I am using "HDSDR" software to demodulate a WFM broadcast ("Classic FM" in the UK, on 100.1 MHz, 192 KHz bandwidth). I'm using an "SDRPlay RSP1" software defined radio and a cheap indoor discone antenna.
I'm trying to understand more about the audio demodulation and I'm puzzled by the spectrum display. Here's a screengrab...
Annotation "A" is (according to wiki) the "Pilot Frequency" at 19 Khz.
Annotation "B" is a rather fascinating pattern that is clearly part of the stereo audio modulation. But it is centred at 38 KHz, which is far above human hearing capability. Although I notice that it's also twice the pilot freq.
Annotation "C" is RBDS.
Question(s)
What is the nature of the pattern in annotation "B"? When music is playing this area is rich with visible modulation. But when it's just a simple human voice (the radio show presenter) speaking this area goes totally invisible and blends into the background without a trace.
I vaguely suspect that a human voice (0 Hz to 4 Khz or thereabouts) lacks the dynamic range to show up on the AF waterfall. In contrast, a human voice speaking with some very quiet backing music (e.g. an advertisement track playing) does show up on the AF waterfall.
Is the station doing something special to the broadcast AF modulation during a musical segment, which is absent when the presenter is speaking? Or is it literally that the presenter's voice isn't rich enough in dynamic range to even show up at all?