2011-03-02 07:37:39 -05:00
|
|
|
/* -*- c-basic-offset: 4 indent-tabs-mode: nil -*- vi:set ts=8 sts=4 sw=4: */
|
|
|
|
|
|
|
|
/*
|
|
|
|
QM DSP Library
|
|
|
|
|
|
|
|
Centre for Digital Music, Queen Mary, University of London.
|
|
|
|
This file 2005-2006 Christian Landone.
|
|
|
|
|
|
|
|
This program is free software; you can redistribute it and/or
|
|
|
|
modify it under the terms of the GNU General Public License as
|
|
|
|
published by the Free Software Foundation; either version 2 of the
|
|
|
|
License, or (at your option) any later version. See the file
|
|
|
|
COPYING included with this distribution for more information.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#ifndef MATHUTILITIES_H
|
|
|
|
#define MATHUTILITIES_H
|
|
|
|
|
|
|
|
#include <vector>
|
|
|
|
|
|
|
|
#include "nan-inf.h"
|
|
|
|
|
2016-10-05 18:16:44 -04:00
|
|
|
/**
|
|
|
|
* Static helper functions for simple mathematical calculations.
|
|
|
|
*/
|
|
|
|
class MathUtilities
|
2011-03-02 07:37:39 -05:00
|
|
|
{
|
2016-10-05 18:16:44 -04:00
|
|
|
public:
|
|
|
|
/**
|
|
|
|
* Round x to the nearest integer.
|
|
|
|
*/
|
2011-03-02 07:37:39 -05:00
|
|
|
static double round( double x );
|
|
|
|
|
2016-10-05 18:16:44 -04:00
|
|
|
/**
|
|
|
|
* Return through min and max pointers the highest and lowest
|
|
|
|
* values in the given array of the given length.
|
|
|
|
*/
|
2011-03-02 07:37:39 -05:00
|
|
|
static void getFrameMinMax( const double* data, unsigned int len, double* min, double* max );
|
|
|
|
|
2016-10-05 18:16:44 -04:00
|
|
|
/**
|
|
|
|
* Return the mean of the given array of the given length.
|
|
|
|
*/
|
2011-03-02 07:37:39 -05:00
|
|
|
static double mean( const double* src, unsigned int len );
|
2016-10-05 18:16:44 -04:00
|
|
|
|
|
|
|
/**
|
|
|
|
* Return the mean of the subset of the given vector identified by
|
|
|
|
* start and count.
|
|
|
|
*/
|
2011-03-02 07:37:39 -05:00
|
|
|
static double mean( const std::vector<double> &data,
|
|
|
|
unsigned int start, unsigned int count );
|
2016-10-05 18:16:44 -04:00
|
|
|
|
|
|
|
/**
|
|
|
|
* Return the sum of the values in the given array of the given
|
|
|
|
* length.
|
|
|
|
*/
|
2011-03-02 07:37:39 -05:00
|
|
|
static double sum( const double* src, unsigned int len );
|
2016-10-05 18:16:44 -04:00
|
|
|
|
|
|
|
/**
|
|
|
|
* Return the median of the values in the given array of the given
|
|
|
|
* length. If the array is even in length, the returned value will
|
|
|
|
* be half-way between the two values adjacent to median.
|
|
|
|
*/
|
2011-03-02 07:37:39 -05:00
|
|
|
static double median( const double* src, unsigned int len );
|
|
|
|
|
2016-10-05 18:16:44 -04:00
|
|
|
/**
|
|
|
|
* The principle argument function. Map the phase angle ang into
|
|
|
|
* the range [-pi,pi).
|
|
|
|
*/
|
2011-03-02 07:37:39 -05:00
|
|
|
static double princarg( double ang );
|
2016-10-05 18:16:44 -04:00
|
|
|
|
|
|
|
/**
|
|
|
|
* Floating-point division modulus: return x % y.
|
|
|
|
*/
|
2011-03-02 07:37:39 -05:00
|
|
|
static double mod( double x, double y);
|
|
|
|
|
|
|
|
static void getAlphaNorm(const double *data, unsigned int len, unsigned int alpha, double* ANorm);
|
|
|
|
static double getAlphaNorm(const std::vector <double> &data, unsigned int alpha );
|
|
|
|
|
|
|
|
static void circShift( double* data, int length, int shift);
|
|
|
|
|
|
|
|
static int getMax( double* data, unsigned int length, double* max = 0 );
|
|
|
|
static int getMax( const std::vector<double> &data, double* max = 0 );
|
|
|
|
static int compareInt(const void * a, const void * b);
|
|
|
|
|
|
|
|
enum NormaliseType {
|
|
|
|
NormaliseNone,
|
|
|
|
NormaliseUnitSum,
|
|
|
|
NormaliseUnitMax
|
|
|
|
};
|
|
|
|
|
2016-10-05 18:16:44 -04:00
|
|
|
static void normalise(double *data, int length,
|
|
|
|
NormaliseType n = NormaliseUnitMax);
|
2011-03-02 07:37:39 -05:00
|
|
|
|
2016-10-05 18:16:44 -04:00
|
|
|
static void normalise(std::vector<double> &data,
|
|
|
|
NormaliseType n = NormaliseUnitMax);
|
2011-03-02 07:37:39 -05:00
|
|
|
|
2016-10-05 18:16:44 -04:00
|
|
|
/**
|
|
|
|
* Threshold the input/output vector data against a moving-mean
|
|
|
|
* average filter.
|
|
|
|
*/
|
2011-03-02 07:37:39 -05:00
|
|
|
static void adaptiveThreshold(std::vector<double> &data);
|
|
|
|
|
2016-10-05 18:16:44 -04:00
|
|
|
/**
|
|
|
|
* Return true if x is 2^n for some integer n >= 0.
|
|
|
|
*/
|
2011-03-02 07:37:39 -05:00
|
|
|
static bool isPowerOfTwo(int x);
|
2016-10-05 18:16:44 -04:00
|
|
|
|
|
|
|
/**
|
|
|
|
* Return the next higher integer power of two from x, e.g. 1300
|
|
|
|
* -> 2048, 2048 -> 2048.
|
|
|
|
*/
|
|
|
|
static int nextPowerOfTwo(int x);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Return the next lower integer power of two from x, e.g. 1300 ->
|
|
|
|
* 1024, 2048 -> 2048.
|
|
|
|
*/
|
|
|
|
static int previousPowerOfTwo(int x);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Return the nearest integer power of two to x, e.g. 1300 -> 1024,
|
|
|
|
* 12 -> 16 (not 8; if two are equidistant, the higher is returned).
|
|
|
|
*/
|
|
|
|
static int nearestPowerOfTwo(int x);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Return x!
|
|
|
|
*/
|
|
|
|
static double factorial(int x); // returns double in case it is large
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Return the greatest common divisor of natural numbers a and b.
|
|
|
|
*/
|
|
|
|
static int gcd(int a, int b);
|
2011-03-02 07:37:39 -05:00
|
|
|
};
|
|
|
|
|
|
|
|
#endif
|