Origin offers an FFT filter, which performs filtering by using Fourier transforms to analyze the frequency components in the input dataset.

There are six types of filters available in the FFT filter function: low-pass, high-pass, band-pass, band-block, threshold and low-pass parabolic. Low-pass filters block all frequency components above the cutoff frequency, allowing only the low frequency components to pass. High-pass filters work in the opposite way: they block frequency components that are below the cutoff frequency.

This tutorial will show you how to perform the low-pass and band-pass filtering using Origin's FFT filter.

What you will learn

This tutorial will show you how to:

Perform low pass filtering.

Perform band pass filtering.

Low-pass Filter

Start with an empty workbook.

Select Data: Import from File: Sound(WAV) menu item and import the file Origin 8 Message.wav located at <Origin EXE Folder>\Samples\Signal Processing\. Accept the default import settings.

Highlight column A(Y) and click the Line button on the 2D Graph toolbar to create a line plot.

This signal is a sound wave and it is already known that the high frequency components can be regarded as noise, and should be blocked. So we will use the Low Pass method in the FFT Filter tool to approximate the low frequency component for further analysis.

Make sure the line plot is active, then select Analysis:Signal Processing:FFT Filters to open the fft_filters dialog box.

Make sure the Filter Type is set to Low Pass.

Check the Auto Preview box to turn on the Preview panel:

The top two images show the signal in the time domain, while the bottom image shows the signal in the frequency domain after Fast Fourier Transform. The X position of the red vertical dot line indicates the cutoff frequency. By moving the vertical line horizontally, you can preview the comparison between the original signal and the filtered signal in real time, in the top part of this panel.

Move the vertical line to the X position of the peak amplitude (as in the image below). Note that human error may be introduced during this step, but it is acceptable since we only want to roughly filter the signal.

Click OK to apply the FFT filter to the original signal.

The signal after filtering will be added to the data plot of original signal. Select Graph:Speed Mode and turn off speed mode in this graph. The resulting graph should look like this:

In the resulting graph, we can see that the high frequency components are blocked by the Low Pass FFT filter.

Band-pass Filter

Start with a new workbook.

Click the Import Single ASCII button and import the file fftfilter3.dat located at <Origin EXE Folder>\Samples\Signal Processing.

Highlight column B and click the Line button on the 2D Graphs toolbar to generate a line plot.

With the graph active, select Analysis:Signal Processing:FFT Filters. This opens the fft_filters dialog box.

Check the Auto Preview box to enable the Preview panel.

From the plot of frequency domain (the image below), we can see that this signal has components at several different frequencies> Now we are going to get the component at about 300Hz. So we will use the Band Pass method.

Set the Filter Type to be Band Pass.

When Band Pass is chosen, there will be two vertical red lines in the preview panel, marking the Lower Cutoff Frequency and Upper Cutoff Frequency. You can similarly move these two lines and get the real time preview of filtering results in the top parts of this panel.

Enter the values of the lower and upper cutoff frequencies according to the image below:

As long as only the desired amplitude peak stays within the cutoff frequency range, the error in filtering could be considered as acceptable even though the values of upper and lower cutoff frequencies may slightly differ for different cases.

Click OK to execute filtering.

We obtain the components at a frequency of around 300 Hz after filtering.