The noise floor in analog formats could be said to be infinite whereas in digital you need to define the range that is used for the noise floor which then has a finite bit depth.* The resolution of the audio band is gigantic in HD digital compared to analog tape.
If you saw one of those youtube videos where someone tries to claim that you hear "stair step" audio with missing holes in it with digital... That's truly a gross misunderstanding of the system. It does not work that way at all. If it did, it would be garish crunchy distortion and just mutilated. Even SD digital delivers smooth continuous waveforms at the DAC output. You can resolve half the frequency of your sampling frequency. That's the bottom line in the physics. Further, HD digital is NOT interested in any data above the range of hearing. The point is having a wide margin between the top of the audio band and the sampling frequency to eliminate said sampling frequency bleeding or aliasing into the audio. The physical circuitry is easier to build to run cleaner at HD.
* More...
The bit depth in digital is finite (24 bits or 16 bits). As you put a lower and lower volume signal into it, there are less bits left to define the volume steps and they get coarser. A simplified example in 16 bit: Take a signal that is half full level. It only gets the bottom 8 bits. (The top 8 are all zero.) So you suddenly have an 8 bit recording for that element. Call the bottom 8 - 12 bits the "noise floor" and then the lowest signal recorded to the system is still 12 - 16 bits in a 24 bit system. 16 bit CAN still hold a very dynamic signal but you DO have to be careful not to waste any bit depth anywhere! And now you know one of the reasons CD's started being peak limited and boosted. Better to lop off a few of the loudest peaks than loose the depth and 'meat' of the sound. Of course the volume war thing went WAY beyond any shortcomings of the 16 bit format! But it actually started innocently with that.
