I am attempting to generate a smooth audio signal such that peaks can be more clearly detected. A portion of an example signal is demonstrated in the image below. I would effectively like to generate a smooth waveform to eliminate false peaks like the second one indicated in red. Is the Savitzky-Golay filter appropriate for this use case, and if so, how can I determine an effective window_length and polyorder? The full signal is also attached in case it would make a difference to these parameters. Examples of false peaks I would like to eliminate are present after every sufficiently large peak.
