Page 1 of 1

Differences between decimater utility and Python

Posted: Fri Feb 22, 2019 7:18 pm
by dgfitch
The decimater utility seems to be doing something beyond simple decimation, but we can't find further documentation. When we compare its output to equivalently decimated files ran through Python libs with a fifth order sinc filter, things look noisier. Almost like the decimater is doing an extra smoothing pass, beyond the documented sinc filter.

I've uploaded some comparisons: http://brainimaging.waisman.wisc.edu/~dfitch/biosemi/

For example, comparing inside BDFReader:

The "decimater" output:

Image

Python output:

Image

Re: Differences between decimater utility and Python

Posted: Sun Feb 24, 2019 10:59 pm
by Coen
The decimator tool does just apply a 5th order cascaded integrator-comb (CIC) filter, no extra filtering/smoothing. Inspect the (open-source) LabVIEW code for confirmation.

Best regards, Coen (BioSemi)

Re: Differences between decimater utility and Python

Posted: Mon Feb 25, 2019 4:04 pm
by dgfitch
Thanks Coen.

You may want to update the downloads page which just says "fifth order sinc filter" and doesn't mention CIC.

Re: Differences between decimater utility and Python

Posted: Mon Feb 25, 2019 5:24 pm
by Robbie303
The information on the website describing the filter has been updated:
"A fifth order cascaded integrator-comb (CIC) filter"