• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
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) 2007, Jonathan Ballard <dzonatas@dzonux.net>
15  * Copyright (c) 2007, Callum Lerwick <seg@haxxed.com>
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 /** @defgroup DWT DWT - Implementation of a discrete wavelet transform */
47 /*@{*/
48 
49 #define OPJ_WS(i) v->mem[(i)*2]
50 #define OPJ_WD(i) v->mem[(1+(i)*2)]
51 
52 /** @name Local data structures */
53 /*@{*/
54 
55 typedef struct dwt_local {
56 	OPJ_INT32* mem;
57 	OPJ_INT32 dn;
58 	OPJ_INT32 sn;
59 	OPJ_INT32 cas;
60 } opj_dwt_t;
61 
62 typedef union {
63 	OPJ_FLOAT32	f[4];
64 } opj_v4_t;
65 
66 typedef struct v4dwt_local {
67 	opj_v4_t*	wavelet ;
68 	OPJ_INT32		dn ;
69 	OPJ_INT32		sn ;
70 	OPJ_INT32		cas ;
71 } opj_v4dwt_t ;
72 
73 static const OPJ_FLOAT32 opj_dwt_alpha =  1.586134342f; /*  12994 */
74 static const OPJ_FLOAT32 opj_dwt_beta  =  0.052980118f; /*    434 */
75 static const OPJ_FLOAT32 opj_dwt_gamma = -0.882911075f; /*  -7233 */
76 static const OPJ_FLOAT32 opj_dwt_delta = -0.443506852f; /*  -3633 */
77 
78 static const OPJ_FLOAT32 opj_K      = 1.230174105f; /*  10078 */
79 static const OPJ_FLOAT32 opj_c13318 = 1.625732422f;
80 
81 /*@}*/
82 
83 /**
84 Virtual function type for wavelet transform in 1-D
85 */
86 typedef void (*DWT1DFN)(opj_dwt_t* v);
87 
88 /** @name Local static functions */
89 /*@{*/
90 
91 /**
92 Forward lazy transform (horizontal)
93 */
94 static void opj_dwt_deinterleave_h(OPJ_INT32 *a, OPJ_INT32 *b, OPJ_INT32 dn, OPJ_INT32 sn, OPJ_INT32 cas);
95 /**
96 Forward lazy transform (vertical)
97 */
98 static void opj_dwt_deinterleave_v(OPJ_INT32 *a, OPJ_INT32 *b, OPJ_INT32 dn, OPJ_INT32 sn, OPJ_INT32 x, OPJ_INT32 cas);
99 /**
100 Inverse lazy transform (horizontal)
101 */
102 static void opj_dwt_interleave_h(opj_dwt_t* h, OPJ_INT32 *a);
103 /**
104 Inverse lazy transform (vertical)
105 */
106 static void opj_dwt_interleave_v(opj_dwt_t* v, OPJ_INT32 *a, OPJ_INT32 x);
107 /**
108 Forward 5-3 wavelet transform in 1-D
109 */
110 static void opj_dwt_encode_1(OPJ_INT32 *a, OPJ_INT32 dn, OPJ_INT32 sn, OPJ_INT32 cas);
111 /**
112 Inverse 5-3 wavelet transform in 1-D
113 */
114 static void opj_dwt_decode_1(opj_dwt_t *v);
115 static void opj_dwt_decode_1_(OPJ_INT32 *a, OPJ_INT32 dn, OPJ_INT32 sn, OPJ_INT32 cas);
116 /**
117 Forward 9-7 wavelet transform in 1-D
118 */
119 static void opj_dwt_encode_1_real(OPJ_INT32 *a, OPJ_INT32 dn, OPJ_INT32 sn, OPJ_INT32 cas);
120 /**
121 Explicit calculation of the Quantization Stepsizes
122 */
123 static void opj_dwt_encode_stepsize(OPJ_INT32 stepsize, OPJ_INT32 numbps, opj_stepsize_t *bandno_stepsize);
