1-D digital filter runtime

Quick question - Is there any computational savings to padding input arrays for the filter function to a power of 2?

Answer 1

[Star Strider]

Not that I’m aware of. It would likely do the opposite. (It could also introduce artefactual ‘ringing’ in the output with a sharp transition to zeros.)

Also, use the filtfilt function if you have the Signal Processing Toolbox. It has the advantage of not introducing any phase distortion. The filter function does.

[Peter Cook]

Good deal. I was messing around with repmat() and padarray() for that padding, but filtfilt on an unpadded array seems to get me what I was looking for.

Thanks much.

[Star Strider]

My pleasure!