`idealfilter`

Applies an ideal frequency-domain filter to a timeseries.

Category: Language Fundamentals > Data Types

Syntax

tsout = idealfilter(ts, interval, 'pass')

Keeps the frequencies inside each interval.

tsout = idealfilter(ts, interval, 'notch')

Removes the frequencies inside each interval.

Description

The filter works in the frequency domain: the data are transformed with an FFT, the unwanted frequency bins are set to zero, and the inverse transform gives the filtered signal. The cut is perfectly sharp, which is why it is called ideal, but also why it can ring near abrupt changes.

Input arguments

ts: timeseries object.
interval: Two-column matrix of frequency bands [low high], in cycles per time unit of the series.
'pass' | 'notch': Whether the bands are kept or removed.

Output arguments

tsout: Filtered timeseries with the same time vector.

Example

Low-pass of two tones

t = (0:7)';
y = sin(2*pi*0.05*t) + sin(2*pi*0.4*t);
f = idealfilter(timeseries(y, t), [0 0.2], 'pass');
disp(round(f.Data' * 1000) / 1000)

Output

   -0.4450   -0.4260   -0.1580    0.2030    0.4450    0.4260    0.1580   -0.2030

idealfilter