Phase Shifting a Time Series in Frequency Domain

[mthapar]

I am trying to shift the phase of time series by a constant angle. The time series does not appear to change except for a phase shift of 90 deg where it is simply opposite polarity (sign) as compared to the input. The phase shift of 90 deg works fine using Hilbert transform.

Here is the code I am using:
ang:=90
shf:=PI*ang/180
FTS:=CFFT(Sct)
amps:=mag(FTS)
phs:=phase(FTS)
phscor:=phs+shf
FTS2:=makeri(amps,phscor)
Sct90:=Re(ICFFT(FTS2))


The Hilbert transform command:
HSct:=hilbert(Sct)
This works fine.

I would appreciate any help.

Thanks,

Mangat

[RichardJ]

Could you post the worksheet.

Richard

[PhilipOakley]

A fixed delay is equivalent to a linearly increasing phase shift (with frequency).

A fixed phase shift will result in a 1/frequency time shift in the signal. The reciprocal term will be dominated by the lowest frequency component.
Try making the amplitude have a similar reciprocal profile (up to some frequency) to see more of an effect....
Philip Oakley

[ptc-1368288]

No work sheet ?



jmG

[ptc-1368288]

What has Hilbert to do in there ?

From what you are attempting, is seems to be a rotation.
We have the program to rotate the spectrum.
What you call "shift" is probably what I call "rotate".



jmG