Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Re: Recovering the Original Signal After Match Filtering On Excel
0 views
Skip to first unread message
Bret Cahill
Feb 3, 2012, 5:05:12 PM2/3/12
to
This seems to work:
1. take the Fourier Transform (FFT) of the noisy signal and of the
kernel.
2. multiply the signal by the kernel [IMPRODUCT()]
3. Take the Inverse FFT (FFT^-1)
4. IMREAL()
5. The convolution will be inside out so cut it in half and splice
the edges to the center and the center to the edges.
6. Take the FFT
7. Take the square root [IMSQRT()]
8. FFT^-1
9. IMREAL
Is there a more elegant way to do this without returning to the time
domain for splicing step 5?
I can Email the Excel file.
Bret Cahill
1. take the Fourier Transform (FFT) of the noisy signal and of the
kernel.
2. multiply the signal by the kernel [IMPRODUCT()]
3. Take the Inverse FFT (FFT^-1)
4. IMREAL()
5. The convolution will be inside out so cut it in half and splice
the edges to the center and the center to the edges.
6. Take the FFT
7. Take the square root [IMSQRT()]
8. FFT^-1
9. IMREAL
Is there a more elegant way to do this without returning to the time
domain for splicing step 5?
I can Email the Excel file.
Bret Cahill
0 new messages