Could someone explain why adding a 0.01uF capacitor in parallel/bypass with a 2.2uF capacitor improves sound quality?@kalali It doesn't! Any capacitor has a certain 'speed' defined by a combination of its dielectric constant and series resistance. This varies from cap to cap (inductance plays a role too; this is caused by the fact that the cap is rolled in a spiral fashion) and even from value to value of caps of similar construction.
BTW this is quite measurable (notwithstanding the known effects of the specifications on paper :) ) and one way to do it is to measure bandwidth. Some caps roll off the signal a lot quicker than others. The bigger the capacitor, the more profoundly its affected by these aspects.
Since smaller caps tend to be faster than larger caps, essentially a phase shift is introduced by the inclusion of a smaller cap in parallel with a larger cap. In a power supply this is advantageous, but as a coupling cap in a nutshell it smears the signal (and being real, the signal is messed with by the cap regardless). The more transparent your setup, the easier it is to hear this!
If this is done in a circuit that employs loop feedback, the phase shift introduced can interfere with the phase margin of the circuit; YIKES! this could lead to oscillation.
So as a result, unless you are compensating for a deficiency elsewhere in the circuit or system, the best approach is to give it your best shot and simply install the best capacitor you can and keep it as low a value as is practical.
This is also why increasing the value of a coupling cap (for example to extend bass response) isn't always a good idea. Circuits behave according to math and the use of coupling caps is a really good example.