Floyd posted, in part:
There is no difference at all between the audio response
necessary for SSB and AM. AM, because it has both sidebands,
will necessarily take up twice the RF spectrum for the same
audio response, but in fact 2.4KHz (400Hz to 2800Hz) is actually
*preferable* to higher fidelity audio response when the purpose
is voice communications. (Ma Bell did a bazillion studies on
this decades ago, so it is not exactly new information.)
---------------------
I'd like to make a few comments about this.
1. AM does not necessarily have two sidebands. The same "quality" is had with
SSB AM and DSB AM, and SSB AM uses half the spectrum.

2. If the transmitted audio passband is 400 to 2800 Hz (2.4 kHz), the receiver
passband still needs to be 2.8 kHz for correct demodulation. That is, the
demodulation carrier must have the identical relationship to the sideband that
existed at the transmitter.

3. During WWII the Bell system used as little bandwidth as 1.6 kHz in their
long-haul channels. The quality was very poor, but you *could* communicate.
Modern telecom channel bandwidth is about 200 Hz to 3450 kHz. Generally this
is called a 3500 Hz channel. In ham radio that bandwidth provides good
fidelity for speech. With anything less it becomes increasingly difficult to
recognize who is the speaker. Using narrow band receive filters (1.8 to 2.4
kHz) improves the ability to communicate by eliminating or reducing QRM/QRN,
but the quality is reduced by this.

4. Ma Bell's studies on fidelity got into the perceptual area of balance
between the low end and high end of the audio spectrum; Having lots of bass
without good treble, or good treble without good base did not sound good.
That's one of the big reasons that the low end of the telecom voice channels
don't start at 100 Hz or lower. I thought I had a copy of the study, but
couldn't find it.

I would like to see hams who like AM migrate to SSB AM and limit the audio
passband to 3500 Hz. For carrierless SSB 3500 Hz should also be the passband
limit.

Don