You guys lost connection to reality, I mean, really totally. Amir is the one to blame in the end, measuring mostly irrelevant stuff and then the unwashed come along and try to make a case out of nothing from graphs and numbers they fail to interpret correctly (this increasing tendency of backseat driving here on ASR is putting me off more and more, btw).
That little 10dB rise of low/mid frequency in one channel, compromising SINAD by only 3dB (because it covers only a small part of the frequency range)... this is completely irrelevant and this is clearly not a fault nor a design issue, it could be just simple 1/f noise from components. Note the level at which this is happening!