ftp.nice.ch/pub/next/unix/graphics/rayshade.4.0.s.tar.gz#/rayshade.4.0/libcommon/transform.c

This is transform.c in view mode; [Download] [Up]

/*
 * transform.c
 *
 * Copyright (C) 1989, 1991, Craig E. Kolb
 * All rights reserved.
 *
 * This software may be freely copied, modified, and redistributed
 * provided that this copyright notice is preserved on all copies.
 *
 * You may not distribute this software, in whole or in part, as part of
 * any commercial product without the express consent of the authors.
 *
 * There is no warranty or other guarantee of fitness of this software
 * for any purpose.  It is provided solely "as is".
 *
 * $Id$
 *
 * $Log$
 */
#include "common.h"

/*
 * Matrices are indexed row-first; that is:
 * matrix[ROW][COLUMN]
 */
/*
 * Allocate new structure that holds both object-to-world and
 * world-to-object space transformation structures.  It probably
 * should hold pointers to these structures.
 */
Trans *
TransCreate(t, it)
Matrix *t, *it;
{
	Trans *res;

	res = (Trans *)Malloc(sizeof(Trans));
	res->trans = *t;
	res->itrans = *it;
	return res;
}

/*
 * Allocate new transformation 'matrix'.
 */
Matrix *
MatrixCreate()
{
	Matrix *res;

	res = (Matrix *)Malloc(sizeof(Matrix));
	MatrixInit(res);
	return res;
}

/*
 * Set up matrix for counter-clockwise rotation about vector (x, y, z) by
 * theta radians.
 */
Matrix *
RotationMatrix(dir, theta)
Vector *dir;
Float theta;
{
	Float n1, n2, n3, sintheta, costheta;
	Vector vector;
	Matrix *trans;

	trans = MatrixCreate();
	vector = *dir;
	if (VecNormalize(&vector) == 0.)
		RLerror(RL_WARN, "Degenerate rotation axis.\n");

	sintheta = sin(theta);
	costheta = cos(theta);

	n1 = vector.x; n2 = vector.y; n3 = vector.z;
	trans->matrix[0][0] = (Float)(n1*n1 + (1. - n1*n1)*costheta);
	trans->matrix[0][1] = (Float)(n1*n2*(1 - costheta) + n3*sintheta);
	trans->matrix[0][2] = (Float)(n1*n3*(1 - costheta) - n2*sintheta);
	trans->matrix[1][0] = (Float)(n1*n2*(1 - costheta) - n3*sintheta);
	trans->matrix[1][1] = (Float)(n2*n2 + (1 - n2*n2)*costheta);
	trans->matrix[1][2] = (Float)(n2*n3*(1 - costheta) + n1*sintheta);
	trans->matrix[2][0] = (Float)(n1*n3*(1 - costheta) + n2*sintheta);
	trans->matrix[2][1] = (Float)(n2*n3*(1 - costheta) - n1*sintheta);
	trans->matrix[2][2] = (Float)(n3*n3 + (1 - n3*n3)*costheta);

	return trans;
}

/*
 * Apply translation by (vec) to trans.
 */
Matrix *
TranslationMatrix(vec)
Vector *vec;
{
	Matrix *trans;

	trans = MatrixCreate();
	trans->translate.x = vec->x;
	trans->translate.y = vec->y;
	trans->translate.z = vec->z;
	return trans;
}

/*
 * Apply scale of (x, y, z) to trans.
 */
Matrix *
ScaleMatrix(x, y, z)
Float x, y, z;
{
	Matrix *trans;

	trans = MatrixCreate();
	trans->matrix[0][0] = x;
	trans->matrix[1][1] = y;
	trans->matrix[2][2] = z;
	return trans;
}

