Department of Electrical & Electronic Engineering — Egerton University

Stereo FM Broadcasting

EEEN 462 — Analog Communications  |  4th Year Study Guide

Module 3  •  FM Multiplex Systems

1. Learning Objectives

By the end of this study unit, the student should be able to:

  1. Explain why FM broadcasting uses a sum-and-difference (matrix) format rather than sending the left (L) and right (R) channels directly.
  2. Write the full time-domain expression of the stereo multiplex (MPX) baseband signal, identifying every component.
  3. Sketch and fully label the stereo multiplex spectrum, including the L+R band, the 19 kHz pilot tone, and the L−R DSB-SC band centred on 38 kHz.
  4. Explain the purpose and operation of pre-emphasis and de-emphasis networks and compute their time constants.
  5. Describe how the L−R channel is modulated onto a 38 kHz subcarrier using a balanced (switching) modulator, and the role of the 19 kHz pilot tone in coherent demodulation.
  6. Draw the block diagram of a stereo receiver and explain the matrix recovery of L and R from L+R and L−R.
  7. Explain mono–stereo compatibility in both transmission and reception directions.
  8. Account for the division of modulation percentage among the components of the MPX signal.
  9. Describe auxiliary services (SCA) and the RDS/RBDS subcarrier.
  10. Solve quantitative problems on stereo multiplex composition, subcarrier frequencies, and modulation percentages.

Contents at a Glance

2. Introduction and Motivation

When FM broadcasting began in the 1940s it was monophonic. When stereo records and tapes arrived in the late 1950s, broadcasters faced a hard constraint: any stereo system had to keep the millions of existing mono FM receivers working. Simply transmitting the left (L) and right (R) microphones as two separate FM carriers would double spectrum usage and orphan every mono receiver.

The solution adopted by the FCC in 1961 and used worldwide (the Zenith–GE stereo multiplex system) is a masterpiece of analog communications engineering:

A stereo receiver recovers L and R by the inverse matrix operation:

Stereo matrix recoveryL = ½(L+R) + ½(L−R)     R = ½(L+R) − ½(L−R)
Key idea: The 38 kHz subcarrier and the DSB-SC choice are deliberate. DSB-SC conserves power (no wasteful subcarrier), and placing the difference information above 23 kHz keeps it clear of the mono L+R band. The 19 kHz pilot is cheap in power but lets the receiver rebuild the 38 kHz reference with exactly the right phase — critical, because a phase error φ rotates L and R into each other (L↔R crosstalk).

3. The Stereo Multiplex (MPX) Signal

Let the left and right audio signals be L(t) and R(t), each band-limited to 15 kHz. The composite baseband signal that frequency-modulates the main carrier (called the multiplex or MPX signal) is:

Stereo multiplex signalsMPX(t) = [L(t) + R(t)] + [L(t) − R(t)]·cos(2π·38k·t) + Ap·cos(2π·19k·t)
ComponentContentFrequency rangeTypical level
L + R (sum)Mono-compatible main channel50 Hz – 15 kHz45% modulation
L − R (difference)DSB-SC on 38 kHz subcarrier23 – 53 kHz (sidebands)45% modulation
Pilot tone19 kHz reference19 kHz (single tone)10% modulation
Modulation budget: 45% + 45% + 10% = 100% of the ±75 kHz deviation. If L−R modulation exceeds 45% while the pilot is at 10%, the instantaneous peak deviation would exceed 75 kHz and splatter into adjacent channels — transmitters therefore apply limiting to the MPX signal before the modulator.

Why send half the subcarrier frequency as the pilot? Because a 38 kHz tone at the required low level would sit exactly on top of the DSB-SC spectrum and was harder to filter cleanly; 19 kHz lies in a quiet gap of the spectrum (between 15 kHz and 23 kHz) and a frequency doubler regenerates 38 kHz exactly in phase with the original carrier.

4. The Multiplex Spectrum

f (kHz) S(f) L + R 0–15 kHz 19 kHz 19 pilot tone (10%) 38 kHz 38 L − R (DSB-SC) LSB 23–38 kHz USB 38–53 kHz 23 53 suppressed 38 kHz subcarrier is regenerated ×2 from the pilot main channel: mono-compatible stereo sub-channel (needs coherent demodulation with 38 kHz carrier)
Figure 1 — Stereo multiplex baseband spectrum: the L+R main channel (0–15 kHz) occupies the same bandwidth (15 kHz) with the same spectral shape as each L−R DSB-SC sideband (23–38 kHz and 38–53 kHz); the 19 kHz pilot sits between them, and the 38 kHz subcarrier is suppressed.
Why DSB-SC and not AM? If the 38 kHz subcarrier were transmitted at full amplitude, it would consume most of the modulation budget without carrying information, and the receiver could demodulate with an envelope detector — but then a phase/frequency error between transmitter and receiver would distort the difference channel instead of merely mixing L into R. Suppressing the carrier and sending a coherent pilot gives precise phase control and saves power.

