I don't know... this test is unscientific... clearly mud and banana can have an unintended side effect that makes audio sound better and needs to be investigated immediately.
on a more serious note.. doesn't seem like the "good" audio was good? there is a huge difference between noise free audio and garbage integrated audio / speakers with hizz imbalance and peaking... if the "good" audio is bad then there obviously won't be a difference between any of them.
which makes me think... banana and mud are noise filters... hmm...
The experiment compared .wav files directly ripped from the cd to these same files played through copper/banana/mud.
So yes, the “good” audio is good.