OK found this:
Bryston Cubed Differences
I can explain the differences and the thinking behind the 3 or Cubed Series amps, but it's a bit complex and long-winded, sorry:
A few years ago, we had a review in a Swedish magazine that pronounced the 14B SST2 to be 'perfect'. That was not based on 'he liked it'; it had nothing to do with auditory preference, it had to do with transparency. The reviewer had and still has a setup where he can literally 'bypass' an amplifier to drop it out of the signal path, (dummy speaker load, gain-reduction L-pad, leading to a second amplifier to the speakers). The 14B SST2 turned out to be impossible to detect in the signal path, on any signal they tried, from simple guitar music to full orchestra ffff, to an electronic metronome with instant rise time and no overshoot. And it was the first audio device in his history to do so.
That does not mean the amplifier is actually 'perfect', of course. It just means that it is not contributing, (or subtracting), anything the ear can detect, to or from the signal path in an otherwise exemplary system.
From that result and many others of our own, we concluded that static distortion measurements do not tell the whole story. It would seem that if static THD and IMD are low enough, 0.001% or -100dB, they cease to be relevant factors in the transparency of the device. More important are time smear, deviations in frequency-response, dynamic distortions, noise within the device or from RF interference, etc.
Thus we did not introduce our new 'Salomie' input stage (which we received a patent on) simply to reduce static distortion. Our newest 'Cubed Series' amplifiers do have lower THD and IMD than the 'Squared Series' that preceded them, but not significantly so. They measure out at about 0.001% or a bit below, over almost the entire frequency band. However, and we consider this to be most important, they are quieter and more musically accurate in a number of ways; First, the 20-20K noise is reduced, but the new input stage has far better PSRR, or power-supply rejection ratio. This means hash on the power-supply is rejected, in this case by more than 140dB. The input stage also has much better CMRR, or common-mode rejection. That means it rejects noise coming in on the signal cable much more completely than before. These are dynamic issues that can show up in a real-world system completely apart from static THD and IMD.
In addition to the above, the new Cubed Series employs highly effective RF rejection in the power supplies, which traps and eliminates radio-frequency hash from either interfering with or emanating from the amplifier via the power cable or signal cables.
The audible improvements we realized with these approaches are rather subtle. They would have to be with amps that started out 'perfect' in at least one reviewer's description, but they show up from very low listening levels, almost whisper-quiet, where the silences are more inky-black than before. It's almost an unconscious realization. They amps also cope better with dirt and hash of all kinds on the power cord, from distorted 60Hz Voltage to RF. Thus the quiet middle-of-the-night listening is even more relaxing and emotionally accessible. The depth of the soundstage and the precision of instrument placement is closer to reality than ever before.
Of course, that's what it really boils down to; recreating the original musical experience in your home. The 'you are there' feeling. It's what we have been pursuing from day one, and are still focused on. The Cubed Series is closer than ever.
