Fixed point number representation and basic math routines. The fixed point numbers are unsigned and occupy 32 bits, of which 16 for the whole part and 16 for the fractional part. More...

#include <stdint.h>

Defines
#define	NB_FRACTIONAL_BITS 16
	The number of fractional bits in the fixed point representation.
#define	FRACTIONAL_MULTIPLIER 65536.0f
	2 ^ (NB_FRACTIONAL_BITS)
#define	FXP_16U16_MAX UINT32_MAX
	The largest representable fixed point 16u16 number.
#define	FXP_16U16_ZERO 0
	The number 0 represented as a fixed point 16u16.
#define	FXP_16U16_ONE ((fxp16u16_t)1 << NB_FRACTIONAL_BITS)
	The number 1 represented as a fixed point 16u16.
#define	FXP_16U16_ONE_HALF (FXP_16U16_ONE >> 1)
	The number 0.5 represented as a fixed point 16u16.
Typedefs
typedef uint32_t	fxp16u16_t
	Fixed point 16u16 are internally represented as 32-bit unsigned integers.
Functions
uint16_t	fxp16u16_to_uint16 (fxp16u16_t x)
	Converts a fixed point 16u16 number into an unsigned 16-bit integer, by rounding to the nearest integer.
float	fxp16u16_to_float (fxp16u16_t x)
	Converts a fixed point 16u16 number into a float.
fxp16u16_t	fxp16u16_from_uint16 (uint16_t x)
	Converts an unsigned 16-bit integer into a fixed point 16u16 number.
fxp16u16_t	fxp16u16_from_float (float x)
	Converts a float into a fixed point 16u16 number.
fxp16u16_t	fxp16u16_from_float_s (float x)
	Converts a float into a fixed point 16u16 number, using saturated arithmetic.
fxp16u16_t	fxp16u16_mul (fxp16u16_t x, fxp16u16_t y)
	Returns the product of two fixed point 16u16 numbers, if that product is within the range of representable fixed point 16u16 numbers.
fxp16u16_t	fxp16u16_div (fxp16u16_t x, fxp16u16_t y)
	Returns the quotient of two fixed point 16u16 numbers.
fxp16u16_t	fxp16u16_div_uint16 (fxp16u16_t x, uint16_t y)
	Returns the quotient of a fixed point 16u16 number divided by an unsigned 16-bit integer.
fxp16u16_t	fxp16u16_square (fxp16u16_t x)
	Returns the square of a fixed point 16u16 number.

Detailed Description

Fixed point number representation and basic math routines. The fixed point numbers are unsigned and occupy 32 bits, of which 16 for the whole part and 16 for the fractional part.

Author:: Pieter Agten (pieter.agten@gmail.com)

Date:: 16 nov 2011

Definition in file fixed_point.h.

Function Documentation

fxp16u16_t fxp16u16_div	(	fxp16u16_t	x,
		fxp16u16_t	y
	)		`[inline]`

Returns the quotient of two fixed point 16u16 numbers.

Parameters:

x	the dividend
y	the divisor

Returns:: x / y

Definition at line 176 of file fixed_point.h.

fxp16u16_t fxp16u16_div_uint16	(	fxp16u16_t	x,
		uint16_t	y
	)		`[inline]`

Returns the quotient of a fixed point 16u16 number divided by an unsigned 16-bit integer.

Parameters:

x	the dividend
y	the divisor

Returns:: x / y

Definition at line 192 of file fixed_point.h.

fxp16u16_t fxp16u16_from_float ( float x ) [inline]

Converts a float into a fixed point 16u16 number.

Parameters:

x	the float to convert

Returns:: The fixed point 16u16 number closest to the given float, if the float is within the representable range of the fixed point 16u16 datatype (0 to (65536 - eps)). The result is undefined otherwise.

Definition at line 121 of file fixed_point.h.

fxp16u16_t fxp16u16_from_float_s ( float x ) [inline]

Converts a float into a fixed point 16u16 number, using saturated arithmetic.

Parameters:

x	the float to convert

Returns:: The fixed point 16u16 number closest to the given float, if the float is within the representable range of the fixed point 16u16 datatype (0 to (65536 - eps)). If the float is less than 0, the fixed point 16u16 representation of 0 is returned. If the float is larger than the maximum representable fixed point 16u16 value, FXP_16U16_MAX is returned.

Definition at line 139 of file fixed_point.h.

fxp16u16_t fxp16u16_from_uint16 ( uint16_t x ) [inline]

Converts an unsigned 16-bit integer into a fixed point 16u16 number.

Parameters:

x	the 16-bit integer to convert

Returns:: The fixed point 16u16 number of which the whole part is equal to x and of which the fractional part is 0.

Definition at line 107 of file fixed_point.h.

fxp16u16_t fxp16u16_mul	(	fxp16u16_t	x,
		fxp16u16_t	y
	)		`[inline]`

Returns the product of two fixed point 16u16 numbers, if that product is within the range of representable fixed point 16u16 numbers.

Parameters:

x	the first factor of the product
y	the second factor of the product

Returns:: (x * y) if that is within the range of representable fixed point 16u16 numbers, or an undefined value otherwise.

Definition at line 161 of file fixed_point.h.

fxp16u16_t fxp16u16_square ( fxp16u16_t x ) [inline]

Returns the square of a fixed point 16u16 number.

Parameters:

x	the number of which to return the square

Returns:: x * x

Definition at line 206 of file fixed_point.h.

float fxp16u16_to_float ( fxp16u16_t x ) [inline]

Converts a fixed point 16u16 number into a float.

Parameters:

x	the fixed point number to convert

Returns:: The float closest to the number represented by x.

Definition at line 94 of file fixed_point.h.

uint16_t fxp16u16_to_uint16 ( fxp16u16_t x ) [inline]

Converts a fixed point 16u16 number into an unsigned 16-bit integer, by rounding to the nearest integer.

Parameters:

x	the fixed point number to convert

Returns:: The integer closest to the number represented by x, in the form of a 16-bit unsigned integer. If the fractional part of x is equal to 0.5, it is rounded up.

Definition at line 82 of file fixed_point.h.

Defines

Typedefs

Functions

Detailed Description

Function Documentation