The only way "Channel Separation" could affect audible performance is if published "Channel Separation" specs are over stated when compared to real-world listening experience.
Separation between L/R ears of the average human being barely (if at all) exceeds single digit dB. Separation is limited by shared head space, distance between the two ears, head/face shape, outer ear shape, gelatinous quality of the material contents of the head, and ear/nerve functions. IIRC separation can not exceed 10-12 dB. Channel separation is further minimized by the listening room, but that's another subject.
"Channel separation" for stereo power amps with one power supply shared between two channels exceed by multiples the separation between the ears of the listener. Multiples as in 4-6 times minimum, about 500%. Even the worst high end amp spec exceeds human ear sensitivity by 40-50 dB.
Whatever is the audible advantage of dual mono vs. single power supply, per the above math, the cause is not likely attributed to increased channel sensitivity, again, unless the test spec is dramatically overstated (optimistic) compared to real world listening experience under dynamic music conditions driving a real loudspeaker.
If the current test for channel separation is accurate, that would seem to indicate dual mono power supplies or mono blocs conduct current quicker and more efficiently, for better dynamic peaks. If this is the case, then what spec more accurately predicts the advantage of DM or mono blocs vs. stereo/shared PS?
Hypex Ncore NC400 + SMPS1200 stereo seems indistinguishable from NC400 + SMPS600 mono. Ditto, the distributor for my all time favorite amp said the designer agrees his properly engineered stereo version (i.e. large enough PS and power transformer) should equal the mono blocs (though they always feature the latter because of buyer preferences).
Suppose we compare one stereo amp with PS capacity equal to sum total power supply of DM or mono version of same amp. Let's arbitrarily define the total PS capacity as "1200W." In every case, no matter how transient, of PS demand being less than 1200W sum total both channels, and one channel requires over 600W, the stereo version could theoretically exceed performance of DM or mono. For instance: 700W required L channel, 500W required R channel. In this case the L channel of the DM and mono are short-changed 100W.
Hugh Dean of AKSA said as much in post about one of his amps.
In the above case the only known potential advantage of mono is shorter speaker cables vs. any stereo amp, and advantage with which I am familiar. This potential advantage is largely eliminated by siting the stereo amp exactly mid point between L/R speakers, and using the shortest possible speaker cable with vanishingly low series resistance. I found affordable 4AWG that works well with a simple parallel run of 30AWG to minimize the 4AWG's self inductance. 6AWG would have been my first choice, but the next closest step available for similar cost was (IIRC) 10AWG.
Experience seems to indicate the on line speaker cable resistance calculators are optimistic. It's critical to input the speaker's lowest impedance; the nominal figure is useless. You really gotta check the speaker's impedance graph for this figure.
I suspect phase angle affects minimum cable gauge spec, yet phase angle is never a calculator data point for this spec.
