mmath.c

#include "mmath.h"

/*
* References
* https://www.nullhardware.com/blog/fixed-point-sine-and-cosine-for-embedded-systems/
* http://www.coranac.com/2009/07/sines/
*/

/*
* Trig constants (2 pi isn't actually 2 pi because it is 0, but the constant is still there)
*/
#define PI  0x8000
#define HPI 0x4000

/*
* tan
*
* @desc
*   Calculates a fixed point tangent output
*   using a 9th degree maclaurian series
*   of tangent
*   The domain is [0, pi / 2], if x is out of the domain
*   it will get converted into the domain by 90-degree rotations
* @param
*   16-bit angle
*   2^16 - 1 is ~ 2 * pi
* @return
*   16-bit fixed point result
*   Higher byte is integer part
*   Lower byte is fractional part
*/
int16_t mm_tan(int16_t x)
{
  /*
  * 1. Convert the input to [0, pi / 2]
  */
  while (x > HPI) x -= HPI;
  return 0;
}

/*
* cot
*
* @desc
*   Because of the restrictions of tan,
*   cot is just tan mirrored about x = pi / 2,
*   so cot just converts the input and calls tan
* @param x
*   Same as tan
* @return
*   Same as tan
*/
int16_t mm_cot(int16_t x)
{
  /*
  * 1. Convert the input to [0, pi / 2]
  */
  while (x > HPI) x -= HPI;
  /*
  * 2. Call tangent
  */
  return mm_tan(HPI - x);
}

/*
* trig tables (mm_[trig]t)
*
* @desc
*   Taylor stuff is hard to figure out, I'm gonna sacrifice
*   ~ 32Kb instead for each table (the co-tables are just
*   mirrors of the other tables in the domain of the functions)
* @param x
*   Must be less than HPI
*/
#include "trig_tables.h"
uint16_t mm_tant(uint16_t x)
{
while (x > HPI) x -= HPI;
return mm_tan_table[x];
}

uint16_t mm_cott(uint16_t x)
{
while (x > HPI) x -= HPI;
return mm_tan_table[HPI - x];
}


uint16_t mm_sect(uint16_t x)
{
  while (x > HPI) x -= HPI;
  return mm_sec_table[x];
}

uint16_t mm_csct(uint16_t x)
{
  return mm_sec_table[HPI - x];
}

uint16_t mm_sint(uint16_t x)
{
  return mm_sin_table[x];
}

uint16_t mm_cost(uint16_t x)
{
  return mm_sin_table[HPI - x];
}