What is the definition of an amplifier? It's an electrical circuit that increases the magnitude of it's input signal. To create an exact replica of it's input, it must increase all frequencies equally. What comes in should come out exactly the same, only bigger.
That's the ideal. Of course, reality is usually something different. If you take a look at John Atkinson's amplifier measurements using the dummy speaker load, you will notice that SS amps generally fit the definition stated above. So why don't some people like them?
If you look at the same graphs for reviewed tube amps, you will notice a pretty big spike centered around 1KHZ-2KHZ, followed by a dip in the 3KHZ-8KHZ range. Remember that decibels are logarithmic in nature, so a 1DB difference is about 20%, 3DB around 50%. Note that this frequency range y coincides with a typical tweeter crossover, which is not surprising since the dummy load is supposed to simulate a real world speaker.
So why do people perceive an amp whose high frequency performance is just about perfect as have an "unrefined" top end, especially when compared to other amps that don't really have a top-end, at least into an inductive load? When did omission and non-linearity become synonyms for refinement?
Given the current fads in speakers, i.e. two way monitors with small drivers, my guess is that tube amps don't deliver enough output at the frequencies near the crossover where the mid/bass driver is starting to misbehave for it to be noticable. Maybe that big peak at 1KHZ is masking the errors, some people are just more sensitive to them.
I was using a Bryston 3B-ST to drive a pair of Proac Response 1.5s. I was pretty unhappy, and was looking to change amps. By coincidence, around the same time Dynaudio revamped their Contour line and I ended up picking up a pair of 3.0s which I had been lusting after.
I never looked back at changing the amp, although I would like to play with a little more power. I believe the real difference is that the Dynaudios have a dedicated midrange driver. That little 6.5 woofer had to work too hard to produce bass, and I think it's long travel was screwing up the upper mids. On some tunes with a lot of low bass, you could actually see the thing flexing asymetrically.
I am not trying to invalidate anyone's choices, or even defend Bryston. I just find it amusing that we use wildly inaccurate devices (loudspeakers/ears) to measure the quality of an electronic circuit, when it probably should be the other way around.
