alt Hacker NewsA difference that long cables make can be heard in this video: https://www.youtube.com/watch?v=SorO-QpqYRU .
Therein, audio from a microphone is sent through progressively-longer cables until the length reaches ~6 miles. It gets pretty muffled-sounding... eventually.
(The longest pair of wires I've sent analog audio through was in the realm of 37 miles, stretching across the countryside. AMA, I guess.)
Hah, that AM-receiving cable was at a theater and only a couple hundred feet long.
In general, with low-level analog audio and non-ridiculous lengths, additive noise effects are likely to become audible long before attenuation or especially frequency-dependent attenuation. As a decent heuristic, as long as the DC resistance is small compared to load impedance, the cable impedance is unlikely to be a problem. For the connection from the amplifier to the speaker, additive noise is unlikely to be a problem, so the DC resistance is even a decent heuristic: keep the round-trip resistance below half an ohm or so and you should be fine with most speakers.