5. Pre-emphasis and De-emphasis

FM noise at the discriminator output has a parabolic (rising with frequency) amplitude spectrum, so high audio frequencies suffer more noise. Broadcasting standards therefore pre-distort the audio at the transmitter and undo it at the receiver:

StandardPre-emphasis time constant τ3 dB corner frequency f_c = 1/(2πτ)
FM broadcasting (Americas, Japan, most of Africa/Europe FM)75 µs≈ 2.12 kHz
FM broadcasting (Europe, Australia; also used by RDS)50 µs≈ 3.18 kHz
Where it sits in the chain: pre-emphasis is applied to L and R before the matrix, so the L+R and L−R paths are both emphasised once. Applying it after matrixing would emphasise twice. De-emphasis is the last audio stage, after stereo matrixing in the receiver.

6. The FM Stereocoder (Transmitter)

L (audio) R (audio) Pre-emp Pre-emp MATRIX Σ (L + R) Δ (L − R) L R L + R main channel (0–15 kHz) L − R Balanced modulator (DSB-SC) 38 kHz osc. BPF 23–53 kHz (sidebands only) ÷ 2 19 kHz pilot tone (10%) Σ MPX FM modulator (MPX in → RF out)
Figure 2 — Stereocoder: pre-emphasis, matrix (sum/difference), balanced modulator generating L−R DSB-SC at 38 kHz, pilot derived by dividing the subcarrier oscillator by 2, and summation to form the MPX signal.

6.1 The Balanced (Switching) Modulator

The DSB-SC signal is produced by multiplying L−R by a 38 kHz switching (square-wave) signal and filtering:

Switching modulator outputv(t) = [L(t)−R(t)] · ssw(t),   ssw = (4/π)[cos ω_c t − (1/3)cos 3ω_c t + (1/5)cos 5ω_c t − …]

The BPF at 23–53 kHz passes only the term (4/π)[L−R]cos(2π·38k·t) — the wanted DSB-SC — and rejects the harmonics at 114 kHz, 190 kHz, … which would otherwise land outside the allocated channel after FM modulation.

6.2 Pilot Injection

Dividing the 38 kHz oscillator by 2 guarantees the pilot is exactly 19 kHz and phase-coherent with the suppressed subcarrier: when the receiver doubles the pilot back to 38 kHz, the regenerated carrier is in precisely the phase the modulator used.

7. The Stereo Receiver

antenna FM tuner + discriminator MPX LPF 15 kHz main channel L + R to L−R path BPF 23–53 kHz (L−R DSB-SC) Product demodulator (coherent) LPF 15 kHz → L − R pilot 19 kHz filter + PLL lock × 2 38 kHz 38 kHz carrier Stereo pilot detector STEREO lamp / blend to mono MATRIX L = ½Σ + ½Δ R = ½Σ − ½Δ De-emp L out De-emp R out
Figure 3 — Stereo receiver from discriminator output: low-pass for L+R, band-pass for the L−R DSB-SC, 19 kHz pilot recovery, ×2 regeneration of the 38 kHz carrier, coherent product demodulation, matrix recovery of L and R, de-emphasis, and stereo/mono switching.

7.1 Coherent Recovery of L−R

The product demodulator multiplies the DSB-SC sidebands by the regenerated 38 kHz carrier:

Product demodulator output (before LPF)v(t) = [L−R]cos(ωct) · cos(ωct + φ) = ½[L−R]cos φ + ½[L−R]cos(2ωct + φ)

The 15 kHz LPF keeps the baseband term: ½[L−R]cos φ. With perfect phase lock φ = 0, the full L−R is recovered; a phase error simply scales it (φ = 90° kills the channel entirely — which is why the 19 kHz pilot must be regenerated with a PLL, not a free-running oscillator).

Crosstalk viewpoint: a receiver phase error φ mixes the channels: the recovered "L" becomes L·cos²(φ/2) + R·sin²(φ/2). At φ = 10° the unwanted channel is already −21 dB down; broadcast specifications keep pilot phase errors below a few degrees.

7.2 Stereo/Mono Switching

If no 19 kHz pilot is detected (mono transmission, or weak signal), the receiver blends to mono: it mutes the L−R path and feeds the L+R output to both loudspeakers. This avoids the noisy "hiss" that the difference channel adds under weak-signal conditions.

8. Mono–Stereo Compatibility

CombinationResultReason
Mono TX → Mono RXMonophonic soundOrdinary FM reception of the L+R main channel
Stereo TX → Mono RXMonophonic sound (L+R)Mono receiver LPFs at ~15 kHz, rejecting pilot and DSB-SC band — plays the sum signal perfectly
Stereo TX → Stereo RXFull stereoPilot detected; matrix recovers L and R
Mono TX → Stereo RXMonophonic (both speakers)No pilot found; receiver blends to mono
Design triumph: the mono receiver never even "sees" the stereo information — it sits above 19 kHz and is removed by the standard 15 kHz audio filter. Hence the 1961 FCC requirement was met: existing receivers needed no modification. The price is that stereo receivers demodulate a noisier difference channel, which is why weak stereo signals are deliberately blended to mono.