124 /**
125 Inverse wavelet transform in 2-D.
126 */
127 static OPJ_BOOL opj_dwt_decode_tile(opj_tcd_tilecomp_t* tilec, OPJ_UINT32 i, DWT1DFN fn);
128 
129 static OPJ_BOOL opj_dwt_encode_procedure(	opj_tcd_tilecomp_t * tilec,
130 										    void (*p_function)(OPJ_INT32 *, OPJ_INT32,OPJ_INT32,OPJ_INT32) );
131 
132 static OPJ_UINT32 opj_dwt_max_resolution(opj_tcd_resolution_t* restrict r, OPJ_UINT32 i);
133 
134 /* <summary>                             */
135 /* Inverse 9-7 wavelet transform in 1-D. */
136 /* </summary>                            */
137 static void opj_v4dwt_decode(opj_v4dwt_t* restrict dwt);
138 
139 static void opj_v4dwt_interleave_h(opj_v4dwt_t* restrict w, OPJ_FLOAT32* restrict a, OPJ_INT32 x, OPJ_INT32 size);
140 
141 static void opj_v4dwt_interleave_v(opj_v4dwt_t* restrict v , OPJ_FLOAT32* restrict a , OPJ_INT32 x, OPJ_INT32 nb_elts_read);
142 
143 #ifdef __SSE__
144 static void opj_v4dwt_decode_step1_sse(opj_v4_t* w, OPJ_INT32 count, const __m128 c);
145 
146 static void opj_v4dwt_decode_step2_sse(opj_v4_t* l, opj_v4_t* w, OPJ_INT32 k, OPJ_INT32 m, __m128 c);
147 
148 #else
149 static void opj_v4dwt_decode_step1(opj_v4_t* w, OPJ_INT32 count, const OPJ_FLOAT32 c);
150 
151 static void opj_v4dwt_decode_step2(opj_v4_t* l, opj_v4_t* w, OPJ_INT32 k, OPJ_INT32 m, OPJ_FLOAT32 c);
152 
153 #endif
154 
155 /*@}*/
156 
157 /*@}*/
158 
159 #define OPJ_S(i) a[(i)*2]
160 #define OPJ_D(i) a[(1+(i)*2)]
161 #define OPJ_S_(i) ((i)<0?OPJ_S(0):((i)>=sn?OPJ_S(sn-1):OPJ_S(i)))
162 #define OPJ_D_(i) ((i)<0?OPJ_D(0):((i)>=dn?OPJ_D(dn-1):OPJ_D(i)))
163 /* new */
164 #define OPJ_SS_(i) ((i)<0?OPJ_S(0):((i)>=dn?OPJ_S(dn-1):OPJ_S(i)))
165 #define OPJ_DD_(i) ((i)<0?OPJ_D(0):((i)>=sn?OPJ_D(sn-1):OPJ_D(i)))
166 
167 /* <summary>                                                              */
168 /* This table contains the norms of the 5-3 wavelets for different bands. */
169 /* </summary>                                                             */
170 static const OPJ_FLOAT64 opj_dwt_norms[4][10] = {
171 	{1.000, 1.500, 2.750, 5.375, 10.68, 21.34, 42.67, 85.33, 170.7, 341.3},
172 	{1.038, 1.592, 2.919, 5.703, 11.33, 22.64, 45.25, 90.48, 180.9},
173 	{1.038, 1.592, 2.919, 5.703, 11.33, 22.64, 45.25, 90.48, 180.9},
174 	{.7186, .9218, 1.586, 3.043, 6.019, 12.01, 24.00, 47.97, 95.93}
175 };
176 
177 /* <summary>                                                              */
178 /* This table contains the norms of the 9-7 wavelets for different bands. */
179 /* </summary>                                                             */
180 static const OPJ_FLOAT64 opj_dwt_norms_real[4][10] = {
181 	{1.000, 1.965, 4.177, 8.403, 16.90, 33.84, 67.69, 135.3, 270.6, 540.9},
182 	{2.022, 3.989, 8.355, 17.04, 34.27, 68.63, 137.3, 274.6, 549.0},
183 	{2.022, 3.989, 8.355, 17.04, 34.27, 68.63, 137.3, 274.6, 549.0},
184 	{2.080, 3.865, 8.307, 17.18, 34.71, 69.59, 139.3, 278.6, 557.2}
185 };
186 
187 /*
188 ==========================================================
189    local functions
190 ==========================================================
