I'd really be interested in learning where that 10 dB figure came from.
I admittedly have very little experience in using NBFM, but 10 dB seems
awfully large to make a perceptable difference. I recall from a
communications course that FM detection has a threshold effect -- below
a certain S/N ratio, the noise effectively multiplies the modulation,
rather than adding to it as it does in low S/N AM. This is described by
Carlson in _Communication Systems_ as "mutilation" of the modulation. It
seems from his analysis that it would take much less than 10 dB to cross
the threshold, that is, to go from noisy but recognizable modulation to
badly distorted, "mutilated" audio.

In a quick experiment with a NBFM receiver and signal generator set for
5 kHz deviation, I could easily tell the difference in quieting
resulting from each 3 dB change in signal level, when the signal was
below full quieting. Subjective examination of the output signal on a
scope showed a transition from a recognizable but noisy sine wave to
apparent noise only, with a signal level change of only 2 dB.

So, what's the basis for the 10 dB figure?

Roy Lewallen, W7EL

Dave Bushong wrote:
. . .
In fact, 10 dB is how much it takes for a mobile listener, listening to
NBFM, to notice the difference.

TEN DB! That means you have to change from 1 watt to 10 watts for me to
hear the difference.
. . .