dt_FFT – library MetaTrader 5

Real author:

klot

Library of fast Fourier transformation functions (FFT).

This library was first implemented in MQL4 and published in CodeBase on October 2, 2006.

The library has seven fast Fourier transformation functions:

1. FFT of a complex function (direct and inverse)
2. FFT of a real function (direct and inverse)
3. FFT of two real functions (direct only)
4. Fast discrete sine transformation
5. Fast discrete cosine transformation
6. Fast simplification with FFT
7. Fast correlation with FFT.

1. FFT of a complex function (direct and inverse)
void fastfouriertransform(double& a[], int nn, bool inversefft);

The algorithm makes fast Fourier transformation of a complex function defined by nn counts on a real axis. Depending on the parameters passed, both direct or inverse transformation can be executed.

Input Parameters:

• nn – Number of values of the function. Must be a power of two!!! The algorithm doesn’t check the correctness of the passed value.
• a – array [0 .. 2*nn-1] of Real. Values of the function. Elements a[2*I] (real part) and a[2*I+1] (imaginary part) correspond to the first value.
• InverseFFT – direction of transformation. True, if it’s reverse; False, if it’s direct.

Input parameters:

• a – result of transformation. For more information, see description at

2. FFT of a real function (direct and inverse)
void realfastfouriertransform(double& a[], int tnn, bool inversefft);

The algorithm makes fast Fourier transformation of a real function defined by n counts on a real axis. Depending on  the parameters passed, both direct or inverse transformation can be executed.

Input Parameters:

• tnn – Number of values of the function. Must be a power of two!!! The algorithm doesn’t check the correctness of the passed value.
• a – array [0 .. nn-1] of Real. Values of the function.
• InverseFFT – direction of transformation. True, if it’s reverse; False, if it’s direct.

Input parameters:

• a – result of transformation. For more information, see description at

3. FFT of two real functions (direct only)
void tworealffts(double a1[], double a2[], double& a[], double& b[], int tn);

The algorithm makes fast Fourier transformation of two real functions each of which is defined by tn counts on a real axis. The algorithm saves your time but performs direct transformation only.

Input Parameters:

• tn – Number of values of the function. Must be a power of two. The algorithm doesn’t check the correctness of the passed value.
• a1 – array [0 .. nn-1] of Real. Value of the first function.
• a2 – array [0 .. nn-1] of Real. Values of the second function.

Input parameters:

• a – The first function Fourier transformation
• b – The second function Fourier transformation (see details on web-site)

4. Fast discrete sine transformation
void fastsinetransform(double& a[], int tnn, bool inversefst);

The algorithm makes fast sine transformation of a real function defined by tnn counts on a real axis. Depending on the parameters passed, both direct or inverse transformation can be executed.

Input Parameters:

• nn – Number of values of the function. Must be a power of two. The algorithm doesn’t check the correctness of the passed value.
• a – array [0 .. nn-1] of Real. Values of the function.
• InverseFST – direction of transformation. True, if it’s reverse; False, if it’s direct.
  

Input parameters:

• a – result of transformation. For more information, see description at

5. Fast discrete cosine transformation
void fastcosinetransform(double& a[],int tnn, bool inversefct);

The algorithm makes fast cosine transformation of a real function defined by nn counts on a real axis. Depending on the parameters passed, both direct or inverse transformation can be executed.

Input Parameters:

• tnn – Number of values of the function minus one. Must be a power of two (for example 1024). The algorithm doesn’t check the correctness of the passed value.
• a – array [0..nn] of Real. Function values (for example, 1025).
  Peculiarity of the preparation of an array for function transfer:
  int element_count2=ArrayResize(array,tnn1+1); //For cosine !!!
  
• InverseFCT – direction of transformation. True, if it’s reverse; False, if it’s direct.
  

Input parameters:

• a – result of transformation. For more information, see description at

6. Fast simplification with FFT
void fastcosinetransform(double& a[],int tnn, bool inversefct);

Simplification. One of the functions is assumed as a signal. The second one is considered a response.

Input:

• Signal – a signal with which simplification is made. Array of real numbers, numbering of elements from 0 to SignalLen-1.
• SignalLen – signal length.
• Response – response function. It consists of two parts corresponding to positive and negative values of an argument.
  Response values in points from -NegativeLen to 0 correspond to array elements with numbers from 0 to NegativeLen.
  Response values in points from 1 to PositiveLen correspond to array elements with numbers from NegativeLen+ to NegativeLen+PositiveLen.
• NegativeLen – “Negative length” of a response.
• PositiveLen -“Positive length” of a response.
  A response is equal to zero beyond [-NegativeLen, PositiveLen].

Output:

• Signal – values of function simplification in points from 0 to SignalLen-1.

7. Fast correlation with FFT
void fastcorellation(double& signal[], int signallen,  double& pattern[],  int patternlen);

Input:

• Signal – array signal with which correlation is made. Numbering of elements from 0 to SignalLen-1
• SignalLen – signal length.
• Pattern – array pattern correlation of a signal with which we are looking for. Numbering of elements from 0 to PatternLen-1
• PatternLen – pattern length

Output:

• Signal – values of correlation in points from 0 to SignalLen-1. Visit the web-site for more detailed description.