191 */
192 
193 /* <summary>			                 */
194 /* Forward lazy transform (horizontal).  */
195 /* </summary>                            */
opj_dwt_deinterleave_h(OPJ_INT32 * a,OPJ_INT32 * b,OPJ_INT32 dn,OPJ_INT32 sn,OPJ_INT32 cas)196 void opj_dwt_deinterleave_h(OPJ_INT32 *a, OPJ_INT32 *b, OPJ_INT32 dn, OPJ_INT32 sn, OPJ_INT32 cas) {
197 	OPJ_INT32 i;
198 	OPJ_INT32 * l_dest = b;
199 	OPJ_INT32 * l_src = a+cas;
200 
201     for (i=0; i<sn; ++i) {
202 		*l_dest++ = *l_src;
203 		l_src += 2;
204 	}
205 
206     l_dest = b + sn;
207 	l_src = a + 1 - cas;
208 
209     for	(i=0; i<dn; ++i)  {
210 		*l_dest++=*l_src;
211 		l_src += 2;
212 	}
213 }
214 
215 /* <summary>                             */
216 /* Forward lazy transform (vertical).    */
217 /* </summary>                            */
opj_dwt_deinterleave_v(OPJ_INT32 * a,OPJ_INT32 * b,OPJ_INT32 dn,OPJ_INT32 sn,OPJ_INT32 x,OPJ_INT32 cas)218 void opj_dwt_deinterleave_v(OPJ_INT32 *a, OPJ_INT32 *b, OPJ_INT32 dn, OPJ_INT32 sn, OPJ_INT32 x, OPJ_INT32 cas) {
219     OPJ_INT32 i = sn;
220 	OPJ_INT32 * l_dest = b;
221 	OPJ_INT32 * l_src = a+cas;
222 
223     while (i--) {
224 		*l_dest = *l_src;
225 		l_dest += x;
226 		l_src += 2;
227 		} /* b[i*x]=a[2*i+cas]; */
228 
229 	l_dest = b + sn * x;
230 	l_src = a + 1 - cas;
231 
232 	i = dn;
233     while (i--) {
234 		*l_dest = *l_src;
235 		l_dest += x;
236 		l_src += 2;
237         } /*b[(sn+i)*x]=a[(2*i+1-cas)];*/
238 }
239 
240 /* <summary>                             */
241 /* Inverse lazy transform (horizontal).  */
242 /* </summary>                            */
opj_dwt_interleave_h(opj_dwt_t * h,OPJ_INT32 * a)243 void opj_dwt_interleave_h(opj_dwt_t* h, OPJ_INT32 *a) {
244     OPJ_INT32 *ai = a;
245     OPJ_INT32 *bi = h->mem + h->cas;
246     OPJ_INT32  i	= h->sn;
247     while( i-- ) {
248       *bi = *(ai++);
249 	  bi += 2;
250     }
251     ai	= a + h->sn;
252     bi	= h->mem + 1 - h->cas;
253     i	= h->dn ;
254     while( i-- ) {
255       *bi = *(ai++);
256 	  bi += 2;
257     }
258 }
259 
260 /* <summary>                             */
261 /* Inverse lazy transform (vertical).    */
262 /* </summary>                            */
opj_dwt_interleave_v(opj_dwt_t * v,OPJ_INT32 * a,OPJ_INT32 x)263 void opj_dwt_interleave_v(opj_dwt_t* v, OPJ_INT32 *a, OPJ_INT32 x) {
264     OPJ_INT32 *ai = a;
265     OPJ_INT32 *bi = v->mem + v->cas;
266     OPJ_INT32  i = v->sn;
267     while( i-- ) {
268       *bi = *ai;
269 	  bi += 2;
270 	  ai += x;
271     }
272     ai = a + (v->sn * x);
273     bi = v->mem + 1 - v->cas;
274     i = v->dn ;
275     while( i-- ) {
276       *bi = *ai;
277 	  bi += 2;
278 	  ai += x;
279     }
280 }
281 
282 
283 /* <summary>                            */
284 /* Forward 5-3 wavelet transform in 1-D. */
285 /* </summary>                           */
opj_dwt_encode_1(OPJ_INT32 * a,OPJ_INT32 dn,OPJ_INT32 sn,OPJ_INT32 cas)286 void opj_dwt_encode_1(OPJ_INT32 *a, OPJ_INT32 dn, OPJ_INT32 sn, OPJ_INT32 cas) {
287 	OPJ_INT32 i;
288 
289 	if (!cas) {
290 		if ((dn > 0) || (sn > 1)) {	/* NEW :  CASE ONE ELEMENT */
