Bob, the "high harmonics/square wave" comment is not actually true. Been through the exercise and the theory of how the analog output wave is constructed from the digital audio stream of pulses and I used to think the same thing, but after studying the actual process it actually does not work the way one would think. The impulse response of each individual sample is not a square wave but rather a ringing filter response. When you add all of those ringing filter responses together, you end up with a smooth analog wave form with the intended frequency response. You do not end up with a bunch of high harmonics that are unintended. This has to do with the output anti-aliasing filter technology that has been present for decades.