Think of it this way, it is a combination of the loudspeaker and the amplifier that creates the dynamic range:
102dB/Wm sensitive speakers unpackage 102 dB range from 1W source material.
From the same 1W source material 82dB/Wm speakers unpackage only 82dB.
No matter how many watts we throw at it, the more sensitive is always giving us a wider dynamic range.
Now, if you need 102dB level with the 82dB/Wm speakers, then you need to bloat the 1W source to 100W, creating x100 distortion along the way compared to the 1W. To get the 100W reasonably error-free, to reduce the errors made by the additional amplification, we need feedback that removes not only the errors but substantially compromises the low level details as well, leading to a collapse of the low level of the dynamic range.