filtering help

Hello

I am trying to filter out a big chunk of data. I have the data in a .csv file. When I read and plot that data, there is a lot of noise along with the pulses. The pulses are not periodic and staggered all over the place. I am getting this data out of an oscilloscope which is connected to the output of a log amp and don't know what frequency the signal is on. The input to the log amp is a 30MHz IF. The output is a base band signal. If I did use the filter command y = filter(b,a,X) how do I decide what the coefficients must be?

Thanks

Answer 1

[Rick Rosson]

I would recommend that you take a look at the signal in the frequency domain, and try to figure out whether the noise is primarily at frequencies that are lower or higher than the signal of interest. Also, is the signal of interest limited to a certain frequency range or band.

Based on this information, you may be able to design an appropriate filter (or sequence of filters) to reduce the noise while retaining the signal of interest.

The easiest way to design a filter initially is to use the Filter Design and Analysis tool or fdatool. For more information:

   >> doc fdatool

HTH.

Rick

[Walter Roberson]

I thought of looking at the fft, but then recalled that vsee specified pulses for the signals. Pulses are square waves or approximations thereof, and so are composed of many frequencies. Especially with the pulses not being periodic, it was not clear to me that an fft would provide useful information, especially not for filtering shot noise ?

[vsee]

True. Pulses are not periodic and therefore there is not set pri(pulse repetition interval).

[Rick Rosson]

Yes, all very true. Still, in order to figure out the type of filter and filter specs that are appropriate for this example, you will likely need to inspect the signal in the frequency domain. Obviously you can use the FFT for this purpose, but there may be other techniques (based on the FFT) that will provide more information. For example, you may want to use one of the PSD estimation techniques, the STFT, or the spectrogram. In any case, the FFT is probably a good place to start, and then you can decide what else you may want to try.