I gather from the debate that nobody here truly appreciate how much of an engineering marvel "carrier-frequency telephony" was back in analog days.
Here are some info to try to put it in proper perspective:
The "L5" system, which for all practical purposes was the final and ultimate in carrier-frequency transmission before everything went digital, modulated 10,800 telephone channels, each 4kHz wide, onto a single coax-cable, and transmited the signal in "toll grade" quality, which includes 44dB S/N ratio, from one side of USA to the other, or rather: up to 4000 US miles.
This required a repeater every single mile, literally 4000 repeaters one after the other, each amplifying the signal 25dB to compensate the loss of one mile of telco-grade coax. Well, "approx 25dB" because the attenuation depends on the temperature...
And yes, that means the total amplification from coast to coast was 100,000 (a hundred thousand) dB at the top frequency near 60 MHz.
Now, imagine if those 4000 amplifiers due to some accident of design or manufacturing all had 0.01dB dip or bulge at the same frequency ?
Or imagine if the coax had a little nick every N meters, due to a piece of debris on a pulley wheel in the factory ?
Ohh, and since the total frequency range is 1.5 to 60 MHz, better watch those I2 and I3 distortion products.
And did I mention that the repeaters live in a damp hole in the ground or boltet to the top of a tall wooden pole, out in the middle of nowhere ?
*Now* you can start to appreciate why HP, W&G and others built instrumentation to those insane specs, in particular with the ability to measure dB to part of thousands.
The HP3336 and HP3586 for example were built as a matched pair, directly focused on the L4 Carrier systems "only" 3600 telephone channels. They were some of the first microcontroller controlled instruments ever, which made it possible to measure all 3600 channels in a matter of minutes.
For more information on L5 see:
https://archive.org/details/bstj53-10-1901 (And the entire rest of the BSTJ december 1974 issue on archive.org)