Perhaps there is a misunderstanding. The average audible band for humans is ~ 20Hz - 20KHz. CDs are sampled at 44.1kHz, (as mentioned above this is a product of my negro Nyquist, since sampling a perfect sinusiod at exactly two times the frequency, gives you a bunch of zeros).
Once the signal is sampled it has to be filtered to remove the aliased frequency components of the original signal. This is where over-sampling helps. By over-sampling you get wider gaps between the aliased signals, and put less stringent requirements on the anti-aliasing filter (so you can nearly uniform frequency response in the sampling band) and potentially a better sample. The quality of the final signal does not indefinitely increase by increasing the sampling rate.
Increasing the resolution of the signal becomes moot too because DACs and playback devices really are not that good (Not 24-bit good at least).
So, aside from the content of the music, you want higher quality still?