But also correct here: > Compression means it's effectively using less of the [dynamic] range and the dynamics [range] could be represented with fewer bits in

But also correct here: > Compression means it's effectively using less of the [dynamic] range and the dynamics [range] could be represented with fewer bits in an optimal encoding. Lower DNR can indeed be encoded with fewer bits per sample.