I am new to the field of DSP. I have a measured signal, and I wanted to apply a low pass filter on the data. I just simply applied FFT, wrote zeroes at the unwanted frequencies then applied inverse FFT. But I had got complex numbers instead of real values. What is the proper way of doing this kind of filtering? How to get back the filtered time signal?
Asked
Active
Viewed 112 times
0
-
4Does this answer your question? Why is it a bad idea to filter by zeroing out FFT bins? – MBaz May 09 '22 at 12:54
-
Did you get complex numbers with significant imaginary parts, or with imaginary parts that were teeny with respect to the real data (i.e., numbers like 0.5 - 1e-18i)? If you did get significant imaginary components, then edit your question to show which bins you zeroed out -- preferably graphically, or with math, not with a bunch of code. – TimWescott May 09 '22 at 14:20