Window Function for measuring Q values

You can ask questions here about Trevor and Gerard's exciting new book on Luthiery.

Moderators: kiwigeo, Jeremy D

Post Reply
User avatar
wadkin
Wandoo
Posts: 19
Joined: Tue Feb 25, 2014 10:08 am
Location: Colchester,Essex,England
Contact:

Window Function for measuring Q values

Post by wadkin » Fri Mar 07, 2014 8:34 am

Hi what would be the most accurate window function for measuring q values?
I'v been using the Hann (Hanning) function but I don't know if another one would be better?

Also I'm not sure whether is should average out the whole sample or what, I'v been using SpectruPlus and setting peak to 'infinite hold'

got some fairly meaning full results but it totally changes if I alter the settings so doesn't install me confidence, but I'm guess this is to do with natural limits of ffts.
I'm using sample size of 262144 sample rate of 96000(I think audacity default sample rate)

User avatar
Trevor Gore
Blackwood
Posts: 1605
Joined: Mon Jun 20, 2011 8:11 pm

Re: Window Function for measuring Q values

Post by Trevor Gore » Fri Mar 07, 2014 9:53 am

In my view the choice of window function is less important than setting the most suitable spectral line resolution. There's a discussion and guidance on setting spectral line resolution in Appendix A-I page AI-2 in the black book. There's a discussion on choice of window function here. As ever, the choice depends on what it is you're trying to do, but I think the limit of accuracy for measuring Q is going to be set more by the spectral line resolution than the window function at the lower end of the frequency range. At the high end of the frequency range you're better off trying to measure sound spectrum absorption in a way that is independent of the shape of the material sample, but don't ask me how to do it! Visual Analyser lets you choose no smoothing window, so one less thing to mess with your data.

User avatar
wadkin
Wandoo
Posts: 19
Joined: Tue Feb 25, 2014 10:08 am
Location: Colchester,Essex,England
Contact:

Re: Window Function for measuring Q values

Post by wadkin » Fri Mar 07, 2014 1:38 pm

Okay so I changed the sample rate to 22050 so with a sample size of 262144 that would give a resolution of 0.0841Hz surely that should be pretty good no?

now I'm getting values for Q over 400 hundred 0_o hmmm
I guess I should just choose a setting and stick with it so at least I get a relative comparison between boards.

Post Reply

Who is online

Users browsing this forum: GregHolmberg and 13 guests