CR4® - The Engineer's Place for News and Discussion®

 Previous in Forum: Ms Excel Next in Forum: Voltage Drop When Loaded LM317T Regulated Power Supply
Member

Join Date: Jun 2017
Posts: 7

# Inverse Fourier Transform

06/04/2017 11:41 AM

Hi friends. İ have freq(Hz)-ampitude values and I want to plot time (s)-acceleration graph.how can I do this with using inverse Fourier function on excel?

Interested in this topic? By joining CR4 you can "subscribe" to

Guru

Join Date: Apr 2010
Posts: 5998
#1

### Re: inverse fourier transform

06/04/2017 2:38 PM

You cannot convert back to time series unless you also have phase information.

Member

Join Date: Jun 2017
Posts: 7
#2

### Re: inverse fourier transform

06/04/2017 2:43 PM

How can I calculate phase? Can you help me?

3
Guru

Join Date: Apr 2010
Posts: 5998
#8

### Re: inverse fourier transform

06/04/2017 5:45 PM

You cannot calculate phase from the amplitude spectrum. Let me explain with an example.

Say you have 1024 data samples. You run it through an FFT and get 1024 complex frequency values, i.e. 1024 real and 1024 imaginary values.

If you now take these complex frequency values through an inverse FFT, you would get your original 1024 sample time series.

To get the amplitude spectrum (which you have) you square each real and imaginary, add them, and take the square root. amplitude = sqrt(real2 + imag2). But from the amplitude spectrum, you cannot go back to get the real and imaginary components to do the inverse FFT.

To get the phase spectrum, you need the real and imaginary frequency components. phase = arctan(imaginary/real).

If you have both the amplitude spectrum and the phase spectrum, you could combine these to recreate the real and imaginary frequency components.

real = amplitude x cos(phase), imag = amplitude x sin(phase)

You could then run the inverse FFT to get back to time series.

Member

Join Date: Jun 2017
Posts: 7
#9

### Re: inverse fourier transform

06/04/2017 5:51 PM

Mr. Rixter can you send me an example xls file about your explains?

Guru

Join Date: Mar 2007
Location: at the beach in Florida
Posts: 20032
#3

### Re: inverse fourier transform

06/04/2017 2:45 PM

..." In Excel, on the "Fourier Analysis" window, select "Inverse" on the

bottom of the window. "...

__________________
Life is like riding a bicycle. To keep your balance you must keep moving. A.E.
Member

Join Date: Jun 2017
Posts: 7
#4

### Re: inverse fourier transform

06/04/2017 2:50 PM

I got inverse FFT of amplitude and I had complex numbers. How can I find acceleration in time by complex numbers ?

Guru

Join Date: Mar 2007
Location: at the beach in Florida
Posts: 20032
#5

### Re: inverse fourier transform

06/04/2017 2:59 PM
__________________
Life is like riding a bicycle. To keep your balance you must keep moving. A.E.
Member

Join Date: Jun 2017
Posts: 7
#6

### Re: inverse fourier transform

06/04/2017 3:22 PM

I mean I want to get "M" column from "K" column and plottime-acceleration-2graph in data as shown in this picture. How can I do that ?

Guru

Join Date: Mar 2007
Location: at the beach in Florida
Posts: 20032
#7

### Re: inverse fourier transform

06/04/2017 3:54 PM
__________________
Life is like riding a bicycle. To keep your balance you must keep moving. A.E.
2
Guru

Join Date: Apr 2010
Posts: 5998
#10

### Re: inverse fourier transform

06/04/2017 6:01 PM

No, nothing you do to your amplitude spectrum will recover lost information. See #8. When you combine real and imaginary to get amplitude, you only have half of the information.

It's synonymous to distance and direction. Think of real as north-south and imaginary as east-west. If real is 3 miles and imaginary is 4 miles, distance is 5 miles. But if you only know distance is 5 miles and don't know the direction, you can't find north-south or east-west.

