Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!rutgers!apple!kanner
From: kanner@Apple.COM (Herbert Kanner)
Newsgroups: sci.electronics
Subject: Re: Stereo vs. Mono: Q about AM Stereo
Message-ID: <1666@internal.Apple.COM>
Date: 5 May 89 01:18:38 GMT
References: <4046@bgsuvax.UUCP> <21000021@m.cs.uiuc.edu> <1972@hp-sdd.hp.com>
Organization: Development Systems Group, Apple Computer
Lines: 55

>>On the AM broadcast band, there is a station frequency every 10KHZ, starting
>>at 550 and going up in increments of 10 to 1600. If a station were at 760KHZ,
>>they would have space 5 below center frequency and 5 above center frequency
>>for their lower and upper sidebands. This is a total of 10 in width.
>>True, each sideband is limited to 5, so, they need to used single sideband
>>and gain 5MHZ in band width, they have the space there, it is just not
>>being utilized to the best advantage. SSB is still AM, I wonder if the
>>FCC would allow it? It would take several years to junk out all of the
>>old AM receivers, maybe 25, but, they would eventually disappear, and,
>>if you wanted the home radio to get either mode of transmission, you
>>may be inclined to upgrade sooner, as the SSB receiver has to reinject
>>the carrier locally, so demodulation can take place.
>>
> I think you misunderstood me when I said that AM is restricted to 5khz.
>What I meant was since the space between stations was 10khz and since 
>envelope detection is used for demodulation, each 5khz sideband of the 
>transmitted signal is mirror image copy of the other. If the transmitted
>AM modutated signal is limited to 10khz, then the baseband signal is
>5khz. Now, I have been told that in LA and San Diego, some stations have
>obtained permission to transmit a baseband signal of 10khz, but that
>means that they are taking up 20khz of the AM spectrum. These are
>probably the stations that seem to take up a large portion of the AM dial.
>
>You are correct there is a 5khz sideband above and below the center 
>frequency, but in the current AM scheme, these sidebands contain the 
>exact same information. SSB AM would be a much better system 
>especially since the coherent detection required also gives a slightly
>better signal to noise ratio -- but the cost of the receiver is greater
>for coherent dectection. Modern electronics may make this cost negligible
>but I am not sure on that point.

I don't believe SSB would be practical for music, although we know
from practical experience that it is acceptable for speech.  The
reason is that music puts very strict requirements on frequency
ratios.  Consider a note from a musical instrument.  With rare
exceptions, which technically are called 'noise' (e.g. a sound from a
vibrating circular membrane without benefit of a resonator: a bass
drum, but not a tympani), all musical sounds have harmonics which are
exact integral multiples of the fundamental frequency.  Now, if you
are receiving SSB and the injected frequency is not spot on, you will
disturb this integral relationship.  For example, consider a 400 herz
note with a second harmonic at 800 herz.  Now, if we have a 1 megaherz
carrier and consider upper sideband, the fundamental will be 1,000,400
herz and the second harmonic will be 1,000,800 herz.  Now make the
signal a SSB signal.  The carrier goes away.  Back at my receiver, I
inject a carrier that is off by one hundreth of a percent, namely
1,000,100 herz.  Now, the two frequencies which I will detect will be
300 herz and 700 herz; the upper note is about 14% too sharp to be a
second harmonic of the lower note.  To understand what 14% means: a
musical half-step is a frequency change of about 6%.
-- 
Herb Kanner
Apple Computer, Inc.
{idi,nsc}!apple!kanner
kanner@apple.com