It contains a set of utilities to implement some digital signal processing. This is a wrapper around Intel Performance Primitives.

A copy of IPP installer can be downloaded from loccioni ftp.

In order to make it works, please add path of Intel library to PATH Windows system variable. For instance an IPP version it's located in:

C:\Program Files (x86)\IntelSWTools\compilers_and_libraries_2019.5.281\windows\redist\intel64_win\ipp

ATTENTION!

• It needs Intel processors. It's not tested for AMD.
• IntelSWTools platform (x64/x86) must be aligned with your executable platform.

An example of the simplest usage:

Butterworth filter = new Butterworth(
    order: 8,                // 8th order it could be enough :-)
    samplingFrequency: 1000, //Hz 
    cutoff: 300);            //Hz

var filteredData = f.Filter(sourceData);

filter.Dispose();

In order to improve performance, you could create the filter one time and apply it at each new data arrival.

Butterworth filter = new Butterworth(
    order: 8,                // 8th order it could be enough :-)
    samplingFrequency: 1000, //Hz 
    cutoff: 300);            //Hz

// later on a data arrival event that could occurs more time...
var filteredData = f.Filter(sourceData);

// on client dispose
filter.Dispose();

If each new data array is not time connected, you should reset internal filter state by calling Reset() method before every Filter call.

It's a simple benchmark to verify performance compared with the (already used and payed :-)) NationalInstruments.Analysis.Enterprise library included in NI Measurement Studio.

Input signal: 20000 samples.

Boundary conditions:

• BenchmarkDotNet=v0.11.5, OS=Windows 10.0.17763.805 (1809/October2018Update/Redstone5)
• Intel Core i7-4770S CPU 3.10GHz (Haswell), 1 CPU, 8 logical and 4 physical cores
  • [Host] : .NET Framework 4.7.2 (CLR 4.0.30319.42000), 64bit RyuJIT-v4.7.3468.0 [AttachedDebugger]
  • DefaultJob : .NET Framework 4.7.2 (CLR 4.0.30319.42000), 64bit RyuJIT-v4.7.3468.0

• Intel IPP dev reference.