(14 June 1994)
Norman McLeod
Getting the Zenith-GE stereo system
to work well is never an easy task.
It produces stereo separation by a process which relies on very flat amplitude and phase characteristics being present across the passband from 30Hz - 53kHz wherever the signal is in multiplex form.
Anywhere on the tortuous journey from transmitter modulator to receiver demodulator things can go awry, and when the trip from coder to decoder includes a double RF path - when a link is in use - the hazards are doubled.
One excellent way of demonstrating the vulnerability of the stereo signal is to listen to the difference signal on its own. This is (L - R) and is transmitted supersonically between 23 and 53 kHz. All you have to do to hear what it sounds like is to disconnect the speakers from your stereo system, and then reconnect one of the speakers between the 'hot' (red) terminals of the left and right channels (either way round).
In the vast majority of amplifiers where the speakers share a common earth, this will reveal the difference signal in all its glory. In the minority of amplifiers with bridged output stages, you may need to connect the other speaker between the earth terminals to get the effect.
To check the system is working, switch your tuner to mono, and almost all the sound should disappear (there will always be a little leakage due to not-quite- perfect gain matching between left and right). In stereo, you should hear some of the music but virtually nothing from the DJ (since they are almost always centre-injected, and cancel out). Most records will sound peculiar in difference mode.
This technique can be applied to any stereo source, not just radio, to see how much use is made of the stereo domain. I have sat through our regional BBC-TV news programme which trumpets itself as being in 'stereo' before it starts.
In fact, there is no difference signal at all (i.e. the programme is in mono) except during the 'stings' played in between items, and during the weather forecast, where the presenter stands to the right of the map, and is duly panned across.
Similarly, BBC local radio transmits little if any stereo information during most speech programming. Almost everything on our local service is centre-injected, and the only time the difference channel is activated is on the rare occasion when a record or promo is played.
Perhaps someone at the BBC can try to justify why BBC local radio needs stereo FM at all, given that the stereo effect is redundant and unused for so much of the time. And the only effect of having the red light on, as far as the listener is concerned, is to make reception more noisy and distorted than it needs to be!
One thing you will notice immediately about the difference signal is that if normal FM reception is noisy, somewhat distorted or subject to interference, the difference signal will sound much worse.
And you may also detect production distortions which are normally masked by the sum signal. Anything that's been through a digital bit-scrambling process (politely known as compression) will often sound strange as a difference component, particularly if attempts have been made to exploit the alleged 'redundancy' between components of a stereo pair.
Can YOU hear the difference?
Any readers wishing to comment on what I have said are welcome to call me on 01 273 684 172 or send e-mail to normac@fastnet.co.uk
Copyright NJ McLeod 1995