291 			for (i = 0; i < dn; i++) OPJ_D(i) -= (OPJ_S_(i) + OPJ_S_(i + 1)) >> 1;
292 			for (i = 0; i < sn; i++) OPJ_S(i) += (OPJ_D_(i - 1) + OPJ_D_(i) + 2) >> 2;
293 		}
294 	} else {
295 		if (!sn && dn == 1)		    /* NEW :  CASE ONE ELEMENT */
296 			OPJ_S(0) *= 2;
297 		else {
298 			for (i = 0; i < dn; i++) OPJ_S(i) -= (OPJ_DD_(i) + OPJ_DD_(i - 1)) >> 1;
299 			for (i = 0; i < sn; i++) OPJ_D(i) += (OPJ_SS_(i) + OPJ_SS_(i + 1) + 2) >> 2;
300 		}
301 	}
302 }
303 
304 /* <summary>                            */
305 /* Inverse 5-3 wavelet transform in 1-D. */
306 /* </summary>                           */
opj_dwt_decode_1_(OPJ_INT32 * a,OPJ_INT32 dn,OPJ_INT32 sn,OPJ_INT32 cas)307 void opj_dwt_decode_1_(OPJ_INT32 *a, OPJ_INT32 dn, OPJ_INT32 sn, OPJ_INT32 cas) {
308 	OPJ_INT32 i;
309 
310 	if (!cas) {
311 		if ((dn > 0) || (sn > 1)) { /* NEW :  CASE ONE ELEMENT */
312 			for (i = 0; i < sn; i++) OPJ_S(i) -= (OPJ_D_(i - 1) + OPJ_D_(i) + 2) >> 2;
313 			for (i = 0; i < dn; i++) OPJ_D(i) += (OPJ_S_(i) + OPJ_S_(i + 1)) >> 1;
314 		}
315 	} else {
316 		if (!sn  && dn == 1)          /* NEW :  CASE ONE ELEMENT */
317 			OPJ_S(0) /= 2;
318 		else {
319 			for (i = 0; i < sn; i++) OPJ_D(i) -= (OPJ_SS_(i) + OPJ_SS_(i + 1) + 2) >> 2;
320 			for (i = 0; i < dn; i++) OPJ_S(i) += (OPJ_DD_(i) + OPJ_DD_(i - 1)) >> 1;
321 		}
322 	}
323 }
324 
325 /* <summary>                            */
326 /* Inverse 5-3 wavelet transform in 1-D. */
327 /* </summary>                           */
opj_dwt_decode_1(opj_dwt_t * v)328 void opj_dwt_decode_1(opj_dwt_t *v) {
329 	opj_dwt_decode_1_(v->mem, v->dn, v->sn, v->cas);
330 }
331 
332 /* <summary>                             */
333 /* Forward 9-7 wavelet transform in 1-D. */
334 /* </summary>                            */
opj_dwt_encode_1_real(OPJ_INT32 * a,OPJ_INT32 dn,OPJ_INT32 sn,OPJ_INT32 cas)335 void opj_dwt_encode_1_real(OPJ_INT32 *a, OPJ_INT32 dn, OPJ_INT32 sn, OPJ_INT32 cas) {
336 	OPJ_INT32 i;
337 	if (!cas) {
338 		if ((dn > 0) || (sn > 1)) {	/* NEW :  CASE ONE ELEMENT */
339 			for (i = 0; i < dn; i++)
340 				OPJ_D(i) -= opj_int_fix_mul(OPJ_S_(i) + OPJ_S_(i + 1), 12993);
341 			for (i = 0; i < sn; i++)
342 				OPJ_S(i) -= opj_int_fix_mul(OPJ_D_(i - 1) + OPJ_D_(i), 434);
343 			for (i = 0; i < dn; i++)
344 				OPJ_D(i) += opj_int_fix_mul(OPJ_S_(i) + OPJ_S_(i + 1), 7233);
345 			for (i = 0; i < sn; i++)
346 				OPJ_S(i) += opj_int_fix_mul(OPJ_D_(i - 1) + OPJ_D_(i), 3633);
347 			for (i = 0; i < dn; i++)
348 				OPJ_D(i) = opj_int_fix_mul(OPJ_D(i), 5038);	/*5038 */
349 			for (i = 0; i < sn; i++)
350 				OPJ_S(i) = opj_int_fix_mul(OPJ_S(i), 6659);	/*6660 */
351 		}
352 	} else {
353 		if ((sn > 0) || (dn > 1)) {	/* NEW :  CASE ONE ELEMENT */
354 			for (i = 0; i < dn; i++)
355 				OPJ_S(i) -= opj_int_fix_mul(OPJ_DD_(i) + OPJ_DD_(i - 1), 12993);
356 			for (i = 0; i < sn; i++)
357 				OPJ_D(i) -= opj_int_fix_mul(OPJ_SS_(i) + OPJ_SS_(i + 1), 434);
358 			for (i = 0; i < dn; i++)
359 				OPJ_S(i) += opj_int_fix_mul(OPJ_DD_(i) + OPJ_DD_(i - 1), 7233);
360 			for (i = 0; i < sn; i++)
