Another important factor is that the input to the vertical amplifier in the
scope is taken from the AGC buss of the receiver, and not from the audio
section. This is for the same reason, the audio response of the receiver
at frequencies of 15 Hz or less is poor, just as your hearing is poor at such
low frequencies. It is _recommended_ to use a DC-coupled oscilloscope. The
original published work on SAH's was done (ca. 1965) by Gordon Nelson using Dumont
304 scopes with DC vert coupling and with scopes using P7 phosphor, so the
horizontal sweep rate could be reduced to maybe 1 or 2 sweeps/second, if needed.
This technique worked best when there were just 2 signals present, so that
the difference would be clearly seen as a single "beat" wave. Also the AGC time
constant could be modified for better LF response. Fourier analysis could be
used to identify multiple sine waves present.
If you were looking at two signals, a few Hz apart, and at similar strengths,
you could hear a "whoosh-whoosh-whoosh" as the two signals beat together.
Nelson had discovered that two signals, a few Hz apart, *and differing by
maybe => 30 dB*, could both be detected by looking for the very small displacement
in the AGC voltage, when it was too small to be detected audibly. Since Nelson
spent time looking for secondary European signals fading in underneath
much stronger big signals, this was useful to him. He used published data
from the EBU monitoring stations that listed European carrier freq's to 1 Hz
resolution as a guide to identifying signals too weak to produce audio. Note,
too that freq tolerances back then were much looser than today.
The existance of a weak "beat" was a precursor indicator of the possibility
of hearing audio as the signal hopefully faded up to readability. Getting an
audible signal was always the goal. This was a guide to where to spend your
DXing effort, as a SAH display of an audible SAH beat is sort of superfluous info.
There is no reason this technique couldn't be done today, except for strong
artifacts from sideband splash from signals 2 to 5 kHz away, masking
the weak SAH trace, being so much worse today than we had in the 1960's.