Files
lmms/include/lmms_math.h
DigArtRoks 5f4d0cab9b Fix for issue #3816 - FM or heavy PM in TripleOscillator makes outputs odd for some target waveforms. (#5651)
The internal waveforms of the class Oscillator produces the wrong amplitude when the input is a
negative phase. When doing PM or FM, negative phases may occur. When a negative phase is e.g. passed
to the the saw sample, it produces values less than -1.0, hence going out of range.

Converted all fraction calls to absFraction calls.

Removed the +2 in the function Oscillator::recalcPhase. The comment here was that it was needed to avoid
negative phases in case of PM. But by converting fraction to absFraction in the waveforms, negative phases
are not an issue anymore. On top of that the m_phase variable gains about 2 extra bits in precision.
As side effect of that, it improves the behaviour of the issue #2047 - TripleOscillator: Oscillators are getting out of sync.
Though I did not investigate it thoroughly over different notes and samplerates.

Add documentation to the fraction and absFraction functions in lmms_math.h as it was not immediately clear by the name what the
functions do. Correct the implementation of the functions in case the flag __INTEL_COMPILER is set. (floorf rounds always down).
2020-09-10 22:47:06 +02:00

332 lines
7.5 KiB
C++

/*
* lmms_math.h - defines math functions
*
* Copyright (c) 2004-2008 Tobias Doerffel <tobydox/at/users.sourceforge.net>
*
* This file is part of LMMS - https://lmms.io
*
* 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.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License for more details.
*
* You should have received a copy of the GNU General Public
* License along with this program (see COPYING); if not, write to the
* Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
* Boston, MA 02110-1301 USA.
*
*/
#ifndef LMMS_MATH_H
#define LMMS_MATH_H
#include <cstdint>
#include "lmms_constants.h"
#include "lmmsconfig.h"
#include <QtCore/QtGlobal>
#include <cmath>
using namespace std;
#ifndef exp10
#define exp10(x) std::pow( 10.0, x )
#endif
#ifdef __INTEL_COMPILER
static inline float absFraction( const float _x )
{
return( _x - floorf( _x ) );
}
static inline float fraction( const float _x )
{
return( _x - floorf( _x ) - ( _x >= 0.0f ? 0.0 : 1.0 ) );
}
#else
/*!
* @brief Returns the wrapped fractional part of a float, a value between 0.0f and 1.0f.
*
* absFraction( 2.3) => 0.3
* absFraction(-2.3) => 0.7
*
* Note that this not the same as the absolute value of the fraction (as the function name suggests).
* If the result is interpreted as a phase of an oscillator, it makes that negative phases are
* converted to positive phases.
*/
static inline float absFraction( const float _x )
{
return( _x - ( _x >= 0.0f ? static_cast<int>( _x ) :
static_cast<int>( _x ) - 1 ) );
}
/*!
* @brief Returns the fractional part of a float, a value between -1.0f and 1.0f.
*
* fraction( 2.3) => 0.3
* fraction(-2.3) => -0.3
*
* Note that if the return value is used as a phase of an oscillator, that the oscillator must support
* negative phases.
*/
static inline float fraction( const float _x )
{
return( _x - static_cast<int>( _x ) );
}
#if 0
// SSE3-version
static inline float absFraction( float _x )
{
unsigned int tmp;
asm(
"fld %%st\n\t"
"fisttp %1\n\t"
"fild %1\n\t"
"ftst\n\t"
"sahf\n\t"
"jae 1f\n\t"
"fld1\n\t"
"fsubrp %%st, %%st(1)\n\t"
"1:\n\t"
"fsubrp %%st, %%st(1)"
: "+t"( _x ), "=m"( tmp )
:
: "st(1)", "cc" );
return( _x );
}
static inline float absFraction( float _x )
{
unsigned int tmp;
asm(
"fld %%st\n\t"
"fisttp %1\n\t"
"fild %1\n\t"
"fsubrp %%st, %%st(1)"
: "+t"( _x ), "=m"( tmp )
:
: "st(1)" );
return( _x );
}
#endif
#endif
#define FAST_RAND_MAX 32767
static inline int fast_rand()
{
static unsigned long next = 1;
next = next * 1103515245 + 12345;
return( (unsigned)( next / 65536 ) % 32768 );
}
static inline double fastRand( double range )
{
static const double fast_rand_ratio = 1.0 / FAST_RAND_MAX;
return fast_rand() * range * fast_rand_ratio;
}
static inline float fastRandf( float range )
{
static const float fast_rand_ratio = 1.0f / FAST_RAND_MAX;
return fast_rand() * range * fast_rand_ratio;
}
