Re: [Phys-L] Statistics question

[Bernard Cleyet <bernardcleyet@redshift.com>]

On 2012, Aug 24, , at 15:30, curtis osterhoudt wrote:

> Depends on who 'we' is, and what you mean by 'generally'. As someone concerned with the quality factors of resonators (of any sort, be they electromagnetic, acoustical, chemical, feedback systems in social networks, etc.), FW/HWHM is used much more often. In distributions which differ significantly from Gaussian, standard deviations often don't even make sense to talk about. 


And then there's HWB.



Then, Q is related to fn,
f1, and f2 by the HPB formula [1]
Q ~ fn/ f2 − f1 = fn/ delta f
.

Because the HPB formula is so convenient, it is tempting
to apply it as if a stable half-power amplitude does exist at
the jump-down frequency


968 JOURNAL OF MICROELECTROMECHANICAL SYSTEMS, VOL. 20, NO. 4, AUGUST 2011
Measuring Quality Factor From a Nonlinear
Frequency Response With Jump Discontinuities
Wyatt O. Davis, Member, IEEE