Do you know what? I've never really thought about it, I just acepted that valves are more linear than transistors.
That is perfectly true: valves are more linear than transistors. However, all too many folks latch onto that one fact and forget all about that output transformer. Linear valve + Nonlinear transformer= Nonlinear output. Valves also are low gain devices, so that there is insufficient open loop gain to linearize the transformer with inverse feedback. So a whole folklore grows up around these misconceptions, the marketing guys get hold of that, and the hype is lapped up like antifreeze at a petting zoo.
It's ok I understand I meant valve amplifiers in general, Valves alone are very linear, like you said it's lack of negative feedback due to their low gain and the output transformer being very non-linear that generally increases the distortion of valve amplifiers. Modern solid state amplifiers have a very high open gain despite the transistors are less linear than valves, the negitive feedback caused by the high open loop gain reduces the distortion, and there's no output transformer to mess things up.
I'm sure a class B valve amplifier (correct me if I'm wrong but valves lack crosover distortion so class AB isn't required) driving a high impedance loudspeaker for a load will have very low distortion, but it'd be bloody expensive and I bet you a cheaper modern solid state design could achieve similar performance.
Nonlinear BJT + Inverse feedback + No output transformer= Linear output. It's really that simple.
In fact the bigest non-linear element in modern audio systems is the loud speaker system, but these have improved a lot too.
Having said that nothing beats traveling wave tubes, klystrons and magnetrons in low noise high powered VHF and microwave amplifier and oscilator applications.
True: that's where "hollow state" beats solid state: RF applications.
I'd like to get my hands on some of these devices one day.
