Extracting a sinewave signal when characteristic phase and frequency are known

[tchetch]

Hi there,

i have this problem:
I have a sinewave signal contaminated by a large amount of noise. I already know the phase of the signal and its characteristic frequency, and I am searching for its magnitude.

If I do a basic Fourier transform, I would obtain the magnitude at its characteristic frequency, but the phase would be generally shifted. Does anyone know a way to extract the magnitude at its specific phase and frequency ?

Thanks in advance

 

[MATLABdude]

Filter out the noise and then do peak detect? Or just take the RMS of the signal, and calculate the peak value (adding or subtracting an offset as necessary)?

 

[Pumblechook]

I don't understand how a sine wave can have absolute phase?

 

[FredGarvin]

Is it truly noise or have you considered it could be an aliasing issue? Have you treid playing with the sampling rate?

 

[FredGarvin]

I was also looking through some old notes I have. Tom Irvine over at Vibrationdata.com published a quick paper regarding a method he wrote a small program for, called sinefind. The paper is entitled "Sine Function Identification and Removal." I can't post it here because it is Tom's work. It pulls out amplitude, frequency and phase by a curve fit method.

If you have any interest in vibration analysis, take a look over at Tom's website.

www.vibrationdata.com

 

[Topher925]

There is a type of numerical filter than can do this, but the name escapes me right now. Is the noise gaussian or normal?