QUOTE (skamp @ May 4 2013, 11:40) <{POST_SNAPBACK}>
First, I plugged the EMU's line out to its line in, and recorded with nothing playing, and with silence playing (with the same Fuze+ input gain).
It would be better to record the noise floor of the line in only, because the quality of the line output of the EMU is not really relevant to testing the Fuze+. That way, the noise levels could be compensated for the EMU noise. But even if the full loopback noise power is subtracted (which gives an optimistic estimate), there is only a ~2 dB improvement in the -36 dB "silence" case:
CODE
Tone Fuze+ Fuze+ Fuze+ Fuze+ EMU Fuze+
0 dB silence -12 dB -24 dB -36 dB loopback -36 comp
L: -1.117 -97.509 -99.638 -100.398 -100.546 -104.494 -102.785
R: -0.868 -97.161 -99.697 -100.134 -100.302 -104.341 -102.481
0 dB silence -12 dB -24 dB -36 dB loopback -36 comp
L: -1.117 -97.509 -99.638 -100.398 -100.546 -104.494 -102.785
R: -0.868 -97.161 -99.697 -100.134 -100.302 -104.341 -102.481
All the values in the above table are A-weighted dB relative to 0 dBFS in the recording (so a constant input gain is assumed), and are from the "silence" version of each test. The "nothing" files do not differ by more than a fraction of a dB.
The noise level only seems to decrease by a few dB at low vs. maximum volume, which explains why there is some audible hiss with very sensitive IEMs, even though the dynamic range is good at full volume (which would be too loud with those IEMs).