/*
 * Multiply m1 and m2, copy result into "res".
 */
void
MatrixMult(t1, t2, res)
Matrix *t1, *t2, *res;
{
	register int i;
	Matrix tmp;

	for (i = 0; i < 3; i++) {
		tmp.matrix[i][0] = t1->matrix[i][0] * t2->matrix[0][0] +
			  	   t1->matrix[i][1] * t2->matrix[1][0] +
			  	   t1->matrix[i][2] * t2->matrix[2][0];
		tmp.matrix[i][1] = t1->matrix[i][0] * t2->matrix[0][1] +
			  	   t1->matrix[i][1] * t2->matrix[1][1] +
			  	   t1->matrix[i][2] * t2->matrix[2][1];
		tmp.matrix[i][2] = t1->matrix[i][0] * t2->matrix[0][2] +
			  	   t1->matrix[i][1] * t2->matrix[1][2] +
		  		   t1->matrix[i][2] * t2->matrix[2][2];
	}

	tmp.translate.x = t1->translate.x * t2->matrix[0][0] +
			  t1->translate.y * t2->matrix[1][0] +
			  t1->translate.z * t2->matrix[2][0] + t2->translate.x;
	tmp.translate.y = t1->translate.x * t2->matrix[0][1] +
			  t1->translate.y * t2->matrix[1][1] +
			  t1->translate.z * t2->matrix[2][1] + t2->translate.y;
	tmp.translate.z = t1->translate.x * t2->matrix[0][2] +
			  t1->translate.y * t2->matrix[1][2] +
			  t1->translate.z * t2->matrix[2][2] + t2->translate.z;
	MatrixCopy(&tmp, res);
}

/*
 * Return transformation information to map the "coordinate system"
 * with the given origin, "up" vector, radius, and up axis lengths to
 * one in which the "up" vector is the Z axis and the x/y/up axes
 * have unit length.  This is useful for transforming a general
 * form of a primitive into a canonical, Z-axis aligned, unit size
 * primitive, facilitating intersection testing.
 */
Trans *
CoordSysTransform(origin, up, r, len)
Vector *origin, *up;
Float r, len;
{
	Matrix trans, itrans, *tmp;
	Vector atmp;

	MatrixInit(&trans);
	tmp = ScaleMatrix(r, r, len);
	MatrixMult(&trans, tmp, &trans);
	free((voidstar)tmp);
	if (fabs(up->z) == 1.) {
		atmp.x = 1.;
		atmp.y = atmp.z = 0.;
	} else {
		atmp.x = up->y;
		atmp.y = -up->x;
		atmp.z= 0.;
	}
	/*
	 * Might want to make sure that |up->z| is < 1.
	 */
	tmp = RotationMatrix(&atmp, -acos(up->z));
	MatrixMult(&trans, tmp, &trans);
	free((voidstar)tmp);
	tmp = TranslationMatrix(origin);
	MatrixMult(&trans, tmp, &trans);
	free((voidstar)tmp);
	MatrixInvert(&trans, &itrans);

	return TransCreate(&trans, &itrans);
}

void
TransCopy(from, into)
Trans *into, *from;
{
	MatrixCopy(&from->trans, &into->trans);
	MatrixCopy(&from->itrans, &into->itrans);
}

void
TransInvert(from, into)
Trans *into, *from;
{
	Matrix ttmp;
	/*
	 * In case into == from...
	 */
	ttmp = from->trans;
	into->trans = from->itrans;
	into->itrans = ttmp;
}

/*
 * Copy a given transformation structure.
 */
void
MatrixCopy(from, into)
Matrix *into, *from;
{
	into->matrix[0][0] = from->matrix[0][0];
	into->matrix[0][1] = from->matrix[0][1];
	into->matrix[0][2] = from->matrix[0][2];
	into->matrix[1][0] = from->matrix[1][0];
	into->matrix[1][1] = from->matrix[1][1];
	into->matrix[1][2] = from->matrix[1][2];
	into->matrix[2][0] = from->matrix[2][0];
	into->matrix[2][1] = from->matrix[2][1];
	into->matrix[2][2] = from->matrix[2][2];
	into->translate = from->translate;
}

void
TransInit(trans)
Trans *trans;
{
	MatrixInit(&trans->trans);
	MatrixInit(&trans->itrans);
}

void
TransCompose(t1, t2, res)
Trans *t1, *t2, *res;
{
	MatrixMult(&t1->trans, &t2->trans, &res->trans);
	MatrixMult(&t1->itrans, &t2->itrans, &res->itrans);
}

