And, which amp would that be?
An amplifier that has the same distortion profile as a tube amp of course.
In case this is mysterious to anyone, the ’sonic signature’ of any amplifier is the distortion it makes.
Tube amps sound the way they do because they have either a fairly prodigious 2nd harmonic (as in the case of SETs in particular), 3rd harmonic (as will be the case with a fully differential push-pull amplifier) or both. These harmonics mask the higher ordered harmonics which has two effects: it makes the amp nice and smooth and tends to bring out a sense of ’presence’ of images; ’musicians in the room’, that sort of thing.
In most solid state amps while there is always some 2nd and 3rd, they are usually suppressed by feedback enough that the higher orders are not masked- and so cause the amp to sound harsh and bright (that is how the ear interprets higher ordered harmonics).
One of the problems of traditional amplifier design has been how feedback is applied to the amplifier. In tube amps this tends to be the cathode of the input tube; in solid state amps usually the input of a transistor that is part of a differential pair at the input of the amplifier. The problem with either approach is that feedback node isn’t linear. So the feedback is distorted before it can really do its job. The feedback node in tube amps is more linear than seen in solid state amps so the feedback tends to be distorted less, but this still causes higher ordered harmonic generation that would not be there if the feedback node was linear.
There was a highly respected engineer named Baxandall who proposed that the non-linearity of the feedback node could be overcome with greater amounts of feedback.
Its been this issue why there was the hunt for greater amounts of feedback in the late 1970s. The problem that so many solid state designers didn’t realize is that the input of the amplifier was outside of the feedback loop, in addition to the distortion always added by the feedback node. This is what caused TIM (transient intermodulation); remember those days?? Really low distortion amps that sounded bad.
So feedback got a bad rap and this was the tip of the iceberg.
Another problem that was never overcome by tube and solid state designers until very recently is the phenomena of distortion rising with frequency.
The significance of this is that the THD number can sweep the actual distortion of the amplifier under the carpet, at the frequencies that are most important to the human ear!
This problem is caused by a lack of gain and bandwidth in the amplifier. Together they form a value known as ’Gain Bandwidth Product’ to engineers. This is the frequency at which, when the amplifier is run with no feedback, that its gain has fallen to a value of 1.
Obviously a gain of 1 is impractical- usually you need 25-30dB. 30dB is a gain of 1000. That’s not trivial. If the amplifier has a gain of 1 at 1MHz, what you do is divide 1 MHz by 1000 (if your gain is 30dB). You get a value of 1KHz; above this frequency the feedback will decrease on a 6dB (or more) per octave slope and so distortion will increase on that same slope.
THD is usually measured at a lower frequency- 100Hz is common. You can see where this is going! At 5KHz the distortion can by much higher than the THD suggests, and this is right in the middle of the ear’s most sensitive range (the Fletcher Munson curve, also known as the loudness curve). So an amplifier with this problem can sound bright!
This is why we didn’t use feedback in our tube amplifiers- we wanted a nice straight line across the audio band when graphing distortion vs frequency. This results in smoother sound!
You can do this with solid state and feedback, but you have to jump through some hoops. Class D offers a way to do this because its very easy to get insanely high Gain Bandwidth Product values- in excess of 20MHz. This allows the feedback to be supported across the audio band; distortion vs frequency is a ruler flat line. This makes the amplifier sound smoother.
I can't speak for all class D amplifiers, but in some I’ve heard, the non-linearities that cause distortion in the amp tend to make lower ordered harmonics (the 2nd and 3rd) and so can sound just like a tube amp.
If you’ve gotten this far through this post, the take away is that there is a direct line between what we can measure and what we hear. That old trope about hearing things we can’t measure was true in the 1980s but it isn’t true now. If you got thru this whole thing, now you have knowledge about exactly why amps sound they way they do. There’s no magic; its all engineering. Whether designers know all this stuff is a different story...