361 				OPJ_D(i) += opj_int_fix_mul(OPJ_SS_(i) + OPJ_SS_(i + 1), 3633);
362 			for (i = 0; i < dn; i++)
363 				OPJ_S(i) = opj_int_fix_mul(OPJ_S(i), 5038);	/*5038 */
364 			for (i = 0; i < sn; i++)
365 				OPJ_D(i) = opj_int_fix_mul(OPJ_D(i), 6659);	/*6660 */
366 		}
367 	}
368 }
369 
opj_dwt_encode_stepsize(OPJ_INT32 stepsize,OPJ_INT32 numbps,opj_stepsize_t * bandno_stepsize)370 void opj_dwt_encode_stepsize(OPJ_INT32 stepsize, OPJ_INT32 numbps, opj_stepsize_t *bandno_stepsize) {
371 	OPJ_INT32 p, n;
372 	p = opj_int_floorlog2(stepsize) - 13;
373 	n = 11 - opj_int_floorlog2(stepsize);
374 	bandno_stepsize->mant = (n < 0 ? stepsize >> -n : stepsize << n) & 0x7ff;
375 	bandno_stepsize->expn = numbps - p;
376 }
377 
378 /*
379 ==========================================================
380    DWT interface
381 ==========================================================
382 */
383 
384 
385 /* <summary>                            */
386 /* Forward 5-3 wavelet transform in 2-D. */
387 /* </summary>                           */
opj_dwt_encode_procedure(opj_tcd_tilecomp_t * tilec,void (* p_function)(OPJ_INT32 *,OPJ_INT32,OPJ_INT32,OPJ_INT32))388 INLINE OPJ_BOOL opj_dwt_encode_procedure(opj_tcd_tilecomp_t * tilec,void (*p_function)(OPJ_INT32 *, OPJ_INT32,OPJ_INT32,OPJ_INT32) )
389 {
390 	OPJ_INT32 i, j, k;
391 	OPJ_INT32 *a = 00;
392 	OPJ_INT32 *aj = 00;
393 	OPJ_INT32 *bj = 00;
394 	OPJ_INT32 w, l;
395 
396 	OPJ_INT32 rw;			/* width of the resolution level computed   */
397 	OPJ_INT32 rh;			/* height of the resolution level computed  */
398 	OPJ_UINT32 l_data_size;
399 
400 	opj_tcd_resolution_t * l_cur_res = 0;
401 	opj_tcd_resolution_t * l_last_res = 0;
402 
403 	w = tilec->x1-tilec->x0;
404 	l = (OPJ_INT32)tilec->numresolutions-1;
405 	a = tilec->data;
406 
407 	l_cur_res = tilec->resolutions + l;
408 	l_last_res = l_cur_res - 1;
409 
410 	l_data_size = opj_dwt_max_resolution( tilec->resolutions,tilec->numresolutions) * (OPJ_UINT32)sizeof(OPJ_INT32);
411 	bj = (OPJ_INT32*)opj_malloc((size_t)l_data_size);
412 	if (! bj) {
413 		return OPJ_FALSE;
414 	}
415 	i = l;
416 
417 	while (i--) {
418 		OPJ_INT32 rw1;		/* width of the resolution level once lower than computed one                                       */
419 		OPJ_INT32 rh1;		/* height of the resolution level once lower than computed one                                      */
420 		OPJ_INT32 cas_col;	/* 0 = non inversion on horizontal filtering 1 = inversion between low-pass and high-pass filtering */
421 		OPJ_INT32 cas_row;	/* 0 = non inversion on vertical filtering 1 = inversion between low-pass and high-pass filtering   */
422 		OPJ_INT32 dn, sn;
423 
424 		rw  = l_cur_res->x1 - l_cur_res->x0;
425 		rh  = l_cur_res->y1 - l_cur_res->y0;
426 		rw1 = l_last_res->x1 - l_last_res->x0;
427 		rh1 = l_last_res->y1 - l_last_res->y0;
428 
429 		cas_row = l_cur_res->x0 & 1;
430 		cas_col = l_cur_res->y0 & 1;
431 
432 		sn = rh1;
433 		dn = rh - rh1;
434 		for (j = 0; j < rw; ++j) {
435 			aj = a + j;
436 			for (k = 0; k < rh; ++k) {
437 				bj[k] = aj[k*w];
438 			}
