Yes, you understand exactly what I was trying to convey when describing why the limiter must be in for full power testing.
In a brief description of how a (conventional) limiter works, when the signal crosses a preset threshold (like the clip threshold of the power amp), a control circuit detects this and depending on the attack, hold, and release time constants used, plus any frequency weighting and averaging of the peaks versus detecting the peaks, the GAIN of the amplifier is reduced for a period of time, thus eliminating clipping. The problem with a bass signal is that the necessary time constants vary across the spectrum and as you move around quickly the signal can pump if the time constants are too fast or can audibly turn down if they are too slow. In many instances, for a simpler limiter, there is no happy medum. Our limiter does not worry about trying to second guess a gain reduction equation but looks specifically at what is happening during the overload process and manages the signal in a way similar to how a tube output stage overloads. So, as the signal approaches clipping the drive signal gradually becomes less efficient at driving the power amp's output stage, gently enters a soft clipping period and when driven harder, enters a more stiff cliping until the power amp actually clips. Since the power amp is a (relatively) low feedback design (also like a tube power amp), it's behavour is very graceful when it does finally clip. All of this happens over a 6dB range.
So to put it in a nutshell, it's not a traditional limiter but it acts like one in that the natural compression/limiting of this circuit emulates quite nicely a tube output stage. Sometimes you have to go back and make sure that the solution (conventional limiting) isn't worse than the original problem. In this case, MY opinion is that it's a more useable, playable solution even if it looks numerically "wrong" by conventional standards.