Also known as: frequency modulation
Frequency modulation (FM) encodes information by varying a carrier’s frequency while its amplitude stays constant.1 The instantaneous frequency swings above and below the resting carrier in proportion to the message, and the peak swing is the deviation. Edwin Armstrong demonstrated in 1933 that this trade of bandwidth for noise immunity gives dramatically cleaner audio than AM.
How it works
Where AM writes the message onto the envelope, FM writes it onto the instantaneous frequency: f(t) = fc + Δf·s(t), with Δf the peak deviation. Because the information lives in frequency, not amplitude, an FM receiver can pass the signal through a limiter that strips away any amplitude variation — noise, fading, ignition sparks — before recovering the audio, and none of that amplitude noise survives. This is the source of FM’s headline advantage: above a threshold SNR, each extra decibel of received signal buys more than a decibel of audio quality, and wide deviation trades more RF bandwidth for still-quieter audio.
The bandwidth an FM signal occupies is not just twice the audio; it depends on the modulation index β = Δf / fmax, the ratio of peak deviation to the highest message frequency. Carson’s rule estimates the occupied bandwidth as roughly 2·(Δf + fmax).2 Broadcast FM in the US uses ±75 kHz deviation with 15 kHz audio, so β ≈ 5 and each station spans about 180–200 kHz — a wideband signal. Two-way land-mobile radio instead uses ±2.5 to ±5 kHz narrowband FM to fit many channels into a crowded band. Both add pre-emphasis that boosts treble before transmission and cuts it after, because FM noise rises with audio frequency.
FM also shows a capture effect: when two signals share a channel, the stronger one by even a few decibels seizes the limiter and the weaker one vanishes, rather than the two mixing audibly as they would in AM. This is why FM two-way traffic sounds clean but distant AM airband stations beat together.
Variants
FM and phase modulation (PM) are close relatives — FM is PM applied to the integral of the message — and both belong to the broader family of angle modulation. The digital descendant of FM is frequency-shift keying (FSK), which switches the carrier among a discrete set of frequencies instead of sweeping it continuously; its constant-envelope, phase-continuous refinements (continuous-phase modulation, GMSK, C4FM) underlie most digital land-mobile radio.
Relevance to SDR
FM broadcast (wide deviation) and narrowband FM two-way voice are everywhere, and FM is
usually the first mode an SDR beginner demods. An SDR recovers FM by differentiating the
phase of the complex IQ baseband — the derivative of phase is
instantaneous frequency — typically as an atan2 discriminator or the algebraically
equivalent (I·Q' − Q·I')/(I²+Q²). Analog narrowband FM voice is the direct analog cousin
of the digital 4FSK voice modes GopherTrunk decodes: P25 Phase 1
C4FM, DMR, and NXDN are all, at
the physical layer, carefully shaped forms of FM, so a good FM discriminator is the front
end of a good digital-voice symbol recovery chain.
Sources
-
Frequency modulation — Wikipedia, for the definition, deviation, modulation index, and the capture effect. ↩
-
Carson bandwidth rule — Wikipedia, for estimating the occupied bandwidth of an FM signal from its deviation and message bandwidth. ↩