Home | History | Annotate | Download | only in libopenjpeg20
      1 /*
      2  * The copyright in this software is being made available under the 2-clauses
      3  * BSD License, included below. This software may be subject to other third
      4  * party and contributor rights, including patent rights, and no such rights
      5  * are granted under this license.
      6  *
      7  * Copyright (c) 2002-2014, Universite catholique de Louvain (UCL), Belgium
      8  * Copyright (c) 2002-2014, Professor Benoit Macq
      9  * Copyright (c) 2001-2003, David Janssens
     10  * Copyright (c) 2002-2003, Yannick Verschueren
     11  * Copyright (c) 2003-2007, Francois-Olivier Devaux
     12  * Copyright (c) 2003-2014, Antonin Descampe
     13  * Copyright (c) 2005, Herve Drolon, FreeImage Team
     14  * Copyright (c) 2008, 2011-2012, Centre National d'Etudes Spatiales (CNES), FR
     15  * Copyright (c) 2012, CS Systemes d'Information, France
     16  * All rights reserved.
     17  *
     18  * Redistribution and use in source and binary forms, with or without
     19  * modification, are permitted provided that the following conditions
     20  * are met:
     21  * 1. Redistributions of source code must retain the above copyright
     22  *    notice, this list of conditions and the following disclaimer.
     23  * 2. Redistributions in binary form must reproduce the above copyright
     24  *    notice, this list of conditions and the following disclaimer in the
     25  *    documentation and/or other materials provided with the distribution.
     26  *
     27  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS `AS IS'
     28  * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     29  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     30  * ARE DISCLAIMED.  IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
     31  * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
     32  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
     33  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
     34  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
     35  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
     36  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
     37  * POSSIBILITY OF SUCH DAMAGE.
     38  */
     39 
     40 #ifdef __SSE__
     41 #include <xmmintrin.h>
     42 #endif
     43 
     44 #include "opj_includes.h"
     45 
     46 /* <summary> */
     47 /* This table contains the norms of the basis function of the reversible MCT. */
     48 /* </summary> */
     49 static const OPJ_FLOAT64 opj_mct_norms[3] = { 1.732, .8292, .8292 };
     50 
     51 /* <summary> */
     52 /* This table contains the norms of the basis function of the irreversible MCT. */
     53 /* </summary> */
     54 static const OPJ_FLOAT64 opj_mct_norms_real[3] = { 1.732, 1.805, 1.573 };
     55 
     56 const OPJ_FLOAT64 * opj_mct_get_mct_norms ()
