It's probably a line level problem, the variance in noise would be due to the Optimod doing its thing.
You would be a brave man to suggest to the ABC audio people they had a problem with shielding or earthing in their studio.
The transmitter and studio are over 20Km apart so RF shielding is not normally an issue. It's more often a line level problem.
Do the ABC use Optimods? ------Ohh, how the mighty have fallen!
On the subject of the ABC:
When I worked for PMG/ Telecom Aust many years back, from time to time when doing the "morning line up" at 6WF/WN, we would sometimes notice that the lineup tone from the Studio was low in level.
Ringing up Master Control on the "Order wire", we would report it.
The answer was "It's alright leaving here", as we watched the VU meter move up to the correct setting!
It wasn't always them, though.
A bit later, I was in the Broadcast Service Centre, when we received a modulation monitor from a country station, which was reported by the local Phone Techs as reading very high, so we sent them a replacement by return freight.
On test, it was "spot on"!
Whilst ruminating about that, the Boss received a call, saying the replacement one was "just as bad", so, unimpressed, he sent myself & another guy off to check.
This was just before 5pm, so by the time we grabbed any useful gear, loaded the station wagon & headed off it was getting to be evening.
Arriving around 10pm, we found that the mod mon's meter was slamming against the stop, but the programme sounded OK.
Weird, so noticing a BWD CRO sitting in the disused workshop, we hung that across the RF monitor point in place of the mod mon.
One look at the mod envelope had us diving for the Desk fader------ the thing was cutting carrier horribly.
Waiting for the next "time pip" we set the level roughly, & waited for closedown.
From time to time, a scheduled lineup was done, using a local source.
8dB pads were switched into the desk VU meter, increasing their range by that amount, +16dBm was supplied from the local source, which then showed as 0VU.
Next, the limiters were bypassed, so that level was fed directly to the Tx input which was adjusted so it reached about 90% modulation.
The limiters were then restored to the chain, fed with the same level, adjusting them if necessary to 3 dB of limiting & their output level to produce the same approx 90% modulation.
The phone guys neglected to switch the 8dB pads in, set the local source to show 0VU on the desk.
This meant that the Tx were set for 90% mod at +8dBm, hence the overmodulation.-----Oops!
The weird thing was, that it sounded OK off air, & nobody seemed to notice the multiple harmonics we must have been radiating.
When I worked for TVW later in the day, we had a couple of community FM stations co sited with us, as part of our "good Corporate Citizen" policy.
One of them had a couple of Auxilary Sub Channels (SCA in Oz talk) as a "nice little earner".
The SCA folks reported interference, apparently from the main FM program, so we got dragged into it just to "stop the whingeing.".
Initially, we found the SCA people had screwed up with their levels, so were under-deviating badly.
This improved things markedly, but there was still "something there".
Winding the poor old, complaining 7L12 spec an down to the bottom of its frequency range, we could just see an out of band signal on the composite stereo coming up the link.
The Tech at the Community station eventually found it--- a switch set incorrectly on the Optimod.