Aw heck, as long as we're talking about features that someone else would have to implement, a ducker would be even better.

For those who never worked in a studio or radio station, a "ducker" is a device that has two inputs. Normally, input A is just copied to the output. When a signal is present on input B, it gets added to some fraction of input A and presented to the output. The audible effect is that of the announcer (usually) talking, and the music "ducking" under his voice. It's usually implemented with a noise gate with separate control inputs, but it's also possible to code in DSP.