439 
440 			(*p_function) (bj, dn, sn, cas_col);
441 
442 			opj_dwt_deinterleave_v(bj, aj, dn, sn, w, cas_col);
443 		}
444 
445 		sn = rw1;
446 		dn = rw - rw1;
447 
448 		for (j = 0; j < rh; j++) {
449 			aj = a + j * w;
450 			for (k = 0; k < rw; k++)  bj[k] = aj[k];
451 			(*p_function) (bj, dn, sn, cas_row);
452 			opj_dwt_deinterleave_h(bj, aj, dn, sn, cas_row);
453 		}
454 
455 		l_cur_res = l_last_res;
456 
457 		--l_last_res;
458 	}
459 
460 	opj_free(bj);
461 	return OPJ_TRUE;
462 }
463 
464 /* Forward 5-3 wavelet transform in 2-D. */
465 /* </summary>                           */
opj_dwt_encode(opj_tcd_tilecomp_t * tilec)466 OPJ_BOOL opj_dwt_encode(opj_tcd_tilecomp_t * tilec)
467 {
468 	return opj_dwt_encode_procedure(tilec,opj_dwt_encode_1);
469 }
470 
471 /* <summary>                            */
472 /* Inverse 5-3 wavelet transform in 2-D. */
473 /* </summary>                           */
opj_dwt_decode(opj_tcd_tilecomp_t * tilec,OPJ_UINT32 numres)474 OPJ_BOOL opj_dwt_decode(opj_tcd_tilecomp_t* tilec, OPJ_UINT32 numres) {
475 	return opj_dwt_decode_tile(tilec, numres, &opj_dwt_decode_1);
476 }
477 
478 
479 /* <summary>                          */
480 /* Get gain of 5-3 wavelet transform. */
481 /* </summary>                         */
opj_dwt_getgain(OPJ_UINT32 orient)482 OPJ_UINT32 opj_dwt_getgain(OPJ_UINT32 orient) {
483 	if (orient == 0)
484 		return 0;
485 	if (orient == 1 || orient == 2)
486 		return 1;
487 	return 2;
488 }
489 
490 /* <summary>                */
491 /* Get norm of 5-3 wavelet. */
492 /* </summary>               */
opj_dwt_getnorm(OPJ_UINT32 level,OPJ_UINT32 orient)493 OPJ_FLOAT64 opj_dwt_getnorm(OPJ_UINT32 level, OPJ_UINT32 orient) {
494 	return opj_dwt_norms[orient][level];
495 }
496 
497 /* <summary>                             */
498 /* Forward 9-7 wavelet transform in 2-D. */
499 /* </summary>                            */
opj_dwt_encode_real(opj_tcd_tilecomp_t * tilec)500 OPJ_BOOL opj_dwt_encode_real(opj_tcd_tilecomp_t * tilec)
501 {
502 	return opj_dwt_encode_procedure(tilec,opj_dwt_encode_1_real);
503 }
504 
505 /* <summary>                          */
506 /* Get gain of 9-7 wavelet transform. */
507 /* </summary>                         */
opj_dwt_getgain_real(OPJ_UINT32 orient)508 OPJ_UINT32 opj_dwt_getgain_real(OPJ_UINT32 orient) {
509 	(void)orient;
510 	return 0;
511 }
512 
513 /* <summary>                */
514 /* Get norm of 9-7 wavelet. */
515 /* </summary>               */
opj_dwt_getnorm_real(OPJ_UINT32 level,OPJ_UINT32 orient)516 OPJ_FLOAT64 opj_dwt_getnorm_real(OPJ_UINT32 level, OPJ_UINT32 orient) {
517 	return opj_dwt_norms_real[orient][level];
518 }
519 
opj_dwt_calc_explicit_stepsizes(opj_tccp_t * tccp,OPJ_UINT32 prec)520 void opj_dwt_calc_explicit_stepsizes(opj_tccp_t * tccp, OPJ_UINT32 prec) {
521 	OPJ_UINT32 numbands, bandno;
522 	numbands = 3 * tccp->numresolutions - 2;
523 	for (bandno = 0; bandno < numbands; bandno++) {
524 		OPJ_FLOAT64 stepsize;
525 		OPJ_UINT32 resno, level, orient, gain;
526 
527 		resno = (bandno == 0) ? 0 : ((bandno - 1) / 3 + 1);
