Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

How to shift a feature in frequency domain

Status
Not open for further replies.
L

lallu3003

Newbie level 4
Joined
Jul 10, 2012
Messages
5
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Activity points
1,321
hi
I am new in signal processing. I have a impulse response of system A which has cut off frequency at f. now i need to shift it to f2 and still the response should be a impulse response. mathematically i can do it by multiplying it with exp(2*pi*f2-f*t) but that does not give you the impulse response in the time domain. Is there any way to do it.
 

Hi,

I think you made a mistake in analysis. you said you have a impulse response of system A and cutoff frequency at f. It means the output is only for this system. If you pass the output through the inverse of the system you will get impulse in the time domain also.

But if you change the output of the system A and pass through the inverse of the same system, is it reasonable to get the same time domain result as previous one? I think you need to change the system to get the time domain impulse response.
 
rongo024 said:
Hi,

I think you made a mistake in analysis. you said you have a impulse response of system A and cutoff frequency at f. It means the output is only for this system. If you pass the output through the inverse of the system you will get impulse in the time domain also.

But if you change the output of the system A and pass through the inverse of the same system, is it reasonable to get the same time domain result as previous one? I think you need to change the system to get the time domain impulse response.
Click to expand...

thanks rongo024 for your response .i mean i need to change the cut off frequency from f to f2 and then need to find the time domain response. but time domain response should be a impulse response. i do not mean same response. it should look like similar. how could i do that?
 

Hi,

I think you need to do frequency scaling/time scaling. If you do shifting that means your dc frequency is shifted to f2-f.
 

No, dc freq is not shifted.

If you want to change your cut-off freq from f to 2*f for example, you just double your input signal sample rate.
Try to do this in MATLAB:
Initial values for FIR estimation using FDATool: 48kHz sample rate, LFP FIR, Fpass = 9600, Fstop = 12000, Apass = 1 dB, Astop = 80 dB.\
When you change frequencies to 96000-19200-24000, your filter coefficients will be exact the same. So it means when you double sample rate - you double cut-off (from 9600 to 19200 in the example). Nothing else is changed.
 

you multiplied f by 2 and got 2*f. this is scaling. right.
 

Yes, but dc remains.

So for scaling a passband from f to f2 all that needed is scaling input signal sample rate by (f2/f) ratio.
 

actually i want to shift it a small bit. so mathmatically i can multiply the time signal by exp(2*f*t) but it will give you some imaginary value. but time signals are always real . is there a way to do it so that i can have only real value.

thanks
 

thanks for the reply. yeah you are right. what i like to do is shift the everything in frequency domain and then do the inverse fourier transform to get the time domain for that spectra. Now initially the time domain do not have any phase information. but after ifft it is producing complex value . si now it come with magnitude and phase. is there any way to shift the spectra in frequency domain and get the time domain data in real (without the imaginary part). thanks.
rongo024 said:
Hi,

I think you made a mistake in analysis. you said you have a impulse response of system A and cutoff frequency at f. It means the output is only for this system. If you pass the output through the inverse of the system you will get impulse in the time domain also.

But if you change the output of the system A and pass through the inverse of the same system, is it reasonable to get the same time domain result as previous one? I think you need to change the system to get the time domain impulse response.
Click to expand...
 

Doing fft you will always get complex values. To get a real world data just reject the imaginary part (or sum the I and Q components like it is done in quadrature modulators - there will be sqrt(2) increase in magnitude).
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top