No, I'm not thank goodness, I'm sure you were annoying to her as well.
With your pestering attitude I shouldn't answer your question now that I have the data.
But as a courtesy to the rest of the members here they are:
The Focusrite 2i2 output is feeding an unbalanced consumer preamp.
The measurement sample is taken from the headphone out as a matter of convenience.
There are numerous pieces of bench equipment connected to the preamp all RCA connections e.g. additional converters, a tuner etc.
One of those converters, a Roland Quad Capture, is also connected to USB on the same computer and preamp. So we're seeing its' USB hash contribution too.
The monitor level is set at death-defying level so there's lots of gain.
The same Focusrite 2i2's A/D is sampling the headphone output.
The blue trace is without the iDefender which has direct connection to the USB port and USB bus power.
The green trace is with the iDefender using an external Meanwell power supply and the iDefender's USB power injector.
The peaks at 7 and 16kHz(?) are about the same level in both - the green trace overlays the blue as it was recorded last.
When AudioTester is running the FFT I hear the noise signature change and that is likely causing the peaks - particularly when using the iDefender - to increase in the last octave or two.
Improve what? I bet if you test it there is no real improvement, and certainly no audible improvement.
It's 20 dB quieter, in this case, with the iDefender.
As I said your mileage may vary.
Balanced measurements look far, far better when I use the 2i2 on the test bench and don't have this dirty of a noise floor signature.
I think a lot of what we hear in USB-based converter idle channel noise is not in-band noise but modulated RF on the data lines and internal out-of-band noise being rectified and folded back in-band. I think that's why we see the overall noise floor rise. It damn sure sounds noisier.
