My simple listening experience with solid-state and tube amps was that I couldn't really detect a difference between the two at low gain(linear) levels, assuming similar bandwidth and group delay characteristics. With my electronics work many years ago in doing experiments on breadboarded circuits where I had scoped the outputs from these circuits I could see the departure from linearity when these circuits were overdriven. The solid state amp would clip, depart from linearity, with a sharper knee, equivalent to superposing a square wave onto the linear curve to a first approximation. The Fourier representation of this square wave is a series of odd harmonics that doesn't converge rapidly, meaning that the amplitudes of these odd frequencies diminish slowly. Thus the output of the overdriven solid state amp will be rich in high order odd harmonics, not musically consonant with the fundamental. The tube triodes that I scoped exhibited a softer knee at the point of departure from linearity. The type of Fourier approximation to this curve was more like a quadratic (even order) superposition that was composed of even harmonics, consonant with the fundamental.
Whether the even or odd harmonic generation by different types of amplification is a good or bad thing is, I think, a matter of personal taste. If I crank the gain on the ss preamp circuit and blend the output with a clean signal from my parallel tube circuit, I can generate some interesting results. I would say, in my experience, that I have to be more careful with overdriving my ss amp section as compared with my tube section. As somebody else mentioned, EQing out some of the high frequency odd harmonics from an overdriven ss amp might give an acceptable type of OD distortion. If I had a HF tweeter in my cab I would want to be especially careful about those HF odd order harmonics from an overdriven ss amp.
