I've seen the low-end IC audio amp datasheets recommending a simple (polarized) electrolytic on signal input, even for a split supply circuit that doesn't DC-bias the input. For example the datasheet of TDA2030 does that.
I've actually built that circuit out of curiosity if the capacitor C1's anode might get (significantly) DC biased by the opamp input, but I've measured a rather insignificant 1.7mV in my circuit at C1's cathode. Looking at the datasheet, the max input current for this IC is 2uA, so going through 22kOhm would max at 4.4mV. So neither theoretically nor in practice is it enough to ensure that the electrolytic capacitor is always biased the right way on sound/AC signals. My guess is it doesn't matter because the AC swings of the typical audio line are below the 1.5V reverse voltage usually needed to make electrolytics blow up fast enough to be entertaining/detectable. Am I correct or is there something else going on?
I've also connected C1 the "wrong way around" (on purpose, as a 2nd experiment) in this circuit, to see if anything bad might happen, and... nothing.
I've even put on it the max AC signal my sound card can dish (which exceeds my phone etc.), at a respectable 1.7V RMS, which is way more than what's needed to make the TDA2030 output clip like crazy, because it has high input sensitivity. Still nothing measurable/observable happened to the input cap.