/* VipsMath --- call various -lm functions (trig, log etc.) on images * * Copyright: 1990, N. Dessipris, based on im_powtra() * Author: Nicos Dessipris * Written on: 02/05/1990 * Modified on: * 5/5/93 JC * - adapted from im_lintra to work with partial images * - incorrect implementation of complex logs removed * 1/7/93 JC * - adapted for partial v2 * - ANSIfied * 24/2/95 JC * - im_logtra() adapted to make im_sintra() * - adapted for im_wrapone() * 26/1/96 JC * - im_asintra() added * 30/8/09 * - gtkdoc * - tiny cleanups * - use im__math() * 19/9/09 * - im_sintra() adapted to make math.c * 4/11/11 * - redone as a class * 11/8/15 * - log/log10 zero-avoid */ /* Copyright (C) 1991-2005 The National Gallery This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library 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 Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ /* These files are distributed with VIPS - http://www.vips.ecs.soton.ac.uk */ /* #define DEBUG */ #ifdef HAVE_CONFIG_H #include #endif /*HAVE_CONFIG_H*/ #include #include #include #include #include #include "unary.h" typedef struct _VipsMath { VipsUnary parent_instance; VipsOperationMath math; } VipsMath; typedef VipsUnaryClass VipsMathClass; G_DEFINE_TYPE( VipsMath, vips_math, VIPS_TYPE_UNARY ); static int vips_math_build( VipsObject *object ) { VipsObjectClass *class = VIPS_OBJECT_GET_CLASS( object ); VipsUnary *unary = (VipsUnary *) object; if( unary->in && vips_check_noncomplex( class->nickname, unary->in ) ) return( -1 ); if( VIPS_OBJECT_CLASS( vips_math_parent_class )->build( object ) ) return( -1 ); return( 0 ); } #define LOOP( IN, OUT, OP ) { \ IN * restrict p = (IN *) in[0]; \ OUT * restrict q = (OUT *) out; \ \ for( x = 0; x < sz; x++ ) \ q[x] = OP( p[x] ); \ } #define SWITCH( OP ) \ switch( vips_image_get_format( im ) ) { \ case VIPS_FORMAT_UCHAR: \ LOOP( unsigned char, float, OP ); break; \ case VIPS_FORMAT_CHAR: \ LOOP( signed char, float, OP ); break; \ case VIPS_FORMAT_USHORT: \ LOOP( unsigned short, float, OP ); break; \ case VIPS_FORMAT_SHORT: \ LOOP( signed short, float, OP ); break; \ case VIPS_FORMAT_UINT: \ LOOP( unsigned int, float, OP ); break; \ case VIPS_FORMAT_INT: \ LOOP( signed int, float, OP ); break; \ case VIPS_FORMAT_FLOAT: \ LOOP( float, float, OP ); break; \ case VIPS_FORMAT_DOUBLE: \ LOOP( double, double, OP ); break;\ \ default: \ g_assert_not_reached(); \ } /* If there's asinh, assume we have the other two as well. */ #if HAVE_ASINH #define ASINH( X ) (asinh( X )) #define ACOSH( X ) (acosh( X )) #define ATANH( X ) (atanh( X )) #else #define ASINH( X ) (LOGZ( (X) + sqrt( (X) * (X) + 1.0 ) )) #define ACOSH( X ) (LOGZ( (X) + sqrt( (X) * (X) - 1.0 ) )) #define ATANH( X ) (0.5 * LOGZ( (1.0 + (X)) / (1.0 - (X)) )) #endif /* sin/cos/tan in degrees. */ #define DSIN( X ) (sin( VIPS_RAD( X ) )) #define DCOS( X ) (cos( VIPS_RAD( X ) )) #define DTAN( X ) (tan( VIPS_RAD( X ) )) #define ADSIN( X ) (VIPS_DEG( asin( X ) )) #define ADCOS( X ) (VIPS_DEG( acos( X ) )) #define ADTAN( X ) (VIPS_DEG( atan( X ) )) /* exp10() is a gnu extension, use pow(). */ #define EXP10( X ) (pow( 10.0, (X) )) /* Zero-avoiding log, cf. zero-avoiding behaviour of /. */ #define LOGZ( X ) ((X) == 0.0 ? 0.0 : log( X )) #define LOGZ10( X ) ((X) == 0.0 ? 