Cut-off frequency

[Arnaud Froidmont]

Hello,

I'm currently working on 3d coordinates of sprinters and I would like to smooth the data. I'm using a second order Butterworth filter with the Challis (1990) method to find the cut-off frequency.

The program I am using accepts cut-off frequencies between 0 and half of the Nyquist frequency. In the literature (Winter 1990), I found that the frequency can be between 0 and the Nyquist frequency. I believe taking half of it is a precaution but does anyone know the actual reason for this, or has any references talking about it.

Thanks a lot for your help

Arnaud

[Edmund Cramp]

Re: Cut-off frequency

I would recheck the program documentation - I suspect that it allows the filter frequency to be set at any value up to half the sampling frequency which makes more sense. Of course the documentation could be wrong so you could just try filtering the data at just below half the sampling frequency and see what happens.