smooth dft

subsidiary question: how can I set one dft wave so that it is directly ploted in Continuous Line Type instead of Spectral Type when I double click on it in ADE-XL ?

I suspect you really want to use psd instead of dft in that case, to compute a Power Spectral Density. If you want a moving average function, you could use article 11613392 (although I've never tested this on non-time domain data), and article 11609072 gives a way of changing the plot style so it's continuous for example.

Andrew.

Thanks Andrew,

Indeed for the psd, but it is not a big deal to convert the dft to psd by a simple multiplication.

Obviously the verilog-A function you mention computes the moving average of a signal over time; but in my case I already have a dft and would like to smooth it, by computing the average of the 10th previous points each 10 points...

The psd function isn't just a multiplied dft. It allows you to average a number of smaller windows over a larger number of samples - please see the documentation for psd:

Mathematically, it is defined as the Fourier Transform of the auto correlation sequence of the time series (signal). The waveform is first interpolated to generate evenly spaced data points in time. The spacing of the data points is the inverse of the dft sampling frequency. The psd is computed by first breaking up the time interval into overlapping segments. Each segment is multiplied, time point by time point, by the specified windowing function. The dft is performed on each windowed segment of the baseband waveform. At each frequency, the dfts from all segments are averaged together and the squared modulus of these averages gives the psd.

The function I pointed to is not a Verilog-A model - it's a calculator function which works over the x-axis and so probably doesn't really care whether it's time or not (the comments in the code say this). I've not tried doing it on frequency domain data, but I think it would work. The code was inspired by a Verilog-A model (the article references that) but apart from that, it doesn't require you to do it on time domain data.

Andrew.

Thanks a lot Andrew,

the abMovingAvg function is quite effective! 

Best regards