     57 {
     58 	return opj_mct_norms;
     59 }
     60 
     61 const OPJ_FLOAT64 * opj_mct_get_mct_norms_real ()
     62 {
     63 	return opj_mct_norms_real;
     64 }
     65 
     66 /* <summary> */
     67 /* Foward reversible MCT. */
     68 /* </summary> */
     69 void opj_mct_encode(
     70 		OPJ_INT32* restrict c0,
     71 		OPJ_INT32* restrict c1,
     72 		OPJ_INT32* restrict c2,
     73 		OPJ_UINT32 n)
     74 {
     75 	OPJ_UINT32 i;
     76 	for(i = 0; i < n; ++i) {
     77 		OPJ_INT32 r = c0[i];
     78 		OPJ_INT32 g = c1[i];
     79 		OPJ_INT32 b = c2[i];
     80 		OPJ_INT32 y = (r + (g * 2) + b) >> 2;
     81 		OPJ_INT32 u = b - g;
     82 		OPJ_INT32 v = r - g;
     83 		c0[i] = y;
     84 		c1[i] = u;
     85 		c2[i] = v;
     86 	}
     87 }
     88 
     89 /* <summary> */
     90 /* Inverse reversible MCT. */
     91 /* </summary> */
     92 void opj_mct_decode(
     93 		OPJ_INT32* restrict c0,
     94 		OPJ_INT32* restrict c1,
     95 		OPJ_INT32* restrict c2,
     96 		OPJ_UINT32 n)
     97 {
     98 	OPJ_UINT32 i;
     99 	for (i = 0; i < n; ++i) {
    100 		OPJ_INT32 y = c0[i];
    101 		OPJ_INT32 u = c1[i];
    102 		OPJ_INT32 v = c2[i];
    103 		OPJ_INT32 g = y - ((u + v) >> 2);
    104 		OPJ_INT32 r = v + g;
    105 		OPJ_INT32 b = u + g;
    106 		c0[i] = r;
    107 		c1[i] = g;
    108 		c2[i] = b;
    109 	}
    110 }
    111 
    112 /* <summary> */
    113 /* Get norm of basis function of reversible MCT. */
    114 /* </summary> */
    115 OPJ_FLOAT64 opj_mct_getnorm(OPJ_UINT32 compno) {
    116 	return opj_mct_norms[compno];
    117 }
    118 
    119 /* <summary> */
    120 /* Foward irreversible MCT. */
    121 /* </summary> */
    122 void opj_mct_encode_real(
    123 		OPJ_INT32* restrict c0,
    124 		OPJ_INT32* restrict c1,
    125 		OPJ_INT32* restrict c2,
    126 		OPJ_UINT32 n)
    127 {
    128 	OPJ_UINT32 i;
    129 	for(i = 0; i < n; ++i) {
    130 		OPJ_INT32 r = c0[i];
    131 		OPJ_INT32 g = c1[i];
    132 		OPJ_INT32 b = c2[i];
    133 		OPJ_INT32 y =  opj_int_fix_mul(r, 2449) + opj_int_fix_mul(g, 4809) + opj_int_fix_mul(b, 934);
    134 		OPJ_INT32 u = -opj_int_fix_mul(r, 1382) - opj_int_fix_mul(g, 2714) + opj_int_fix_mul(b, 4096);
    135 		OPJ_INT32 v =  opj_int_fix_mul(r, 4096) - opj_int_fix_mul(g, 3430) - opj_int_fix_mul(b, 666);
    136 		c0[i] = y;
    137 		c1[i] = u;
    138 		c2[i] = v;
    139 	}
    140 }
    141 
    142 /* <summary> */
    143 /* Inverse irreversible MCT. */
    144 /* </summary> */
    145 void opj_mct_decode_real(
    146 		OPJ_FLOAT32* restrict c0,
    147 		OPJ_FLOAT32* restrict c1,
    148 		OPJ_FLOAT32* restrict c2,
    149 		OPJ_UINT32 n)
    150 {
    151 	OPJ_UINT32 i;
    152 #ifdef __SSE__
    153 	// Mantis BUGID: 0056291. The address must be 16-byte aligned.
    154 	// TestFile: fuzz-signal_sigsegv_6e9e7f_5076_5265.pdf
    155 	if ((uintptr_t)c0 % 16 == 0 && (uintptr_t)c1 % 16 == 0 && (uintptr_t)c2 % 16 == 0){
