Digital recording has a hard ceiling and a hard floor. Meaning, when your recording level is all '1s' anything a ove that is clipped, totally. Anything below all '0s' doesn't exist. Analog had 3-6dB headroom above 0dB and at least 10dB below the noise floor. Recording and mastering engineers had to re.earn a careers full of technique to go digital.
Similarly, low level information, ambiance, string sounds, the difference between a Stradivarius and other violins, a Strat vs. a Les Paul, has fewer bits with which to be described, thus less detail is captured. Note this is y-axis data, and has nothing to with sample rate and Nyquist Theory - that's all x-axis.
Digital has this limited operating 'space' to which the recordings must be confined. Analog is much more forgiving in that regard. To get a digital recording space that is greater than analog requires enlarging both bit depth and sample rate. 24-bit and 96KHz sampling accomplish that, but that was not technically oe economically feasible for a mass market product when the 16-bit, 44.1Khz sample rate CD was developed.
The final piece of the puzzle is the DAC. Analogous to the role of the phono cart, it is the transducer between the digital and analog domains, just as the phono cart is the transducer between mechanical and analog domains. And while there's are lots of ways to convert all the bits successfully, reconstructing them into an analog signal absent distortion artifacts require a 'reconstruction filter'. It is in the execution of that where a DAC adds the vast majority of its audible signature.