9. Subsidiary Communications Authority (SCA) and RDS

The FM channel has room for more than stereo:

Spectrum usage order (baseband, kHz): 0–15 (L+R) → 19 (pilot) → 23–53 (L−R DSB-SC) → 53–75 (guard) → 57 (RDS) / 59.5–74.5 (SCA). The total must still respect the ±75 kHz deviation limit shared among all components.

10. Worked Examples

Example 10.1 — Composition of the MPX signal

During a quiet passage L(t) = 0.6 V and R(t) = 0.4 V (slowly varying). Find the amplitudes of the components of s_MPX(t) entering the FM modulator (ignore pre-emphasis).

L+R = 1.0 V (baseband, 45% modulation budget)  |  L−R = 0.2 V (modulates the 38 kHz DSB-SC)
sMPX(t) = 1.0 + 0.2 cos(2π·38k·t) + Apcos(2π·19k·t) — the stereo receiver will recover L = ½(1.0)+½(0.2) = 0.6 V and R = ½(1.0)−½(0.2) = 0.4 V ✓

Example 10.2 — Where do the sidebands fall?

The difference channel contains a 5 kHz tone. Where does it appear in the baseband spectrum?

DSB-SC: 38 kHz ± 5 kHz → 33 kHz and 43 kHz. (In general the L−R band fills 23–53 kHz.)

Example 10.3 — Pilot and subcarrier relationship

A receiver pilot filter passes 19.0 kHz. After the ×2 doubler, what frequency and phase accuracy is expected, and why does it matter?

Output: 38.0 kHz, phase locked to the transmitter's suppressed subcarrier via the pilot. A phase error φ scales L−R by cos φ; φ = 0° → full stereo separation, φ = 90° → no difference channel (mono result).

Example 10.4 — Modulation budget check

A transmitter sets L+R at 40%, L−R at 45%, and the pilot at 10%. Can it also add an SCA subcarrier at 4% modulation without exceeding 100%?

40 + 45 + 10 + 4 = 99% ✓ — just compliant, with only 1% headroom for peaks (in practice transmitters keep ~5% reserve and use peak limiting).

Example 10.5 — De-emphasis corner frequency

Compute the −3 dB corner of the standard 50 µs and 75 µs de-emphasis networks.

fc = 1/(2πτ): 50 µs → 1/(2π·50×10⁻⁶) ≈ 3.18 kHz;  75 µs → ≈ 2.12 kHz

11. Summary of Key Results

ItemResult / Value
MPX signals = (L+R) + (L−R)cos(2π·38k t) + Apcos(2π·19k t)
Main (mono) channelL+R, 50 Hz – 15 kHz
Stereo sub-channelL−R as DSB-SC, sidebands 23–53 kHz on 38 kHz subcarrier
Pilot19 kHz, 10% modulation, phase-coherent, ×2 → 38 kHz
Modulation budget45% (L+R) + 45% (L−R) + 10% (pilot) = 100% of ±75 kHz
Matrix recoveryL = ½(L+R) + ½(L−R); R = ½(L+R) − ½(L−R)
Phase error effectL−R scaled by cos φ; φ = 90° destroys the difference channel
Pre-/de-emphasis τ75 µs (Americas/Japan) → fc ≈ 2.12 kHz; 50 µs (Europe/AU) → ≈ 3.18 kHz
Auxiliary subcarriersSCA ≈ 67 kHz; RDS = 57 kHz digital data
CompatibilityStereo TX → mono RX plays L+R without modification

12. Review Questions

  1. Write the complete expression for the stereo multiplex signal sMPX(t) and identify each term with its frequency range and typical modulation percentage.
  2. Explain why the L−R channel is transmitted as DSB-SC rather than with the 38 kHz subcarrier present. What two benefits does suppression provide?
  3. Why is the pilot tone at 19 kHz rather than 38 kHz? Explain how the receiver regenerates the 38 kHz carrier and why phase coherence matters (use the cos φ argument).
  4. A difference-channel tone at 7 kHz is present. Sketch the baseband spectrum up to 60 kHz and mark all resulting spectral components.
  5. Show that a mono receiver reproduces the correct programme from a stereo transmission, and explain what a stereo receiver does when tuned to a mono station.
  6. Pre-emphasis is applied before the matrix, and de-emphasis after dematrixing. Explain what would go wrong if this order were reversed.
  7. A station allocates 42% to L+R, 45% to L−R, 10% to the pilot. Is 5% SCA modulation permissible? Justify with the modulation budget.
  8. Describe the role of every block in the stereocoder diagram (Figure 2), including the purpose of the 23–53 kHz band-pass filter.
  9. Explain the capture/blend behaviour: why do stereo receivers switch to mono under weak-signal conditions, and which signal path is muted?
  10. Compute the −3 dB corner frequencies of 50 µs and 75 µs de-emphasis, and explain the noise-shaping argument that motivates pre-emphasis.
Selected answers: Q3 — pilot sits in the quiet 15–23 kHz gap; ×2 doubler regenerates 38 kHz in phase; φ error scales L−R by cos φ. Q4 — sidebands at 31 and 45 kHz plus pilot at 19 kHz and L+R content at 7 kHz. Q7 — 42+45+10+5 = 102% → not permissible; peak limiting or reduced SCA level required.

Recommended Reading