And this is your first mistake, Amir - what Swenson is saying is that the output from the DAC is perceived as lower in noise - in other words with music signals the perception is that everything emerges from a quieter background.
[emphasis added by me]
I can measure the noise floor on a recording using a spectrum analyzer. This includes recordings I have personally made. Noise floor on playback will be the worst element in the chain, either of the recording itself or the combined elements in the playback chain.
One of the microphones I use is a Neumann TLM 103, which has the following stats:
Equivalent noise level, CCIR1): 17.5 dB
Equivalent noise level, A-weighted1): 7 dB-A
Signal-to-noise ratio, CCIR1) (rel. 94 dB SPL): 76.5 dB
Signal-to-noise ratio, A-weighted1) (rel. 94 dB SPL): 87 dB
Maximum SPL for THD 0.5%2): 138 dB
The microphone's self-noise in this case is 7 dBA (quite good). This gets added to the ambient acoustic noise of the recording venue, which in the case of a local jazz venue is about 40 dbA, about as quiet as a library. This gives the following noise floor for the recording:
7 dBA (microphone self noise) + 40 dbA (venue) = 47 dBA
Peak SPL at the TLM 103 during recording (in my last real life example) = 92 dBA
Raw (uncompressed) dynamic range at the microphone = 92dBA - 47dBA = 45 dBA
This puts the recording noise floor WAAAY above the noise floor of digital silence on a 24bit recording (-144 dB), WAAAY above the jitter floor of most good DACs (> -100 dBFS), and WAAAY above the noise floor of other items in a quality electronics playback chain (> 100dBA SNR).
So given all this,
please explain how this device will make it so that "everything emerges from a quieter background" when the weak link the quietness chain is the recording itself?