Audio comes from a output of a transformer from my audio interface. I use the schottkey diode to remove the negative half of the waveform, since I don't need it. Then I use the LM358 as a lowpass filter (the 10k pot controls the cutoff frequency from about 100 Hz to 500 Hz) and the 50k pot controlls the gain from about 1 to 100. I can tweak the gain knob to position the positive half of the audio waveform at just the right level for the 555 operating as a Schmitt trigger to convert any of the low-pass-filtered audio waves exceeded 2/3 of VCC into clean square waves of the same frequency.
hmm...stumbled upon this tidbit that the original variant of the timer ("NE555") has a 0.1% timing drift per volt of power supply:
https://components101.com/articles/lm555-vs-ne555-timer-ics
That could make a perceptible intonation issue. So if say want to generate a 100 Hz square wave, which has a period of 10 milliseconds, then 0.1% timing drift could make for a square wave's period be as low as 9.99ms or as high as 10.01 ms, which corresponds to frequency of 100.1 Hz to 99.9 Hz or +/- 2 cents. But the supply voltage probably doesn't change that much.