Also known as: pre-emphasis, de-emphasis, emphasis
Pre-emphasis and de-emphasis are a complementary filter pair used with frequency modulation: the transmitter boosts high audio frequencies before modulating (pre-emphasis), and the receiver applies an exactly inverse cut after demodulating (de-emphasis).1 The two cancel to leave a flat overall response, but because the receiver’s cut also attenuates the noise that FM concentrates at high frequencies, the net effect is a quieter, cleaner signal.
How it works
An FM demodulator produces output noise whose power rises with the square of frequency — a “triangular” noise spectrum — so without correction the high end of recovered audio is much noisier than the low end. The fix is to pre-distort at the source: a simple first-order high-pass shelving network boosts treble before the audio modulates the carrier, characterized by a time constant τ that sets the corner frequency f = 1/(2πτ). Broadcast FM uses τ = 75 µs in the Americas and Korea and 50 µs elsewhere; narrowband land-mobile FM commonly uses 750 µs.
At the receiver a matching low-pass network with the same τ cuts the treble back by the identical amount. On the wanted audio the boost and cut cancel exactly, so the listener hears flat response. But the receiver’s cut is applied after the demodulator, so it also attenuates that rising high-frequency noise — and the parts of the spectrum where FM noise is worst are exactly where de-emphasis attenuates most. The result is a meaningful improvement in high-frequency signal-to-noise ratio for free, at the cost of a little headroom (loud treble transients deviate the carrier further, which the pre-emphasis stage must be limited to control).
The reason a fixed time constant works so well is that the FM output-noise spectrum is predictable — it rises at a fixed 6 dB/octave — so a first-order network with a matching corner can track it almost exactly across the audio band. Because the pre-emphasis and de-emphasis transfer functions are exact reciprocals, the pairing is transparent to the wanted signal regardless of program content; only the noise, which is added after the transmit pre-emphasis and before the receive de-emphasis, is affected. The catch is headroom: pre-emphasis can lift a cymbal or sibilant by 15-17 dB at the top of the band, and if unchecked that would drive the transmitter far past its allowed deviation. Broadcasters therefore place a fast pre-emphasized peak limiter after the boost so treble transients are clamped before they overdeviate — which subtly ties emphasis to the loudness processing that gives commercial FM its characteristic sound.
Relevance to SDR
Any software FM demodulator that wants to sound right must apply de-emphasis with the correct time constant — a digital filter is a one-pole IIR — or the audio sounds harsh and bright. Getting τ wrong (using 75 µs on a 50 µs signal, say) leaves a mild treble tilt. For broadcast FM the de-emphasis is applied to the recovered mono/stereo audio, not to the composite baseband, so a receiver must de-emphasize after stereo decoding. In land-mobile digital modes the concept mostly falls away: P25 and DMR carry digitized, vocoded audio rather than analog FM audio, so there is no analog emphasis stage in GopherTrunk’s digital decode path. Emphasis matters to GopherTrunk-adjacent work only when demodulating conventional analog FM voice, where the standard land-mobile 750 µs de-emphasis should be used.
In practice
The regional split matters when decoding real signals. A broadcast receiver hard-wired for the 75 µs Americas standard fed a 50 µs European signal leaves a gentle high-frequency lift of a few decibels — audible as extra brightness or sibilance but not fatal — while the reverse sounds slightly dull. In software this is trivial to make configurable: the de-emphasis is one first-order IIR section whose single coefficient encodes τ, so a demod can offer 75/50/750 µs as a menu choice. Land-mobile analog FM’s 750 µs constant corners much lower (~212 Hz), reflecting the narrower 3 kHz voice band and heavier deviation limiting of communications audio versus hi-fi broadcast.
Sources
-
Emphasis (telecommunications) — Wikipedia, for the pre/de-emphasis pairing, time constants, and FM noise-reduction rationale. ↩