Suspected line fault, how to contact Sky/BT?

My family has had connectivity problems for about a year, with high

High line attenuation (50-56db downstream) leading either very frequent disconnects or very slow speeds. Currently reliable but 1Mbit download only. 1 year ago we could get reliable 5Mbit download.

The main symptom that suggests a line fault is that noise from the ADSL modem is audible in phone line. This issue has been there to some extent for a long time but may have got worse.

This is tested from the master socket with everything downstream disconnected, tested with two different Billion/Netgear routers and different ADSL filters. So the problem must be before the master socket.

We have already called out a BT phone engineer but he was only interested in the phone system and there aren't significant problems with phone calls when internet equipment is disconnected and his tests didn't show major problems. He knew nothing about internet and networking (to the extent that he suggested the fact that we have a wired Ethernet network was causing the problem and that we should go wireless only to fix it!).

We pay BT for phone and line rental and Sky for unlimited broadband. I understand we need an BT openreach engineer but should we contact Sky or BT and how?

I think it will be a difficult problem for them to fix in our neighbourhood so I just have a feeling they will try to avoid fixing it one way or another. My family is not technically competent so if I am not there they may be easy to trick. But maybe I am being too cynical and BT openreach engineers are reliable and competent?

Thanks very much for any advice!

Most likely a high resistance fault, which would need a broadband qualified openreach engineer and that would have to be booked via Sky.

Another option is that the slow speeds are because of some occasional instability that the Sky DLM has slowed the line down to cope with, a full set of the line stats would help. Plus is the 50 to 56 dB a guess range, or does it vary across those values if so then that in itself is odd, attenuation is a largely fixed value.