PMAPs code confuses RMS and standard deviation

[atrettin]

When the rms attribute of a Peak is calculated, the code actually returns the standard deviation. The repr of the Peak prints out RMS: {self.rms / units.mus} µs, which is an incorrect labeling. The given quantity is actually the standard deviation. The RMS should correspond to the square of the standard deviation and have units of µs^2, accordingly.

I propose that the attribute rms should contain the square of the standard deviation, and that a separate std attribute is added to contain the standard deviation.

IC/invisible_cities/evm/pmaps.py

Line 42 in cda031c

self.rms = self. rms_above_threshold(0)

IC/invisible_cities/evm/pmaps.py

Lines 62 to 69 in cda031c

	def rms_above_threshold(self, thr):
	i_above_thr = self.pmts.where_above_threshold(thr)
	times_above_thr = self.times[i_above_thr]
	wf_above_thr = self.pmts.sum_over_sensors[i_above_thr]
	if np.size(i_above_thr) < 2 or np.sum(wf_above_thr) == 0:
	return 0
	return weighted_mean_and_std(times_above_thr, wf_above_thr)[1]

IC/invisible_cities/evm/pmaps.py

Line 83 in cda031c

RMS: {self.rms / units.mus} µs

[gonzaponte]

RMS stands for root mean square, i.e. the square root of the variance. RMS and std coincide when the average of the population is 0 and, in an abuse of the language, the terms are usually used interchangeably. Here we push these definitions a bit further by applying it to a waveform where time is the variable and the waveform amplitude its weight.

That being said, I agree that the correct term would be std and should replace RMS.

[atrettin]

Oh, this is actually a bit embarrassing, yes, you're right. I've removed the "bug" flag, because this isn't one. It would still be good to replace RMS with "std" to avoid this confusion, and while we are at it we could also add a "var" property containing the variance. Even though the difference is just a square root, I think that it would be useful because we have a direct linear relationship between the variance and the drift time, so the variance is physically more interesting IMO.

[gonzaponte]

Agreed.