528 		orient = (bandno == 0) ? 0 : ((bandno - 1) % 3 + 1);
529 		level = tccp->numresolutions - 1 - resno;
530 		gain = (tccp->qmfbid == 0) ? 0 : ((orient == 0) ? 0 : (((orient == 1) || (orient == 2)) ? 1 : 2));
531 		if (tccp->qntsty == J2K_CCP_QNTSTY_NOQNT) {
532 			stepsize = 1.0;
533 		} else {
534 			OPJ_FLOAT64 norm = opj_dwt_norms_real[orient][level];
535 			stepsize = (1 << (gain)) / norm;
536 		}
537 		opj_dwt_encode_stepsize((OPJ_INT32) floor(stepsize * 8192.0), (OPJ_INT32)(prec + gain), &tccp->stepsizes[bandno]);
538 	}
539 }
540 
541 /* <summary>                             */
542 /* Determine maximum computed resolution level for inverse wavelet transform */
543 /* </summary>                            */
opj_dwt_max_resolution(opj_tcd_resolution_t * restrict r,OPJ_UINT32 i)544 OPJ_UINT32 opj_dwt_max_resolution(opj_tcd_resolution_t* restrict r, OPJ_UINT32 i) {
545 	OPJ_UINT32 mr	= 0;
546 	OPJ_UINT32 w;
547 	while( --i ) {
548 		++r;
549 		if( mr < ( w = (OPJ_UINT32)(r->x1 - r->x0) ) )
550 			mr = w ;
551 		if( mr < ( w = (OPJ_UINT32)(r->y1 - r->y0) ) )
552 			mr = w ;
553 	}
554 	return mr ;
555 }
556 
557 /* <summary>                            */
558 /* Inverse wavelet transform in 2-D.     */
559 /* </summary>                           */
opj_dwt_decode_tile(opj_tcd_tilecomp_t * tilec,OPJ_UINT32 numres,DWT1DFN dwt_1D)560 OPJ_BOOL opj_dwt_decode_tile(opj_tcd_tilecomp_t* tilec, OPJ_UINT32 numres, DWT1DFN dwt_1D) {
561 	opj_dwt_t h;
562 	opj_dwt_t v;
563 
564 	opj_tcd_resolution_t* tr = tilec->resolutions;
565 
566 	OPJ_UINT32 rw = (OPJ_UINT32)(tr->x1 - tr->x0);	/* width of the resolution level computed */
567 	OPJ_UINT32 rh = (OPJ_UINT32)(tr->y1 - tr->y0);	/* height of the resolution level computed */
568 
569 	OPJ_UINT32 w = (OPJ_UINT32)(tilec->x1 - tilec->x0);
570 
571 	h.mem = (OPJ_INT32*)
572 	opj_aligned_malloc(opj_dwt_max_resolution(tr, numres) * sizeof(OPJ_INT32));
573 	if (! h.mem){
574 		return OPJ_FALSE;
575 	}
576 
577 	v.mem = h.mem;
578 
579 	while( --numres) {
580 		OPJ_INT32 * restrict tiledp = tilec->data;
581 		OPJ_UINT32 j;
582 
583 		++tr;
584 		h.sn = (OPJ_INT32)rw;
585 		v.sn = (OPJ_INT32)rh;
586 
587 		rw = (OPJ_UINT32)(tr->x1 - tr->x0);
588 		rh = (OPJ_UINT32)(tr->y1 - tr->y0);
589 
590 		h.dn = (OPJ_INT32)(rw - (OPJ_UINT32)h.sn);
591 		h.cas = tr->x0 % 2;
592 
593 		for(j = 0; j < rh; ++j) {
594 			opj_dwt_interleave_h(&h, &tiledp[j*w]);
595 			(dwt_1D)(&h);
596 			memcpy(&tiledp[j*w], h.mem, rw * sizeof(OPJ_INT32));
597 		}
598 
599 		v.dn = (OPJ_INT32)(rh - (OPJ_UINT32)v.sn);
600 		v.cas = tr->y0 % 2;
601 
602 		for(j = 0; j < rw; ++j){
603 			OPJ_UINT32 k;
604 			opj_dwt_interleave_v(&v, &tiledp[j], (OPJ_INT32)w);
605 			(dwt_1D)(&v);
606 			for(k = 0; k < rh; ++k) {
607 				tiledp[k * w + j] = v.mem[k];
608 			}
609 		}
610 	}
611 	opj_aligned_free(h.mem);
612 	return OPJ_TRUE;
613 }
614 
opj_v4dwt_interleave_h(opj_v4dwt_t * restrict w,OPJ_FLOAT32 * restrict a,OPJ_INT32 x,OPJ_INT32 size)615 void opj_v4dwt_interleave_h(opj_v4dwt_t* restrict w, OPJ_FLOAT32* restrict a, OPJ_INT32 x, OPJ_INT32 size){
