Filtering is the operation which removes or changes specific frequency components from the
signal. In SIGVIEW, you can simply apply a filter to the signal by defining frequency segment to
be removed, or the segment you want to leave in the signal. That way, you can create bandstop,
bandpass, highpass and lowpass filters.
For more powerful filtering method, see 3D-filter function.
When you choose Signal tools/Filters option from the main menu or press
toolbar button, a
dialog will appear where you can enter segment boundaries (in Hz) and determine if you want to
remove that frequency segment (bandstop), leave only that frequency segment (bandpass) or
you want to remove all frequencies up to (highpass) or above some frequency (lowpass).
For Bandstop and Bandpass filter types, you can also include all higher harmonics of defined
frequency segment. If the defined segment is [x,y], it will also include all segments [N*x, N*y],
There is also an option to use custom filter curves for signal filtering. For details, please see Custom filter curves section.
Advanced options include:
"Filter slope width" selection: If you choose a wider slope, the filter will have a slower transition
from bandstop to bandpass frequencies, but it will also have less unwanted frequency artifacts,
ripples in spectrum etc.
"Reduce bandstop amplitude for" value will determine which amount of energy will be reduced in
the bandstop part of of the signal. The default value, -140 dB, is equivalent to a full removal of the
Modifying filter properties
After pressing OK, a new window with filtered signal will appear. You can change filter’s
frequency boundaries or bandstop amplitude later by choosing Edit/Properties on filtered signal
window from the menu. You will be able to see instantly how these changes affect the filtered
Background information & filter quality
SIGVIEW uses FFT-based filtering algorithm. It is a combination of the FFT calculation,
modifications in the FFT result (filling some frequency bands with zeros or applying custom filter
curve) and the inverse FFT calculation. This method is usually faster then other filtering methods,
but is also quite accurate.
The filter will deliver much better results if it can access a part of the signal before and after the
time segment you are actually filtering. If this is not provided, you will sometime observe signal
artifacts at the beginning and at the end of the filtered signal.
Therefore, it is recommended not to filter the whole available signal but to zoom-in to a smaller
part of the signal and then apply the filter function.