.\" Copyright (c) 2006 Apple Computer
.\"
.Dd December 11, 2006
.Dt COMPLEX 3
.Os
.Sh NAME
.Nm complex
.Nd complex floating-point functions
.Sh SYNOPSIS
.Fd #include
.Sh DESCRIPTION
The header file complex.h provides function prototypes and macros for working with C99 complex
floating-point values. The functions conform to the ISO/IEC 9899:1999(E) standard. In particular,
arguments with infinite real or imaginary parts are regarded as infinities, even if the other
part is a NaN.
.Pp
complex.h defines the macro
.Fa complex
for use as a type specifier, and the macro
.Fa I
to be the imaginary unit, which can be used to construct complex
floating-point numbers from two real floating-point numbers. For example:
.Bd -literal -offset indent
#include
double complex z = 1.0 + 1.0 * I; // z = 1 + i
.Ed
.Pp
Each of the functions that use complex floating-point values are provided in single, double,
and extended precision; the double precision prototypes are listed here. The man pages for the
individual functions provide more details on their use, special cases, and prototypes for their
single and extended precision versions.
.Pp
The double-precision functions defined in complex.h are:
.Pp
.Ft double
.Fn creal "double complex z"
.Pp
.Ft double
.Fn cimag "double complex z"
.Pp
.Fn creal
and
.Fn cimag
take a complex floating-point number and return its real and imaginary part,
respectively, as real floating-point numbers.
.Pp
.Ft double
.Fn cabs "double complex z"
.Pp
.Ft double
.Fn carg "double complex z"
.Pp
.Fn cabs
and
.Fn carg
take a complex floating-point number and return its norm and argument (phase angle),
respectively, as real floating-point numbers. They are used to convert between
rectangular and polar coordinates, and are fully specified in terms of real functions:
.Bd -literal -offset indent
cabs(x + iy) = hypot(x,y)
.br
carg(x + iy) = atan2(y,x)
.Ed
.Pp
.Ft double complex
.Fn conj "double complex z"
.Pp
.Fn conj
takes a complex floating-point number and returns its complex conjugate.
.Pp
.Ft double complex
.Fn cproj "double complex z"
.Pp
.Fn cproj
takes a complex floating-point number and returns its projection onto the
Riemann sphere, as defined in C99. For non-infinite inputs, the return value is
equal to the input value.
.Pp
.Ft double complex
.Fn csqrt "double complex z"
.Pp
.Fn csqrt
takes a complex floating-point number and returns its square root, with a branch cut on
the negative real axis.
.Pp
.Ft double complex
.Fn cexp "double complex z"
.Pp
.Ft double complex
.Fn clog "double complex z"
.Pp
.Fn cexp
and
.Fn clog
take a complex floating-point number and return its base-e exponential and logarithm,
respectively.
.Fn clog
has a branch cut on the negative real axis.
.Pp
.Ft double complex
.Fn cpow "double complex z" "double complex w"
.Pp
.Fn cpow
takes two complex floating-point numbers, and returns the first raised to the power of the second,
with a branch cut for the first parameter along the negative real axis.
.Pp
.Ft double complex
.Fn csin "double complex z"
.Pp
.Ft double complex
.Fn ccos "double complex z"
.Pp
.Ft double complex
.Fn ctan "double complex z"
.Pp
.Fn csin ,
.Fn ccos ,
and
.Fn ctan
take a complex floating-point number and return its sine, cosine, and tangent, respectively.
.Pp
.Ft double complex
.Fn casin "double complex z"
.Pp
.Ft double complex
.Fn cacos "double complex z"
.Pp
.Ft double complex
.Fn catan "double complex z"
.Pp
.Fn casin ,
.Fn cacos ,
and
.Fn catan
take a complex floating-point number and return its inverse sine, cosine, and tangent, respectively.
.Pp
.Fn casin
and
.Fn cacos
have branch cuts outside the interval
.Bq -1, 1
on the real axis,
and
.Fn catan
has a branch cut outside the interval
.Bq -i, i
on the imaginary axis.
.Pp
.Ft double complex
.Fn csinh "double complex z"
.Pp
.Ft double complex
.Fn ccosh "double complex z"
.Pp
.Ft double complex
.Fn ctanh "double complex z"
.Pp
.Fn csinh ,
.Fn ccosh ,
and
.Fn ctanh
take a complex floating-point number and return its hyperbolic sine, cosine, and tangent, respectively.
.Pp
.Ft double complex
.Fn casinh "double complex z"
.Pp
.Ft double complex
.Fn cacosh "double complex z"
.Pp
.Ft double complex
.Fn catanh "double complex z"
.Pp
.Fn casinh ,
.Fn cacosh ,
and
.Fn catanh
take a complex floating-point number and return its inverse hyperbolic sine, cosine, and tangent, respectively.
.Pp
.Fn casinh
has a branch cut outside the interval
.Bq -i, i
on the imaginary axis.
.Fn cacosh
has a branch cut at values less than 1 on the real axis.
.Fn catanh
has a branch cut outside the interval
.Bq -1, 1
on the real axis.
.Sh NOTE
Note that the complex math functions are not, in general, equivalent to their real counterparts for inputs on the
real axis. For example, csqrt(-1 + 0i) is 0 + i, whereas sqrt(-1) is NaN.
.Sh SEE ALSO
.Xr cabs 3 ,
.Xr cacos 3 ,
.Xr cacosh 3 ,
.Xr carg 3 ,
.Xr casin 3 ,
.Xr casinh 3 ,
.Xr catan 3 ,
.Xr catanh 3 ,
.Xr ccos 3 ,
.Xr ccosh 3 ,
.Xr cexp 3 ,
.Xr cimag 3 ,
.Xr clog 3 ,
.Xr conj 3 ,
.Xr cpow 3 ,
.Xr cproj 3 ,
.Xr creal 3 ,
.Xr csin 3 ,
.Xr csinh 3 ,
.Xr csqrt 3 ,
.Xr ctan 3 ,
.Xr ctanh 3 ,
.Xr math 3
.Sh STANDARDS
The functions conform to ISO/IEC 9899:1999(E).