I'd have thought crackling audio which sometimes works is more likely to be a capacitor than a damaged track - which would normally just be audio on or off.
This is unlikely given the roles of the electrolytic capacitors in the circuit; which are AC decoupling on the supply lines (i.e. removing noise on the DC supply) and in the case of the 22µF ones, they provide AC coupling - essentially separating the DC bias of the op-amps from the audio to the output jacks. In this application, capacitors are an open circuit to DC and a short circuit to AC. And because of the very short short time constants involved (the audio frequency changes much faster than the capacitor can charge/discharge), then no, you don't need to use non-polarised capacitors or anything stupid which some people dream up for reasons unknown.
When capacitors leak, they usually don't stop working, it's often a reduction in capacitance; meaning the circuit will still mostly work normally. The leaking is caused by the internal liquid electrolyte becoming unstable over time.
What happens in reality is that the capacitors leak as everyone knows, which often damages something like the -12V supply to one or both op-amps by corroding a track or via open circuit. So instead of the op-amp supply being connected normally to +12V and -12V, it's then connected to +12V and nothing. What this means is that the negative-going output half of the AC waveform can't be correctly produced by the op-amp. Also it tends to ruin the DC bias of the whole IC. As a result, parts of the audio output waveform become clipped (missing), which users interpret as "crackling". Which is something I like on my pork roasts.
