I feel like it would work, mostly. But then, one day, you'll potentially be next to something emitting the wrong EMI/RFI and your rig will sound like a swarm of bees in a sawmill.
You can't bandpass everything because you'll have no signal so any noise that does find it's way there between 20-20k, it's going to be there.
I've seen interference do weird things in and out of the audio spectrum.
I had an unshielded cat5 hooked to a peripheral monitoring system for an MRI cause major image artifacts a few months ago. That single weak point leeched noise across the ground plane of the MRI magnet itself and wreaked havoc. Those systems have so many filter and fail-safes, you couldn't count them all and operate in faraday cages. But a new weakness was implimented that the system wasn't designed to handle.
Don't underestimate noise, I guess is what I'm saying.
Use high quality cables, cross AC at 180° and use filters.
This implementation may work fine 99% of the time(as did the 3rd party monitoring system) but when it fails, it won't be pretty.
Or maybe it works fine in the basement but at the one local club, it causes mayhem because there's a transformer or high voltage line under the stage...
It's possible, if done correctly, and a cool bar trick. But it's not anything I'd rely on.
Also, keep in mind, any noise will start at the begining of the chain and pass through every amplification stage in your chain. Even a tiny bit could be a major issue.
And, if the amp doesn't block DC properly due to some failure, 9v will probably punch the cones through the grill cloth, briefly. Until the voice coils melt and dead short. Then the output transformer goes...
Not worth not running an extra cable. Too many potential failure points.
Or, as I said earlier, make a 5 pin cable (shield being 1 pin). Drive it with a good buffer and use quality cable. Or ziptie two cable together. This is all a solution looking for a problem, to be frank.