/*
 * Initialize transformation structure.
 */
void
MatrixInit(trans)
Matrix *trans;
{
	trans->matrix[0][0] = trans->matrix[1][1] = trans->matrix[2][2] = 1.;
	trans->matrix[0][1] = trans->matrix[0][2] = trans->matrix[1][0] =
	trans->matrix[1][2] = trans->matrix[2][0] = trans->matrix[2][1] = 0.;
	trans->translate.x = trans->translate.y = trans->translate.z = 0.;
}

/*
 * Calculate inverse of the given transformation structure.
 */
void
MatrixInvert(trans, inverse)
Matrix *inverse, *trans;
{
	Matrix ttmp;
	int i;
	Float d;
	extern int yylineno;

	ttmp.matrix[0][0] = trans->matrix[1][1]*trans->matrix[2][2] -
			    trans->matrix[1][2]*trans->matrix[2][1];
	ttmp.matrix[1][0] = trans->matrix[1][0]*trans->matrix[2][2] -
			    trans->matrix[1][2]*trans->matrix[2][0];
	ttmp.matrix[2][0] = trans->matrix[1][0]*trans->matrix[2][1] -
			    trans->matrix[1][1]*trans->matrix[2][0];

	ttmp.matrix[0][1] = trans->matrix[0][1]*trans->matrix[2][2] -
			    trans->matrix[0][2]*trans->matrix[2][1];
	ttmp.matrix[1][1] = trans->matrix[0][0]*trans->matrix[2][2] -
			    trans->matrix[0][2]*trans->matrix[2][0];
	ttmp.matrix[2][1] = trans->matrix[0][0]*trans->matrix[2][1] -
			    trans->matrix[0][1]*trans->matrix[2][0];

	ttmp.matrix[0][2] = trans->matrix[0][1]*trans->matrix[1][2] -
			    trans->matrix[0][2]*trans->matrix[1][1];
	ttmp.matrix[1][2] = trans->matrix[0][0]*trans->matrix[1][2] -
			    trans->matrix[0][2]*trans->matrix[1][0];
	ttmp.matrix[2][2] = trans->matrix[0][0]*trans->matrix[1][1] -
			    trans->matrix[0][1]*trans->matrix[1][0];

	d = trans->matrix[0][0]*ttmp.matrix[0][0] -
	    trans->matrix[0][1]*ttmp.matrix[1][0] +
	    trans->matrix[0][2]*ttmp.matrix[2][0];

	if (fabs(d) < EPSILON*EPSILON)
		RLerror(RL_PANIC, "Singular matrix.\n",yylineno);

	ttmp.matrix[0][0] /= d;
	ttmp.matrix[0][2] /= d;
	ttmp.matrix[1][1] /= d;
	ttmp.matrix[2][0] /= d;
	ttmp.matrix[2][2] /= d;

	d = -d;

	ttmp.matrix[0][1] /= d;
	ttmp.matrix[1][0] /= d;
	ttmp.matrix[1][2] /= d;
	ttmp.matrix[2][1] /= d;

	ttmp.translate.x = -(ttmp.matrix[0][0]*trans->translate.x +
			     ttmp.matrix[1][0]*trans->translate.y +
			     ttmp.matrix[2][0]*trans->translate.z);
	ttmp.translate.y = -(ttmp.matrix[0][1]*trans->translate.x +
			     ttmp.matrix[1][1]*trans->translate.y +
			     ttmp.matrix[2][1]*trans->translate.z);
	ttmp.translate.z = -(ttmp.matrix[0][2]*trans->translate.x +
			     ttmp.matrix[1][2]*trans->translate.y +
			     ttmp.matrix[2][2]*trans->translate.z);

