Description

Components Overview
Units
Global functions
Data types
Examples

Components Overview

The library contains two components:

TdspIIRFilter - Implementation of the Infinite Impulse Response filter
TdspFFT - Implementation of Fast Fourier Transform

Units

dspIIRFilters.pas - contains the implementation of TdspIIRFilter component
dspFFT.pas - implementation of TdspFFT component
dspTypes.pas - contains the base types
dspComplex.pas - set of functions to work with complex numbers
dspWindow.pas - implementation of window functions

Global functions

Unit dspWindow

function dspGetWindowValue(Window: TdspWindow; Index, Length: Integer): TdspSingle;

This function returns the window value for specified window function at the Index position. The Length parameter specifies the window length.

procedure dspCalculateWindow(Window: TdspWindow; Data: PdspSingle; Length: Integer);

This procedure calculates the window values for specified Window function. The values will be stored in the Data parameter. The window length and length of data are specified in the length parameter.

procedure dspApplyWindow(Window: TdspWindow; Data: PdspSingle; Length: Integer);

This procedure applies the specified window to the data. The length parameter is the window length.

Unit dspComplex

function cmpxInit (Re: TdspFloat; Im: TdspFloat = 0.0): TdspComplex; overload;
This function converts Re and Im to complex

procedure cmpxInit (var Z: TdspComplex; Re: TdspFloat; Im: TdspFloat = 0.0); overload;
This procedure converts Re and Im to complex Z

function cmpxNeg (Z: TdspComplex) : TdspComplex;
This function returns negated value of complex number

function cmpxSqr (Z: TdspComplex) : TdspComplex;
This function returns the square of Z

function cmpxSqrt (Z: TdspComplex) : TdspComplex;
This function returns the square root of Z

function cmpxSqrtx (V: TdspFloat) : TdspFloat;
This function returns the square root of V. If V < 0 then function returns 0.

function cmpxExp (Z: TdspComplex) : TdspComplex;
This function returns exponent raised to Z (e^Z)

function cmpxExpj (V: TdspFloat) : TdspComplex;
This function returns exponent raised to jV (e^jV)

function cmpxHypot (Z: TdspComplex) : TdspFloat;
This function returns the hypotenuse of Z.

function cmpxArcTan2 (Z: TdspComplex) : TdspFloat;
This function returns the arctan of Z

function cmpxConj (Z: TdspComplex) : TdspComplex;
This function returns conjugate of Z

function cmpxMul (Z: TdspComplex; V: TdspFloat) : TdspComplex; overload;
This function multiplies a complex and a real

function cmpxMul (Z1, Z2: TdspComplex) : TdspComplex; overload;
This function multiplies two complex numbers

function cmpxDiv (Z: TdspComplex; V: TdspFloat) : TdspComplex; overload;
This function divides Z with R

function cmpxDiv (Z1, Z2: TdspComplex) : TdspComplex; overload;
This function divides two complex numbers

function cmpxAdd (Z1, Z2: TdspComplex) : TdspComplex;
This function adds two complex numbers

function cmpxSub (Z1, Z2: TdspComplex) : TdspComplex;
This function substracts two complex numbers

Data types

TdspIIRFilterOrder = 1..8
The order of IIR filter.

TdspIIRFilterKind = (fkButterworth, fkChebyshev, fkBessel)
The characteristics of IIR filter.

TdspIIRFilterResponse = (ftLowpass, ftHighpass, ftBandpass, ftBandstop)
The frequency response of IIR filter. Lowpass, Highpass, Bandpass or Bandstop, respectively.

TdspIIRFilterProc = function(V: TdspFloat): TdspFloat of object
The filter procedure.

TdspFFTPower = 2..15
The power of 2. Uses in FFT.

TdspWindow = (fwNone, fwHamming, fwHanning, fwBlackman, fwGaussian, fwWelch, fwParzen)
The window function.

