The differences can be measured and the measurements often imply that the differences can be heard.
Whenever an audio product produces a distortion component that is reproduced at a level above the threshold of hearing, there is a chance that it will be heard. The reality is that most amplifiers produce distortion at levels that are significantly above 0 dB SPL in a typical system. In contrast, the distortion produced by the AHB2 is at or below 0 dB SPL (do the math if you doubt this statement).
The sound of an amplifier's distortion is not masked by room noise. We can hear tones that are as much as 30 dB lower than the ambient noise.
An amplifier's distortion may also not be masked by the music, especially when the distortion is separated from the fundamental by several octaves. In musical signals, low frequencies normally have much higher voltage swings than high frequencies. The high-order harmonics produced by bass content can reach levels that are comparable to high-frequency musical content. How loud is that flute that you hear and how loud are the harmonic cues that tell you that it is a flute?
Zero crossing distortion produces many high-order harmonics and it is most audible when playing at low levels.
IMD is particularly problematic because it does not resemble the harmonics that are produced by musical instruments. With IMD, the distortion does not occur at harmonic frequencies and it may be well separated from any musical content that would mask it. Many class-D amplifiers have IMD problems. Class-AB amplifiers can produce significant IMD when they have zero-crossing distortion. They can also produce significant levels of IMD when presented with a pair of high-frequency tones.
The distortion produced by electronics is different than the distortion produced by speakers. The differences in the spectrum can be seen on an FFT. Electronics are more prone to creating high-order harmonics which are not well masked.
Electronics produce harmonic distortion at perfect integer-related frequency ratios. Musical instruments do not. For example, the piano produces harmonics that are a bit further apart than integer ratios. The tuning of a piano is normally stretched to compensate for this non-integer harmonic spacing. The overtones produced by a piano string will beat against the integer-ratio harmonic distortion produced by electronics. A "warm-sounding" amplifier will make a piano sound like it is out of tune.
Harmonic distortion may change the sound of a musical instrument long before we can recognize the fact that the music is distorted. The ratios of harmonics to the fundamental give each musical instrument its unique voice. Any time you add harmonic distortion, you make changes to this voicing.
A non-linear phase response may be more audible than a non-linear frequency response. Errors in the phase response may create the impression that the frequency response is different. The frequency response of the AHB2 extends down to 0.1 HZ and up to 500 kHz so that the phase response is linear within the 20 Hz to 20 kHz audible band.
Many amplifiers distort when driving the low-impedance portions of the speaker's impedance vs. frequency curve. This doesn't show up in steady-state 8-Ohm and 4-Ohm tests. It is important to look at the distortion when driving very low impedances. The AHB2 stays clean when driving difficult loads and phase angles. In contrast, most amplifiers do very poorly into these difficult loads.
Here is an application note about a double-blind ABX listening test that we did between the AHB2 and a typical class-AB amplifier with decent specifications. This test was examining the audibility of zero-crossing distortion when playing a single tone through loudspeakers at a low level (0.01 watt producing 67 dB SPL at the listening position). It was very easy to hear the difference between the two amplifiers. I scored 25 correct out of 25 trials on my first attempt.
ABX listening test:
https://benchmarkmedia.com/blogs/application_notes/power-amplifiers-the-importance-of-the-first-watt