	MatrixCopy(&ttmp, inverse);
}

/*
 * Apply a transformation to a point (translation affects the point).
 */
void
PointTransform(vec, trans)
Vector *vec;
Matrix *trans;
{
	Vector tmp;

	tmp.x = vec->x * trans->matrix[0][0] + vec->y * trans->matrix[1][0] +
			vec->z * trans->matrix[2][0] + trans->translate.x;
	tmp.y = vec->x * trans->matrix[0][1] + vec->y * trans->matrix[1][1] +
			vec->z * trans->matrix[2][1] + trans->translate.y;
	tmp.z = vec->x * trans->matrix[0][2] + vec->y * trans->matrix[1][2] +
			vec->z * trans->matrix[2][2] + trans->translate.z;
	*vec = tmp;
}

/*
 * 'c1x' is the X (0th) component of the first column, and so on.
 */
Matrix *
ArbitraryMatrix(c1x, c2x, c3x, c1y, c2y, c3y, c1z, c2z, c3z, tx, ty, tz)
Float c1x, c1y, c1z, c2x, c2y, c2z, c3x, c3y, c3z, tx, ty, tz;
{
	Matrix *trans;

	trans = MatrixCreate();

	trans->matrix[0][0] = c1x;
	trans->matrix[1][0] = c1y;
	trans->matrix[2][0] = c1z;

	trans->matrix[0][1] = c2x;
	trans->matrix[1][1] = c2y;
	trans->matrix[2][1] = c2z;

	trans->matrix[0][2] = c3x;
	trans->matrix[1][2] = c3y;
	trans->matrix[2][2] = c3z;

	trans->translate.x = tx;
	trans->translate.y = ty;
	trans->translate.z = tz;

	return trans;
}

/*
 * Apply transformation to a vector (translations have no effect).
 */
void
VecTransform(vec, trans)
Vector *vec;
Matrix *trans;
{
	Vector tmp;

	tmp.x = vec->x*trans->matrix[0][0] +
		vec->y*trans->matrix[1][0] + vec->z*trans->matrix[2][0];
	tmp.y = vec->x*trans->matrix[0][1] +
		vec->y*trans->matrix[1][1] + vec->z*trans->matrix[2][1];
	tmp.z = vec->x*trans->matrix[0][2] +
		vec->y*trans->matrix[1][2] + vec->z*trans->matrix[2][2];

	*vec = tmp;
}

/*
 * Transform normal -- multiply by the transpose of the given
 * matrix (which is the inverse of the 'desired' transformation).
 */
void
NormalTransform(norm, it)
Vector *norm;
Matrix *it;
{
	Vector onorm;

	onorm = *norm;

	norm->x = onorm.x*it->matrix[0][0] + onorm.y*it->matrix[0][1] +
				onorm.z*it->matrix[0][2];
	norm->y = onorm.x*it->matrix[1][0] + onorm.y*it->matrix[1][1] +
				onorm.z*it->matrix[1][2];
	norm->z = onorm.x*it->matrix[2][0] + onorm.y*it->matrix[2][1] +
				onorm.z*it->matrix[2][2];
	(void)VecNormalize(norm);
}

/*
 * Transform "ray" by transforming the origin point and direction vector.
 */
Float
RayTransform(ray, trans)
Ray *ray;
Matrix *trans;
{
	PointTransform(&ray->pos, trans);
	VecTransform(&ray->dir, trans);
	return VecNormalize(&ray->dir);
}

These are the contents of the former NiCE NeXT User Group NeXTSTEP/OpenStep software archive, currently hosted by Netfuture.ch.