Most receivers and audio gear have lower noise floor than a spectrum analyser. So it's plausible.
How could they have determined this noise floor then?
Don't forget this 350 Euro ground wire.
Well, to determine the noise floor at baseband, they would have done a conventional signal/noise test,
similar to that which we did as a normal routine, except they would use a calibrated standard modulator.
That said, I don't think the noise floor is that spectacular with any intercarrier sound receiver.
I think the main problem with the SA was lack of resolution at the thirty odd MHz IF frequency, &, for the baseband case, running out of range at the low end of its coverage.
The FFT range on a DSO may have done the trick, but such Devices were only in their infancy at that time.
In both the cases I related, there was no question of the problem being something that could only be heard by the "gifted".
They were real faults which were not immediately apparent in the relatively limited time available to busy Technical staff, but did appear to an enthusiast listening in a good monitoring situation.
In both cases, the people concerned were staff members, not random audiophiles "off the street".