Moving RMS calculation

[ebastler]

The signal is audio.
The RMS value will obviously be for a set period, likely 1-10 ms (TBD)

But that's an RMS averaging time which is less than the period of many of your audio components?! I don't think that will give you meaningful results.

I assume you have a mix of frequency components in your signal, hence can't average over exactly one period length. In that case you will need to average over "many" periods of the slowest signals to get a stable RMS readout. Or use the exponential decay approach, but again with a time constant that is significantly larger than the period of the slower signal components.