libdspl-2.0/dspl/src/math_std/asin_cmplx.c

162 wiersze
5.0 KiB
C
Czysty Wina Historia

This file contains ambiguous Unicode characters!

This file contains ambiguous Unicode characters that may be confused with others in your current locale. If your use case is intentional and legitimate, you can safely ignore this warning. Use the Escape button to highlight these characters.

/*
* Copyright (c) 2015-2024 Sergey Bakhurin
* Digital Signal Processing Library [http://dsplib.org]
*
* This file is part of DSPL.
*
* 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 3 of the License, or
* (at your option) any later version.
*
* DSPL 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 Foobar. If not, see <http://www.gnu.org/licenses/>.
*/
#include <stdio.h>
#include <stdlib.h>
#include "dspl.h"
#ifdef DOXYGEN_ENGLISH
/*! ****************************************************************************
\ingroup SPEC_MATH_TRIG_GROUP
\brief The inverse of the sine function the complex vector argument `x`.
Function calculates the inverse of the sine function as: \n
\f[
\textrm{Arcsin}(x) = j \textrm{Ln}\left( j x + \sqrt{1 - x^2} \right)
\f]
\param[in] x
Pointer to the argument vector `x`. \n
Vector size is `[n x 1]`. \n\n
\param[in] n
Input vector `x` and the inverse sine vector `y` size. \n\n
\param[out] y
Pointer to the output complex vector `y`,
corresponds to the input vector `x`.\n
Vector size is `[n x 1]`. \n
Memory must be allocated. \n\n
\return
`RES_OK` if function calculated successfully.\n
Else \ref ERROR_CODE_GROUP "code error". \n
Example: \n
\code{.cpp}
complex_t x[3] = {{1.0, 2.0}, {3.0, 4.0}, {5.0, 6.0}};
complex_t y[3];
int k;
asin_cmplx(x, 3, y);
for(k = 0; k < 3; k++)
printf("asin_cmplx(%.1f%+.1fj) = %.3f%+.3fj\n",
RE(x[k]), IM(x[k]), RE(y[k]), IM(y[k]));
\endcode
\n
Output is: \n
\verbatim
asin_cmplx(1.0+2.0j) = 0.427+1.529j
asin_cmplx(3.0+4.0j) = 0.634+2.306j
asin_cmplx(5.0+6.0j) = 0.691+2.749j
\endverbatim
\author Sergey Bakhurin www.dsplib.org
***************************************************************************** */
#endif
#ifdef DOXYGEN_RUSSIAN
/*! ****************************************************************************
\ingroup SPEC_MATH_TRIG_GROUP
\brief Арксинус комплексного аргумента `x`.
Функция рассчитывает значения арксинуса комплексного аргумента,
заданного вектором `x` длины `n`: \n
\f[
\textrm{Arcsin}(x) = j \textrm{Ln}\left( j x + \sqrt{1 - x^2} \right)
\f]
\param[in] x
Указатель на вектор аргумента комплексного арксинуса. \n
Размер вектора `[n x 1]`. \n \n
\param[in] n
Размер входного и выходного векторов `x` и `y`. \n \n
\param[out] y
Указатель на вектор значений комплексного арксинуса,
соответствующего входному вектору `x`. \n
Размер массива `[n x 1]`. \n
Память должна быть выделена. \n \n
\return
`RES_OK` если значение функции рассчитано успешно . \n
В противном случае \ref ERROR_CODE_GROUP "код ошибки": \n
\note
Функция может использоваться для расчета арксинуса аргумента
большего единицы, когда вещественная функция `acos` не определена.
Например при выполнении следующего кода
\code{.cpp}
complex_t x[3] = {{1.0, 2.0}, {3.0, 4.0}, {5.0, 6.0}};
complex_t y[3];
int k;
asin_cmplx(x, 3, y);
for(k = 0; k < 3; k++)
printf("asin_cmplx(%.1f%+.1fj) = %.3f%+.3fj\n",
RE(x[k]), IM(x[k]), RE(y[k]), IM(y[k]));
\endcode
\n
Результатом работы будет
\verbatim
asin_cmplx(1.0+2.0j) = 0.427+1.529j
asin_cmplx(3.0+4.0j) = 0.634+2.306j
asin_cmplx(5.0+6.0j) = 0.691+2.749j
\endverbatim
\author Бахурин Сергей www.dsplib.org
***************************************************************************** */
#endif
int DSPL_API asin_cmplx(complex_t* x, int n, complex_t *y)
{
int k;
complex_t tmp;
if(!x || !y)
return ERROR_PTR;
if(n < 1)
return ERROR_SIZE;
for(k = 0; k < n; k++)
{
RE(tmp) = 1.0 - CMRE(x[k], x[k]); /* 1-x[k]^2 */
IM(tmp) = - CMIM(x[k], x[k]); /* 1-x[k]^2 */
sqrt_cmplx(&tmp, 1, y+k); /* sqrt(1 - x[k]^2) */
RE(y[k]) -= IM(x[k]); /* j * x[k] + sqrt(1 - x[k]^2) */
IM(y[k]) += RE(x[k]); /* j * x[k] + sqrt(1 - x[k]^2) */
log_cmplx(y+k, 1, &tmp); /* log( j * x[k] + sqrt(1 - x[k]^2) ) */
RE(y[k]) = IM(tmp); /* -j * log( j * x[k] + sqrt(1 - x[k]^2) ) */
IM(y[k]) = -RE(tmp); /* -j * log( j * x[k] + sqrt(1 - x[k]^2) ) */
}
return RES_OK;
}