This is wholly, completely, and utterly false, speculative, and prejudiced not only have some of us "listened' we have also actually "measured" using reliable, repeatable, objective measurement techniques that are recognized by engineers, scientists, and industry as fitting, proper, and appropriate but of course if you are happy with Class D you should absolutely enjoy it but you're reasoning, arguments, and suspicions are unfounded and false.Wow. Because of the unqualified way in which this is stated, it renders the statement false.
I have a little Topping class D amp that makes 30 watts per channel. It has a lower signal to noise ratio than a Realistic SA-175 amp (despite the fact that the Realistic only makes 10 watts) we have in the shop that is rebuilt (being nearly 50 years old). In case anyone is wondering why I have an old Realistic amplifier hanging around, I put myself through college working at the regional Radio Shack repair center in the Twin Cities, and I enjoy troubleshooting and rebuilding work as a hobby; this amp is definitely a bit of nostalgia and its cute. I rebuilt the matching tuner too.
In our work the main thing that we've seen that contributes to noise in a class D has to do with the encoding scheme- in our case, Pulse Width Modulation (PWM). PWM relies on the use of a triangle wave generator and a comparitor that compares the incoming audio signal to the triangle wave and thus has an output that is either on or off. If the triangle wave generator is not perfectly steady in its frequency, or if there is a bit of DC offset at the input of the comparitor, the result can be a bit of white noise hiss in the loudspeaker. With fairly simple techniques this noise can be reduced to noise floors that are less than conventional amplifiers.