You can tailor the sound through the topology. For example, a purely-single-ended design will exhibit a 2nd order harmonic distortion which will give it a rich sound. If no feedback is used, it will not be harsh until pushed hard, but will feature dynamic character far above its actual power. Yet it will also have great inner detail as distortion vanishes at lower power levels. Sound familiar?
Sunn, a well-known guitar amplifier company, used to make a line of transistor amplifiers back in the 1970s that were recognized for their rich sound. They employed a single-ended zero feedback FET preamp which was primarily responsible- that 2nd harmonic was the richness.
A tube amp with a single-ended input and a push pull output will exhibit a 2nd harmonic with a prominent 5th. The 2nd will be less than that of an SET so it will be less rich, and the 5th will contribute to brightness and some hardness by comparison.
Choosing what topology IOW says a lot about how the resulting circuit will sound. The art part of it is understanding the human hearing perceptual rules, as in general the audio industry tries its best to ignore them. So designers have to educate themselves about that subject, and sort out what it is exactly that they are trying to do. IMO this is still engineering, but I can see some calling it 'art' and maybe it is...