    156 		__m128 vrv, vgu, vgv, vbu;
    157 		vrv = _mm_set1_ps(1.402f);
    158 		vgu = _mm_set1_ps(0.34413f);
    159 		vgv = _mm_set1_ps(0.71414f);
    160 		vbu = _mm_set1_ps(1.772f);
    161 		for (i = 0; i < (n >> 3); ++i) {
    162 			__m128 vy, vu, vv;
    163 			__m128 vr, vg, vb;
    164 
    165 			vy = _mm_load_ps(c0);
    166 			vu = _mm_load_ps(c1);
    167 			vv = _mm_load_ps(c2);
    168 			vr = _mm_add_ps(vy, _mm_mul_ps(vv, vrv));
    169 			vg = _mm_sub_ps(_mm_sub_ps(vy, _mm_mul_ps(vu, vgu)), _mm_mul_ps(vv, vgv));
    170 			vb = _mm_add_ps(vy, _mm_mul_ps(vu, vbu));
    171 			_mm_store_ps(c0, vr);
    172 			_mm_store_ps(c1, vg);
    173 			_mm_store_ps(c2, vb);
    174 			c0 += 4;
    175 			c1 += 4;
    176 			c2 += 4;
    177 
    178 			vy = _mm_load_ps(c0);
    179 			vu = _mm_load_ps(c1);
    180 			vv = _mm_load_ps(c2);
    181 			vr = _mm_add_ps(vy, _mm_mul_ps(vv, vrv));
    182 			vg = _mm_sub_ps(_mm_sub_ps(vy, _mm_mul_ps(vu, vgu)), _mm_mul_ps(vv, vgv));
    183 			vb = _mm_add_ps(vy, _mm_mul_ps(vu, vbu));
    184 			_mm_store_ps(c0, vr);
    185 			_mm_store_ps(c1, vg);
    186 			_mm_store_ps(c2, vb);
    187 			c0 += 4;
    188 			c1 += 4;
    189 			c2 += 4;
    190 		}
    191 		n &= 7;
    192 	} else {
    193 		for(i = 0; i < n; ++i) {
    194 			OPJ_FLOAT32 y = c0[i];
    195 			OPJ_FLOAT32 u = c1[i];
    196 			OPJ_FLOAT32 v = c2[i];
    197 			OPJ_FLOAT32 r = y + (v * 1.402f);
    198 			OPJ_FLOAT32 g = y - (u * 0.34413f) - (v * (0.71414f));
    199 			OPJ_FLOAT32 b = y + (u * 1.772f);
    200 			c0[i] = r;
    201 			c1[i] = g;
    202 			c2[i] = b;
    203 		}
    204 	}
    205 
    206 #endif
    207 	for(i = 0; i < n; ++i) {
    208 		OPJ_FLOAT32 y = c0[i];
    209 		OPJ_FLOAT32 u = c1[i];
    210 		OPJ_FLOAT32 v = c2[i];
    211 		OPJ_FLOAT32 r = y + (v * 1.402f);
    212 		OPJ_FLOAT32 g = y - (u * 0.34413f) - (v * (0.71414f));
    213 		OPJ_FLOAT32 b = y + (u * 1.772f);
    214 		c0[i] = r;
    215 		c1[i] = g;
    216 		c2[i] = b;
    217 	}
    218 }
    219 
    220 /* <summary> */
    221 /* Get norm of basis function of irreversible MCT. */
    222 /* </summary> */
    223 OPJ_FLOAT64 opj_mct_getnorm_real(OPJ_UINT32 compno) {
    224 	return opj_mct_norms_real[compno];
    225 }
    226 
    227 
    228 OPJ_BOOL opj_mct_encode_custom(
    229 					   OPJ_BYTE * pCodingdata,
    230 					   OPJ_UINT32 n,
    231 					   OPJ_BYTE ** pData,
    232 					   OPJ_UINT32 pNbComp,
    233 					   OPJ_UINT32 isSigned)
    234 {
    235 	OPJ_FLOAT32 * lMct = (OPJ_FLOAT32 *) pCodingdata;
    236 	OPJ_UINT32 i;
    237 	OPJ_UINT32 j;
    238 	OPJ_UINT32 k;
    239 	OPJ_UINT32 lNbMatCoeff = pNbComp * pNbComp;
    240 	OPJ_INT32 * lCurrentData = 00;
    241 	OPJ_INT32 * lCurrentMatrix = 00;
    242 	OPJ_INT32 ** lData = (OPJ_INT32 **) pData;
    243 	OPJ_UINT32 lMultiplicator = 1 << 13;
    244 	OPJ_INT32 * lMctPtr;
