OK, we'll I've spent a bit more time thinking about the whole automatic switch scenario, and I've uploaded a diagram at
http://www.btinternet.com/~avantg/schematic.jpg.

Assuming all the units being used have common L and R signal grounds, you should only need to switch the positive side of the signal, so only one DPDT relay is needed. D1 is included to try and eliminate any back EMF generated by the relay coil. The close proximity of the coil to the signal cable might mean that noise is induced into the signal cables, in which case, something like a 1uF cap strapped across the coil, in parallel with the diode ought to help. I've also included resistors in the Empeg signal feed to try and bring the level down a bit. My thinking is that, in series with the input impedance of the head unit, these ought to form a voltage divider so the head unit's input stage 'sees' about half the full 3.7v the Empeg can generate.

Needless to say, all of this is completely theoretical, and it'd probably be wise to run the whole lot with a dummy load in place of the head unit initially, checking the signal at points X and Y with a scope just to make sure you're not getting any significant DC content as the relay switches.

I'd be interested to hear any comments anyone might have about this...

Dominic
(Needless to say, all of this is totally unconnected with my employers, and any use is at the individual's risk!)

Cambridge Car Audio
http://www.cambscaraudio.co.uk