| Author |
 Topic  |
|
|
couturier
France
291 Posts |
 Posted - 05/19/2008 : 07:57:32 AM
|
Origin Version (Select Help-->About Origin): 8 Pro
Operating System: XP pro
There's a reviewer, for a submitted paper, asking me some more details about filters I used. I'm not so sure about the answers to give:
- type of filter: I guess FFT filters are IIR, right ?
- Filter order. I'm used to orders in butterworth, Tchebychev ... filters but don't know how to answer for Origin FFT filters.
- Phase shift of the filter.
|
|
|
Fay_Guo
China
Posts |
|
|
couturier
France
291 Posts |
Posted - 05/20/2008 : 03:56:58 AM
|
Thanks for the answer.
How about phase shift ? if they are ideal filter, is there absolutely no phaseshift ?
Is there a phaseshift for the low pass parabolic ?
|
 |
|
|
Deanna
China
Posts |
Posted - 05/21/2008 : 02:23:53 AM
|
Within the pass-band, an ideal low-pass filter is characterisized by linear phase. Within the stop-band, the phase can be regarded as meaningless. Therefore, we can say there is no phase shift. Take the ideal low-pass filter for example. When we input an ideal pulse (t), the response is
where wc=w*pi*fc and fc is the cut-off frequency we set in the dialog.
If the input is a shifted pulse 
the response will be:
As for the parabolic filter, there is no phase shift for frequencies less than the Pass Frequency. Within the stop band (where frequencies are greater than the Stop Frequency), the amplitude is zero and the phase is meaningless. Between the Pass Frequency and Stop Frequency, there is phase shift. We can calculate the inverse Fourier transform of the following equation:
where wc1=2*pi*fc1, wc2=2*pi*fc1; fc1 is the pass frequency which you set in the dialog and fc2 is the stop frequency.
The result is the response corresponds to an ideal pulse input. Then you can use this result to calculate the phase shift.
Deanna
OriginLab Technical Services
Edited by - Deanna on 05/21/2008 02:31:55 AM |
|
|
| |
Topic |
|
|
|
FFT filters