The Nyquist criteria calls for sampling to be at twice the highest frequency of interest: if this is true the analog waveform is represented without error. BUT...Nyquist was talking about sine waves. Music is not a sine wave.
That's not quite true. The Nyquist Theorem states:
It is true that sampling at less than 1/2 the highest frequency gives rise to aliasing, creating components in the output after DAC that were not there to begin with. Obviously undesirable. But Nyquist used the phrase "at least" - sampling could be higher.
To represent in the digital domain a signal containing frequency components up to FHz it is necessary to sample it at LEAST at 2F samples per second.
In theory, the waveform does not really matter. Music is a periodic waveform and the Fourier Theorem states that:
The mathematics is solid. Problems arise in the implementation.
Any periodic waveform can be represented as a sum of harmonically related sine waves, each with a particular amplitude and phase ...
Regards,