There are some detailed requirements for ATC communications systems in the UK which are not found in international standards. The UK is sometimes accused of over-engineering things and in this case I think it is with good reason. Amongst other things, there is a requirement (I think, possibly it's a strong recommendation only) for headsets to be available. When headsets are used, the normal operational mode is that when RTF transmissions from aircraft are received they are heard in both ears, if a landline is open then the RTF transmissions are in one ear and the landline in the other. When a landline is open, when the controller talks the person on the other end of the landline(s) hears what is said; this includes anything said while the controller is transmitting to an aircraft. In the old days, when transmitting, the controller also got a low-volume feed into the headset, usually called a sidetone, from a receiver separate from the rest of the comms system. This allowed the controller to detect whether his/her transmissions were actually being broadcast and also to pick up instances of cross-transmission where the controller transmitted at the same time as another station (because they would hear the whine as the two signals interfered with each other).
With the introduction of some VoIP VCCS delays in broadcasting a controller's transmission mean that an off-air sidetone can produce a disconcerting 'echo' in the headset and the sidetone can be derived from the VCCS at a point close to the transmitter - if I understand things correctly, this will continue to provide a nice warm feeling to the controller but, in fact, provides no confirmation that the transmission actually went out and no protection from crossed transmissions which, potentially, could go completely undetected. I have no involvement in the implementation of any such systems but I'd be interested in any thoughts (or corrections to my current understanding) from Gonzo or anyone else with hands-on experience with modern systems.