    245 
    246     OPJ_ARG_NOT_USED(isSigned);
    247 
    248 	lCurrentData = (OPJ_INT32 *) opj_malloc((pNbComp + lNbMatCoeff) * sizeof(OPJ_INT32));
    249 	if (! lCurrentData) {
    250 		return OPJ_FALSE;
    251 	}
    252 
    253 	lCurrentMatrix = lCurrentData + pNbComp;
    254 
    255 	for (i =0;i<lNbMatCoeff;++i) {
    256 		lCurrentMatrix[i] = (OPJ_INT32) (*(lMct++) * (OPJ_FLOAT32)lMultiplicator);
    257 	}
    258 
    259 	for (i = 0; i < n; ++i)  {
    260 		lMctPtr = lCurrentMatrix;
    261 		for (j=0;j<pNbComp;++j) {
    262 			lCurrentData[j] = (*(lData[j]));
    263 		}
    264 
    265 		for (j=0;j<pNbComp;++j) {
    266 			*(lData[j]) = 0;
    267 			for (k=0;k<pNbComp;++k) {
    268 				*(lData[j]) += opj_int_fix_mul(*lMctPtr, lCurrentData[k]);
    269 				++lMctPtr;
    270 			}
    271 
    272 			++lData[j];
    273 		}
    274 	}
    275 
    276 	opj_free(lCurrentData);
    277 
    278 	return OPJ_TRUE;
    279 }
    280 
    281 OPJ_BOOL opj_mct_decode_custom(
    282 					   OPJ_BYTE * pDecodingData,
    283 					   OPJ_UINT32 n,
    284 					   OPJ_BYTE ** pData,
    285 					   OPJ_UINT32 pNbComp,
    286 					   OPJ_UINT32 isSigned)
    287 {
    288 	OPJ_FLOAT32 * lMct;
    289 	OPJ_UINT32 i;
    290 	OPJ_UINT32 j;
    291 	OPJ_UINT32 k;
    292 
    293 	OPJ_FLOAT32 * lCurrentData = 00;
    294 	OPJ_FLOAT32 * lCurrentResult = 00;
    295 	OPJ_FLOAT32 ** lData = (OPJ_FLOAT32 **) pData;
    296 
    297     OPJ_ARG_NOT_USED(isSigned);
    298 
    299 	lCurrentData = (OPJ_FLOAT32 *) opj_malloc (2 * pNbComp * sizeof(OPJ_FLOAT32));
    300 	if (! lCurrentData) {
    301 		return OPJ_FALSE;
    302 	}
    303 	lCurrentResult = lCurrentData + pNbComp;
    304 
    305 	for (i = 0; i < n; ++i) {
    306 		lMct = (OPJ_FLOAT32 *) pDecodingData;
    307 		for (j=0;j<pNbComp;++j) {
    308 			lCurrentData[j] = (OPJ_FLOAT32) (*(lData[j]));
    309 		}
    310 		for (j=0;j<pNbComp;++j) {
    311 			lCurrentResult[j] = 0;
    312 			for	(k=0;k<pNbComp;++k)	{
    313 				lCurrentResult[j] += *(lMct++) * lCurrentData[k];
    314 			}
    315 			*(lData[j]++) = (OPJ_FLOAT32) (lCurrentResult[j]);
    316 		}
    317 	}
    318 	opj_free(lCurrentData);
    319 	return OPJ_TRUE;
    320 }
    321 
    322 void opj_calculate_norms(	OPJ_FLOAT64 * pNorms,
    323 							OPJ_UINT32 pNbComps,
    324 							OPJ_FLOAT32 * pMatrix)
    325 {
    326 	OPJ_UINT32 i,j,lIndex;
    327 	OPJ_FLOAT32 lCurrentValue;
    328 	OPJ_FLOAT64 * lNorms = (OPJ_FLOAT64 *) pNorms;
    329 	OPJ_FLOAT32 * lMatrix = (OPJ_FLOAT32 *) pMatrix;
    330 
    331 	for	(i=0;i<pNbComps;++i) {
    332 		lNorms[i] = 0;
    333 		lIndex = i;
    334 
    335 		for	(j=0;j<pNbComps;++j) {
    336 			lCurrentValue = lMatrix[lIndex];
    337 			lIndex += pNbComps;
    338 			lNorms[i] += lCurrentValue * lCurrentValue;
    339 		}
    340 		lNorms[i] = sqrt(lNorms[i]);
    341 	}
    342 }
    343