TdspFloat = Extended

TdspFloatArray = Array[0..High(Integer) div Sizeof(TdspFloat) - 1] of TdspFloat
PdspFloatArray = ^TdspFloatArray

TdspSingle = Single;
PdspSingle = ^TdspSingle;

TdspSingleArray = Array[0..High(Integer) div Sizeof(TdspSingle) - 1] of TdspSingle
PdspSingleArray = ^TdspSingleArray

TdspIntArray = Array[0..High(Integer) div Sizeof(Integer)-1] of Integer
PdspIntArray = ^TdspIntArray

TdspIIRFilter

This component implements an infinite impulse response digital filter. Lowpass, highpass, bandpass and bandstop filters, with Butterworth, Bessel or Chebyshev characteristics, are designed using the bilinear transform.

property Kind: TdspIIRFilterKind
The characteristics of IIR filter. You can design the filter with Butterworth, Bessel or Chebyshev characteristics.

property Response: TdspIIRFilterResponse
The frequency response of the filter. You can design the filter with Lowpass, Highpass, Bandpass or Bandstop frequency response.

property Order: TdspIIRFilterOrder
The order of the filter.

property Frequency1: TdspFloat
property Frequency2: TdspFloat
The corner (–3 dB) frequencies. For lowpass and highpass filters, only Frequency1 is required. For bandpass and bandstop filters, both Frequency1 (the lower corner frequency) and Frequency2 (the upper corner frequency) are required. The valid values: 0 < Frequency1 < SampleRate / 2; 0 < Frequency2 < SampleRate / 2; Frequency1 < Frequency2.

property SampleRate: TdspFloat
The sampling rate of signal.

property Ripple: TdspFloat
The passband ripple in dB, meaningful for Chebyshev filter only.

Filter: TdspIIRFilterProc
Or function Filter(InputValue: TdspFloat): TdspFloat;
Processing input value according to filter parameters. The result is filtered value. Data can be processed in real time. For example, you can process the audio stream in real time.
Note: In some cases the filter is astable. This effect occurs when the value of Alpha1=Frequency1/SampleRate is small (less then 0.001) or the value of equation (Frequency2 - Frequency1) / SampleRate (for filters with bandpass and bandstop frequency responses) is small (<0.01).

procedure Reset
This procedure clears internal buffers.

function    GetFrequencyResponse(Frequency: TdspSingle): TdspSingle
This procedure returns a frequency response.

property ManualReset: Boolean
Set this property to true if you want to change the filter parameters (frequencies and filter kind) during data processing. Usually (ManualReset = false) when the filter parameters are changed the internal buffers are cleared. So you can get some distortions at the processed signal.

property Info: String
This property contains filter information: gain, input and output coefficients.

property LastError: String
This description of the error.

TdspFFT

This component implements a fast fourier transform and inverse fast fourier transform with single floating point precision.

RealIn, ImagIn: PdspSingleArray
Input data for FFT.

RealOut, ImagOut: PdspSingleArray
Output data.

procedure Clear
This procedure zeros the input and output data.

procedure FFT
This procedure does FFT.

procedure IFFT
This procedure does inverse FFT.

procedure CalculateMagnitudes
This procedure calculates the spectrum of signal.

property BufferSize2: Integer
A half of buffer size.

property Power: TdspFFTPower
A power of 2. The buffer size = 2^Power.

property BufferSize: Longint
The buffer size. The value must be a power of 2.

property Window: TdspWindow
The window function for input data.

Examples

Simple FFT demo - Using FFT to get spectrum of signal
This demo shows how to work with FFT component. The program generates the signal and fills the FFT buffer with the data. Then the program calculates FFT and spectrum of signal. And finally, it displays the input signal and spectrum using TChart component.

Simple IIR Filter demo - Setup IIR filter and filtering the signal in real time
This demo shows how to work with IIR filter component to process realtime signal. The program generates a signal. The every value of signal is passed through the filter. The input and output values are displayed on the screen. Change the parameters of the filter and you will see the changes of filtered signal on the chart.

 IIR Plot demo - Plot the frequency response of IIR filter
This demo shows how to plot the frequency response of IIR filter. You can change the parameters of filter and plot the frequency respomse on a chart.

Perfomance - Test the perfomance of IIR filter and FFT
This demo tests perfomance of calculations for IIR filter and FFT on your system.

Simple spectrum analyzer - Sound visualization during recording
This demo shows how to calculate the spectrum of recorded signal in realtime.

Simple player - Sound processing and visualization during playing wav-files
This demo shows how to build a timbre block, graphical equalizer and spectrum analyzer. The timbre block uses IIR Filter components. The graphical equalizer uses FFT components and FFT convolution method to filter signal. The spectrum analyzer uses FFT components. Please see the source code of demos for more details.