Stereo FM Radio.

As most radio listeners are aware, one of the main advantages of FM radio broadcasting over AM/Medium Wave is the availability of Stereo sound. This involves transmitting and receiving two related audio signals, the ‘left’ and ‘right’ speaker signals required to produce a stereo image. In principle, we can use two complete transmitter/receiver systems to send stereo sound. In practice, however, it makes more sense to use multiplexing.

When introducing an improved service — stereo for FM, or colour for TV — broadcasters are often stuck with the ‘old grannie problem’. This is the fact that there will be thousands/millions of listeners/viewers who can't afford (or don't want the new improved service). They want to be able to go on receiving mono/black & white just as if nothing had changed. For this reason one of the main engineering requirements when introducing stereo to FM radio was to do it in a way which went unnoticed by anyone using cheap old mono radios. This means that the system chosen is more complex than otherwise it would need to be.

The standard FM stereo system which has been adopted by broadcasters around the world is illustrated in figure 21.1. It is called the GEC Zenith system after its inventors.
The system uses frequency division multiplexing to combine the two signals destined for the left and right hand loudspeakers. The signals are first passed through filters which only allow through frequencies up to 15 kHz. The L and R signals are then added to produce a sum signal and subtracted one from the other to produce a difference signal. The sum is essentially a monophonic signal which is what we would send for playing through a single loudspeaker. The difference signal is used to DSBSC modulate a 38 kHz sinewave.

The DSBSC output is added to the sum (mono) signal and the combination is sent on the transmitter's FM modulator. A monophonic receiver can now ignore the stereo information simply by using a filter after its FM demodulator to block everything above 15 kHz. A stereo receiver has to have an additional circuit after the FM demodulator which can detect and demodulate the DSBSC wave. Once it has done this it has recovered the difference information and can recreate the left and right signals by adding/subracting

Now, as we saw in the last section, demodulating a DSBSC signal can be difficult due to the absence of the carrier whose frequency & phase we need to perform demodulation. In the stereo system this problem is dealt with by including in the broadcast signal a 19 kHz pilot tone. This comes from a 19 kHz master oscillator at the transmitter to which the 38 kHz subcarrier oscillator is phase locked. The 19 kHz pilot tone falls in a spectral region above the mono sum signal and below the DSBSC difference signal information. (The DSBSC signal extends ±15 kHz around 38 kHz since the input modulating signals are band limited to 15 kHz.)

The stereo receiver looks at incoming FM demodulated signals. It knows that the 15-23 kHz range should be vacant unless a stereo signal is being transmitted. If a 19 kHz pilot tone is present it can be recognised and used to control the frequency and phase of a 38 kHz oscillator in the receiver's stereo decoder. This can then demodulate the difference information and combine it with the mono (sum) signal to recover the stereo sound. The presence of the 19 kHz tone is also often used to make the receiver light up an indicator which flags that a stereo multiplexed signal is being received. (This doesn't, of course, guarantee the sound is in stereo. Broadcasters often leave the system on when transmitting mono material! Then L always equals R and the difference DSBSC signal is always zero.)

From earlier sections on FM we know that the noise power spectral density of a demodulated FM signal tends to increase with the square of the modulation frequency. This means that there will be more noise in the 23 - 53 kHz band used for the difference signal than for the 0 - 15 kHz band used for the sum signal. As a result we require a significantly higher input signal level to receive a stereo transmission with the same output S/N as a mono signal. A careful analysis shows that we need around a tenfold increase in the received signal level to obtain the same S/N for stereo as mono. This is why many radios & FM tuners sound noisier in stereo than in mono. A similar argument applies to interference. Stereo reception is significantly more susceptible to interference than mono. The warbling/twittering sounds produced by interfering with stereo FM/DSBSC signals are sometimes called ‘birdies’ by radio engineers, so you may see this term mentioned in magazines and books.

Content and pages maintained by: Jim Lesurf (
using HTMLEdit3 on a StrongARM powered RISCOS machine.
University of St. Andrews, St Andrews, Fife KY16 9SS, Scotland.