Member

Join Date: Jun 2017
Posts: 7
#11

### Re: inverse fourier transform

06/04/2017 6:05 PM

http://dosya.co/a7e53b8akm4n/data.xlsx.html

This is my xls file

Guru

Join Date: Mar 2007
Location: at the beach in Florida
Posts: 20032
#12

### Re: inverse fourier transform

06/04/2017 6:08 PM
__________________
Life is like riding a bicycle. To keep your balance you must keep moving. A.E.
Guru

Join Date: Oct 2009
Posts: 1252
#13

### Re: inverse fourier transform

06/05/2017 8:35 AM

Anyone fancy doing a Fourier analysis of either the red line or blue line?

Off Topic (Score 5)
Guru

Join Date: Mar 2011
Location: Lubbock, Texas
Posts: 14345
#15

### Re: inverse fourier transform

06/05/2017 3:25 PM

__________________
If it ain't broke, don't fix it. Just build a better one.
Off Topic (Score 5)
Guru

Join Date: Oct 2009
Posts: 1252
#16

### Re: inverse fourier transform

06/05/2017 6:24 PM

I'll pass that on to my fellow church bellringers For the record, it is actually a standard "method" (piece of bellringing music) with the title of Cambridge Surprise Minor.

Off Topic (Score 5)
Guru

Join Date: Mar 2011
Location: Lubbock, Texas
Posts: 14345
#18

### Re: inverse fourier transform

06/06/2017 10:15 AM

To me, it is still pig DNA, or might as well be. I can't read bell-ringer music.

Why isn't that already automated? We want to automate everything else....just kidding.

BTW, are we talking about those who stand at a table with bells and ring out songs or those who pull on the ropes linked to the belfry?

__________________
If it ain't broke, don't fix it. Just build a better one.
Off Topic (Score 5)
Guru

Join Date: Oct 2009
Posts: 1252
#19

### Re: inverse fourier transform

06/06/2017 11:39 AM

Automation? Think of the parallel with an orchestra. We could automate that just as easily as a band of church bellringers, but it is actually an occupation where people meet and cooperate.

Church bell ringers stand in a belfry and pull ropes. The "music" is "change ringing" as in the piggy diagram, or "call changes."

Handbell ringers sit or stand with bells in their hands. Unlike church bell ringers, who almost invariably pull only one rope, handbell ringers generally have one in each hand. The music may be change ringing but is more often "proper" music.

Off Topic (Score 5)
Guru

Join Date: Mar 2011
Location: Lubbock, Texas
Posts: 14345
#20

### Re: inverse fourier transform

06/06/2017 12:31 PM

I wanna hear: "Rollover Beethoven."

__________________
If it ain't broke, don't fix it. Just build a better one.
Off Topic (Score 5)
Guru

Join Date: Mar 2007
Location: at the beach in Florida
Posts: 20032
#21

### Re: inverse fourier transform

06/06/2017 6:51 PM

Just like ringin' a bell....

__________________
Life is like riding a bicycle. To keep your balance you must keep moving. A.E.
Off Topic (Score 5)
Guru

Join Date: Mar 2007
Location: at the beach in Florida
Posts: 20032
#17

### Re: inverse fourier transform

06/06/2017 1:40 AM

__________________
Life is like riding a bicycle. To keep your balance you must keep moving. A.E.
Off Topic (Score 5)
Guru

Join Date: Dec 2008
Location: Long Island NY
Posts: 12442
#14

### Re: Inverse Fourier Transform

06/05/2017 12:28 PM

As many have pointed out, one cannot perform an inverse Fourier transform to return to the time domain from just the amplitude information. You also forgot to identify if this amplitude data is positional or velocity information.

__________________
"Don't disturb my circles." translation of Archimedes last words
Guru

Join Date: Mar 2011
Location: Lubbock, Texas
Posts: 14345
#23

### Re: Inverse Fourier Transform

06/08/2017 12:00 PM

I think I recall that Microsoft Excel FFT produces frequency and phase information columns, but I still would not count on any fidelity in the reverse transform.

__________________
If it ain't broke, don't fix it. Just build a better one.
Member

Join Date: Jun 2017
Posts: 7
#22

### Re: Inverse Fourier Transform

06/08/2017 12:08 AM

Interested in this topic? By joining CR4 you can "subscribe" to