//! @brief Takes advantage of fmal() function if present in hardware
static inline long double fastFmal( long double a, long double b, long double c )
{
#ifdef FP_FAST_FMAL
#ifdef __clang__
return fma( a, b, c );
#else
return fmal( a, b, c );
#endif
#else
return a * b + c;
#endif
}
//! @brief Takes advantage of fmaf() function if present in hardware
static inline float fastFmaf( float a, float b, float c )
{
#ifdef FP_FAST_FMAF
#ifdef __clang__
return fma( a, b, c );
#else
return fmaf( a, b, c );
#endif
#else
return a * b + c;
#endif
}
//! @brief Takes advantage of fma() function if present in hardware
static inline double fastFma( double a, double b, double c )
{
#ifdef FP_FAST_FMA
return fma( a, b, c );
#else
return a * b + c;
#endif
}
// source: http://martin.ankerl.com/2007/10/04/optimized-pow-approximation-for-java-and-c-c/
static inline double fastPow( double a, double b )
{
union
{
double d;
int32_t x[2];
} u = { a };
u.x[1] = static_cast<int32_t>( b * ( u.x[1] - 1072632447 ) + 1072632447 );
u.x[0] = 0;
return u.d;
}
// sinc function
static inline double sinc( double _x )
{
return _x == 0.0 ? 1.0 : sin( F_PI * _x ) / ( F_PI * _x );
}
//! @brief Exponential function that deals with negative bases
static inline float signedPowf( float v, float e )
{
return v < 0
? powf( -v, e ) * -1.0f
: powf( v, e );
}
//! @brief Scales @value from linear to logarithmic.
//! Value should be within [0,1]
static inline float logToLinearScale( float min, float max, float value )
{
if( min < 0 )
{
const float mmax = qMax( qAbs( min ), qAbs( max ) );
const float val = value * ( max - min ) + min;
float result = signedPowf( val / mmax, F_E ) * mmax;
return isnan( result ) ? 0 : result;
}
float result = powf( value, F_E ) * ( max - min ) + min;
return isnan( result ) ? 0 : result;
}
//! @brief Scales value from logarithmic to linear. Value should be in min-max range.
static inline float linearToLogScale( float min, float max, float value )
{
static const float EXP = 1.0f / F_E;
const float valueLimited = qBound( min, value, max);
const float val = ( valueLimited - min ) / ( max - min );
if( min < 0 )
{
const float mmax = qMax( qAbs( min ), qAbs( max ) );
float result = signedPowf( valueLimited / mmax, EXP ) * mmax;
return isnan( result ) ? 0 : result;
}
float result = powf( val, EXP ) * ( max - min ) + min;
return isnan( result ) ? 0 : result;
}
//! @brief Converts linear amplitude (0-1.0) to dBFS scale. Handles zeroes as -inf.
//! @param amp Linear amplitude, where 1.0 = 0dBFS.
//! @return Amplitude in dBFS. -inf for 0 amplitude.
static inline float safeAmpToDbfs( float amp )
{
return amp == 0.0f
? -INFINITY
: log10f( amp ) * 20.0f;
}
//! @brief Converts dBFS-scale to linear amplitude with 0dBFS = 1.0. Handles infinity as zero.
//! @param dbfs The dBFS value to convert: all infinites are treated as -inf and result in 0
//! @return Linear amplitude
static inline float safeDbfsToAmp( float dbfs )
{
return isinf( dbfs )
? 0.0f
: exp10( dbfs * 0.05f );
}
//! @brief Converts linear amplitude (>0-1.0) to dBFS scale.
//! @param amp Linear amplitude, where 1.0 = 0dBFS. ** Must be larger than zero! **
//! @return Amplitude in dBFS.
static inline float ampToDbfs( float amp )
{
return log10f( amp ) * 20.0f;
}
//! @brief Converts dBFS-scale to linear amplitude with 0dBFS = 1.0
//! @param dbfs The dBFS value to convert. ** Must be a real number - not inf/nan! **
//! @return Linear amplitude
static inline float dbfsToAmp( float dbfs )
{
return exp10( dbfs * 0.05f );
}
//! returns 1.0f if val >= 0.0f, -1.0 else
static inline float sign( float val )
{
return val >= 0.0f ? 1.0f : -1.0f;
}
//! if val >= 0.0f, returns sqrtf(val), else: -sqrtf(-val)
static inline float sqrt_neg( float val )
{
return sqrtf( fabs( val ) ) * sign( val );
}
// fast approximation of square root
static inline float fastSqrt( float n )
{
union
{
int32_t i;
float f;
} u;
u.f = n;
u.i = ( u.i + ( 127 << 23 ) ) >> 1;
return u.f;
}
//! returns value furthest from zero
template<class T>
static inline T absMax( T a, T b )
{
return qAbs<T>(a) > qAbs<T>(b) ? a : b;
}
//! returns value nearest to zero
template<class T>
static inline T absMin( T a, T b )
{
return qAbs<T>(a) < qAbs<T>(b) ? a : b;
}
#endif