I have chosen a broadband time function (a double exponential) that I can transform to frequency analytically. Then I can analytically transform back to time with a residue integral. I wanted to see if I could use the FFTPACK on such broadband signals. Attached is a PDF of the worksheet along with the SMATH worksheet. Auto Calculate is turned off because the runtime is about 5.6 minutes.

1. My normalization of the frequency function is arbitrary to get the final results to compare with the original time function. Any guidance would be appreciated.

2. Discounting my normalization issue, the ifft not good. Have I made an obvious error?

3. Am I asking too much for the FFTPACK to work well with such broadband signals?

FFT TO COMPARE WITH FILON.pdf (721 КиБ) скачан 39 раз(а).
FFT TO COMPARE WITH FILON WORKING.sm (142 КиБ) скачан 33 раз(а).

You may recall that I also developed a Filon Fourier Transform Worksheet that works well with such broadband signals. I just wanted to compare the FFT.

Thanks,
Reg