice coil is made from, we only need 10 sec of 10W input to finally arrive at a 100°C increase – or – if we have fortissimo playing – a very short 1 sec of 100W input."