0.0 : log10( X )) static void vips_math_buffer( VipsArithmetic *arithmetic, VipsPel *out, VipsPel **in, int width ) { VipsMath *math = (VipsMath *) arithmetic; VipsImage *im = arithmetic->ready[0]; const int sz = width * vips_image_get_bands( im ); int x; switch( math->math ) { case VIPS_OPERATION_MATH_SIN: SWITCH( DSIN ); break; case VIPS_OPERATION_MATH_COS: SWITCH( DCOS ); break; case VIPS_OPERATION_MATH_TAN: SWITCH( DTAN ); break; case VIPS_OPERATION_MATH_ASIN: SWITCH( ADSIN ); break; case VIPS_OPERATION_MATH_ACOS: SWITCH( ADCOS ); break; case VIPS_OPERATION_MATH_ATAN: SWITCH( ADTAN ); break; case VIPS_OPERATION_MATH_SINH: SWITCH( sinh ); break; case VIPS_OPERATION_MATH_COSH: SWITCH( cosh ); break; case VIPS_OPERATION_MATH_TANH: SWITCH( tanh ); break; case VIPS_OPERATION_MATH_ASINH: SWITCH( ASINH ); break; case VIPS_OPERATION_MATH_ACOSH: SWITCH( ACOSH ); break; case VIPS_OPERATION_MATH_ATANH: SWITCH( ATANH ); break; case VIPS_OPERATION_MATH_LOG: SWITCH( LOGZ ); break; case VIPS_OPERATION_MATH_LOG10: SWITCH( LOGZ10 ); break; case VIPS_OPERATION_MATH_EXP: SWITCH( exp ); break; case VIPS_OPERATION_MATH_EXP10: SWITCH( EXP10 ); break; default: g_assert_not_reached(); } } /* Save a bit of typing. */ #define UC VIPS_FORMAT_UCHAR #define C VIPS_FORMAT_CHAR #define US VIPS_FORMAT_USHORT #define S VIPS_FORMAT_SHORT #define UI VIPS_FORMAT_UINT #define I VIPS_FORMAT_INT #define F VIPS_FORMAT_FLOAT #define X VIPS_FORMAT_COMPLEX #define D VIPS_FORMAT_DOUBLE #define DX VIPS_FORMAT_DPCOMPLEX static const VipsBandFormat vips_math_format_table[10] = { /* Band format: UC C US S UI I F X D DX */ /* Promotion: */ F, F, F, F, F, F, F, X, D, DX }; static void vips_math_class_init( VipsMathClass *class ) { GObjectClass *gobject_class = G_OBJECT_CLASS( class ); VipsObjectClass *object_class = (VipsObjectClass *) class; VipsArithmeticClass *aclass = VIPS_ARITHMETIC_CLASS( class ); gobject_class->set_property = vips_object_set_property; gobject_class->get_property = vips_object_get_property; object_class->nickname = "math"; object_class->description = _( "apply a math operation to an image" ); object_class->build = vips_math_build; aclass->process_line = vips_math_buffer; vips_arithmetic_set_format_table( aclass, vips_math_format_table ); VIPS_ARG_ENUM( class, "math", 200, _( "Operation" ), _( "Math to perform" ), VIPS_ARGUMENT_REQUIRED_INPUT, G_STRUCT_OFFSET( VipsMath, math ), VIPS_TYPE_OPERATION_MATH, VIPS_OPERATION_MATH_SIN ); } static void vips_math_init( VipsMath *math ) { } static int vips_mathv( VipsImage *in, VipsImage **out, VipsOperationMath math, va_list ap ) { return( vips_call_split( "math", ap, in, out, math ) ); } /** * vips_math: (method) * @in: input #VipsImage * @out: (out): output #VipsImage * @math: math operation to perform * @...: %NULL-terminated list of optional named arguments * * Perform various functions in -lm, the maths library, on images. * * Angles are expressed in degrees. The output type is float unless the * input is double, in which case the output is double. * * Non-complex images only. * * See also: vips_math2(). * * Returns: 0 on success, -1 on error */ int vips_math( VipsImage *in, VipsImage **out, VipsOperationMath math, ... ) { va_list ap; int result; va_start( ap, math ); result = vips_mathv( in, out, math, ap ); va_end( ap ); return( result ); } /** * vips_sin: (method) * @in: input #VipsImage * @out: (out): output #VipsImage * @...: %NULL-terminated list of optional named arguments * * Perform #VIPS_OPERATION_MATH_SIN on an image. See vips_math(). * * Returns: 0 on success, -1 on error */ int vips_sin( VipsImage *in, VipsImage **out, ... ) { va_list ap; int result; va_start( ap, out ); result = vips_mathv( in, out, VIPS_OPERATION_MATH_SIN, ap ); va_end( ap ); return( result ); } /** * vips_cos: (method) * @in: input #VipsImage * @out: (out): output #VipsImage * @...: %NULL-terminated list of optional named arguments * * Perform #VIPS_OPERATION_MATH_COS on an image. See vips_math(). * * Returns: 0 on success, -1 on error */ int vips_cos( VipsImage *in, VipsImage **out, ... ) { va_list ap; int result; va_start( ap, out ); result = vips_mathv( in, out, VIPS_OPERATION_MATH_COS, ap ); va_end( ap ); return( result ); } /** * vips_tan: (method) * @in: input #VipsImage * @out: (out): output #VipsImage * @...: %NULL-terminated list of optional named arguments * * Perform #VIPS_OPERATION_MATH_TAN on an image. See vips_math(). * * Returns: 0 on success, -1 on error */ int vips_tan( VipsImage *in, VipsImage **out, ... ) { va_list ap; int result; va_start( ap, out ); result = vips_mathv( in, out, VIPS_OPERATION_MATH_TAN, ap ); va_end( ap ); return( result ); } /** * vips_asin: (method) * @in: input #VipsImage * @out: (out): output #VipsImage * @...: %NULL-terminated list of optional named arguments * * Perform #VIPS_OPERATION_MATH_ASIN on an image. See vips_math(). * * Returns: 0 on success, -1 on error */ int vips_asin( VipsImage *in, VipsImage **out, ... ) { va_list ap; int result; va_start( ap, out ); result = vips_mathv( in, out, VIPS_OPERATION_MATH_ASIN, ap ); va_end( ap ); return( result ); } /** * vips_acos: (method) * @in: input #VipsImage * @out: (out): output #VipsImage * @...: %NULL-terminated list of optional named arguments * * Perform #VIPS_OPERATION_MATH_ACOS on an image. See vips_math(). * * Returns: 0 on success, -1 on error */ int vips_acos( VipsImage *in, VipsImage **out, ... ) { va_list ap; int result; va_start( ap, out ); result = vips_mathv( in, out, VIPS_OPERATION_MATH_ACOS, ap ); va_end( ap ); return( result ); } /** * vips_atan: (method) * @in: input #VipsImage * @out: (out): output #VipsImage * @...: %NULL-terminated list of optional named arguments * * Perform #VIPS_OPERATION_MATH_ATAN on an image. See vips_math(). * * Returns: 0 on success, -1 on error */ int vips_atan( VipsImage *in, VipsImage **out, ... ) { va_list ap; int result; va_start( ap, out ); result = vips_mathv( in, out, VIPS_OPERATION_MATH_ATAN, ap ); va_end( ap ); return( result ); } /** * vips_sinh: (method) * @in: input #VipsImage * @out: (out): output #VipsImage * @...: %NULL-terminated list of optional named arguments * * Perform #VIPS_OPERATION_MATH_SINH on an image. See vips_math(). * * Returns: 0 on success, -1 on error */ int vips_sinh( VipsImage *in, VipsImage **out, ... ) { va_list ap; int result; va_start( ap, out ); result = vips_mathv( in, out, VIPS_OPERATION_MATH_SINH, ap ); va_end( ap ); return( result ); } /** * vips_cosh: (method) * @in: input #VipsImage * @out: (out): output #VipsImage * @...: %NULL-terminated list of optional named arguments * * Perform #VIPS_OPERATION_MATH_COSH on an image. See vips_math(). * * Returns: 0 on success, -1 on error */ int vips_cosh( VipsImage *in, VipsImage **out, ... ) { va_list ap; int result; va_start( ap, out ); result = vips_mathv( in, out, VIPS_OPERATION_MATH_COSH, ap ); va_end( ap ); return( result ); } /** * vips_tanh: (method) * @in: input #VipsImage * @out: (out): output #VipsImage * @...: %NULL-terminated list of optional named arguments * * Perform #VIPS_OPERATION_MATH_TANH on an image. See vips_math(). * * Returns: 0 on success, -1 on error */ int vips_tanh( VipsImage *in, VipsImage **out, ... ) { va_list ap; int result; va_start( ap, out ); result = vips_mathv( in, out, VIPS_OPERATION_MATH_TANH, ap ); va_end( ap ); return( result ); } /** * vips_asinh: (method) * @in: input #VipsImage * @out: (out): output #VipsImage * @...: %NULL-terminated list of optional named arguments * * Perform #VIPS_OPERATION_MATH_ASINH on an image. See vips_math(). * * Returns: 0 on success, -1 on error */ int vips_asinh( VipsImage *in, VipsImage **out, ... ) { va_list ap; int result; va_start( ap, out ); result = vips_mathv( in, out, VIPS_OPERATION_MATH_ASINH, ap ); va_end( ap ); return( result ); } /** * vips_acosh: (method) * @in: input #VipsImage * @out: (out): output #VipsImage * @...: %NULL-terminated list of optional named arguments * * Perform #VIPS_OPERATION_MATH_ACOSH on an image. See vips_math(). * * Returns: 0 on success, -1 on error */ int vips_acosh( VipsImage *in, VipsImage **out, ... ) { va_list ap; int result; va_start( ap, out ); result = vips_mathv( in, out, VIPS_OPERATION_MATH_ACOSH, ap ); va_end( ap ); return( result ); } /** * vips_atanh: (method) * @in: input #VipsImage * @out: (out): output #VipsImage * @...: %NULL-terminated list of optional named arguments * * Perform #VIPS_OPERATION_MATH_ATANH on an image. See vips_math(). * * Returns: 0 on success, -1 on error */ int vips_atanh( VipsImage *in, VipsImage **out, ... ) { va_list ap; int result; va_start( ap, out ); result = vips_mathv( in, out, VIPS_OPERATION_MATH_ATANH, ap ); va_end( ap ); return( result ); } /** * vips_log: (method) * @in: input #VipsImage * @out: (out): output #VipsImage * @...: %NULL-terminated list of optional named arguments * * Perform #VIPS_OPERATION_MATH_LOG on an image. See vips_math(). * * Returns: 0 on success, -1 on error */ int vips_log( VipsImage *in, VipsImage **out, ... ) { va_list ap; int result; va_start( ap, out ); result = vips_mathv( in, out, VIPS_OPERATION_MATH_LOG, ap ); va_end( ap ); return( result ); } /** * vips_log10: (method) * @in: input #VipsImage * @out: (out): output #VipsImage * @...: %NULL-terminated list of optional named arguments * * Perform #VIPS_OPERATION_MATH_LOG10 on an image. See vips_math(). * * Returns: 0 on success, -1 on error */ int vips_log10( VipsImage *in, VipsImage **out, ... ) { va_list ap; int result; va_start( ap, out ); result = vips_mathv( in, out, VIPS_OPERATION_MATH_LOG10, ap ); va_end( ap ); return( result ); } /** * vips_exp: (method) * @in: input #VipsImage * @out: (out): output #VipsImage * @...: %NULL-terminated list of optional named arguments * * Perform #VIPS_OPERATION_MATH_EXP on an image. See vips_math(). * * Returns: 0 on success, -1 on error */ int vips_exp( VipsImage *in, VipsImage **out, ... ) { va_list ap; int result; va_start( ap, out ); result = vips_mathv( in, out, VIPS_OPERATION_MATH_EXP, ap ); va_end( ap ); return( result ); } /** * vips_exp10: (method) * @in: input #VipsImage * @out: (out): output #VipsImage * @...: %NULL-terminated list of optional named arguments * * Perform #VIPS_OPERATION_MATH_EXP10 on an image. See vips_math(). * * Returns: 0 on success, -1 on error */ int vips_exp10( VipsImage *in, VipsImage **out, ... ) { va_list ap; int result; va_start( ap, out ); result = vips_mathv( in, out, VIPS_OPERATION_MATH_EXP10, ap ); va_end( ap ); return( result ); }