ADI-2 uses 12 dB/Octave sloped Hi-/Low-Cut filters.

This kind of filter has a resonance at the crossover point.
Q-factor gives control over the amount this resonance.
It’s not the same as a gain or slope control.

The exact same behavior can be observed everywhere with natural, analog filters, e.g. as can be found in inductor/capacitance circuits, and all types of bass speakers and headphone transducers.

I would not just straight agree with this statement.

I’ve very favorably used high-Q low-cut filters for compensation of bass rolloff e.g. on open baffle transducers.

High Q boosts a certain range in the down-slooped part of their bass frequency response, while the low-cut removes frequencies below the usable part of their range, preventing overload.

The technique can typically shift down the linear bass response by up to one octave.
Doing so it’s important to take into account the power limits of the amp and transducer.

This kind of tuning can be found in most contemporary active loudspeaker designs, even studio monitors like Genelec or Neumann.
Specially half-open, ported speaker constructions like bass-reflex designs benefit very much.

The premise for successful use of high-Q low-cut-filters this way is precise tuning, only achievable with measurement tools.

Finally it’s not the single filter that counts for the audible result, but the transfer characteristics of the complete system.

An example for the RAAL SR-1a, the low boost is a combination of high-Q low cut, and bass shelf:

Gain  Freq.   Q   Type  (comment)
+5.5   130  0.8  Shelf  (adjust gain to taste)
 0.0    25  1.8  LoCut  (high Q boost the bass SR1a is capable, removes below)
-2.0  2.1k  1.5   Peak  (vocal presence)
-1.0  6.3k  3.5   Peak  (harshness)
-2.5  9.1k  5.0   Peak  (Sibilants)

Here’s a measurement showing what can be achieved, a “Harman target” bass response:

The measurement was made on my own head, so quite realistic:
Microphone in the outer part of the ear channel, plots with no compensation.