You could do a FFT on the mic signal and noise shoud be close to constant amplitude across the spectrum. And speech would have content centered around five to eight kHz if I remember correctly depending on the gender of the speaker.

-Finally received my Mark2 no thanks to customs.