616 	OPJ_FLOAT32* restrict bi = (OPJ_FLOAT32*) (w->wavelet + w->cas);
617 	OPJ_INT32 count = w->sn;
618 	OPJ_INT32 i, k;
619 
620 	for(k = 0; k < 2; ++k){
621 		if ( count + 3 * x < size && ((size_t) a & 0x0f) == 0 && ((size_t) bi & 0x0f) == 0 && (x & 0x0f) == 0 ) {
622 			/* Fast code path */
623 			for(i = 0; i < count; ++i){
624 				OPJ_INT32 j = i;
625 				bi[i*8    ] = a[j];
626 				j += x;
627 				bi[i*8 + 1] = a[j];
628 				j += x;
629 				bi[i*8 + 2] = a[j];
630 				j += x;
631 				bi[i*8 + 3] = a[j];
632 			}
633 		}
634 		else {
635 			/* Slow code path */
636 			for(i = 0; i < count; ++i){
637 				OPJ_INT32 j = i;
638 				bi[i*8    ] = a[j];
639 				j += x;
640 				if(j >= size) continue;
641 				bi[i*8 + 1] = a[j];
642 				j += x;
643 				if(j >= size) continue;
644 				bi[i*8 + 2] = a[j];
645 				j += x;
646 				if(j >= size) continue;
647 				bi[i*8 + 3] = a[j]; /* This one*/
648 			}
649 		}
650 
651 		bi = (OPJ_FLOAT32*) (w->wavelet + 1 - w->cas);
652 		a += w->sn;
653 		size -= w->sn;
654 		count = w->dn;
655 	}
656 }
657 
opj_v4dwt_interleave_v(opj_v4dwt_t * restrict v,OPJ_FLOAT32 * restrict a,OPJ_INT32 x,OPJ_INT32 nb_elts_read)658 void opj_v4dwt_interleave_v(opj_v4dwt_t* restrict v , OPJ_FLOAT32* restrict a , OPJ_INT32 x, OPJ_INT32 nb_elts_read){
659 	opj_v4_t* restrict bi = v->wavelet + v->cas;
660 	OPJ_INT32 i;
661 
662 	for(i = 0; i < v->sn; ++i){
663 		memcpy(&bi[i*2], &a[i*x], (size_t)nb_elts_read * sizeof(OPJ_FLOAT32));
664 	}
665 
666 	a += v->sn * x;
667 	bi = v->wavelet + 1 - v->cas;
668 
669 	for(i = 0; i < v->dn; ++i){
670 		memcpy(&bi[i*2], &a[i*x], (size_t)nb_elts_read * sizeof(OPJ_FLOAT32));
671 	}
672 }
673 
674 #ifdef __SSE__
675 
opj_v4dwt_decode_step1_sse(opj_v4_t * w,OPJ_INT32 count,const __m128 c)676 void opj_v4dwt_decode_step1_sse(opj_v4_t* w, OPJ_INT32 count, const __m128 c){
677 	__m128* restrict vw = (__m128*) w;
678 	OPJ_INT32 i;
679 	/* 4x unrolled loop */
680 	for(i = 0; i < count >> 2; ++i){
681 		*vw = _mm_mul_ps(*vw, c);
682 		vw += 2;
683 		*vw = _mm_mul_ps(*vw, c);
684 		vw += 2;
685 		*vw = _mm_mul_ps(*vw, c);
686 		vw += 2;
687 		*vw = _mm_mul_ps(*vw, c);
688 		vw += 2;
689 	}
690 	count &= 3;
691 	for(i = 0; i < count; ++i){
692 		*vw = _mm_mul_ps(*vw, c);
693 		vw += 2;
694 	}
695 }
696 
opj_v4dwt_decode_step2_sse(opj_v4_t * l,opj_v4_t * w,OPJ_INT32 k,OPJ_INT32 m,__m128 c)697 void opj_v4dwt_decode_step2_sse(opj_v4_t* l, opj_v4_t* w, OPJ_INT32 k, OPJ_INT32 m, __m128 c){
698 	__m128* restrict vl = (__m128*) l;
699 	__m128* restrict vw = (__m128*) w;
700 	OPJ_INT32 i;
701 	__m128 tmp1, tmp2, tmp3;
702 	tmp1 = vl[0];
703 	for(i = 0; i < m; ++i){
704 		tmp2 = vw[-1];
705 		tmp3 = vw[ 0];
706 		vw[-1] = _mm_add_ps(tmp2, _mm_mul_ps(_mm_add_ps(tmp1, tmp3), c));
707 		tmp1 = tmp3;
708 		vw += 2;
709 	}
710 	vl = vw - 2;
711 	if(m >= k){
712 		return;
713 	}
714 	c = _mm_add_ps(c, c);
715 	c = _mm_mul_ps(c, vl[0]);
716 	for(; m < k; ++m){
717 		__m128 tmp = vw[-1];
718 		vw[-1] = _mm_add_ps(tmp, c);
719 		vw += 2;
720 	}
721 }
722 
723 #else
724 
opj_v4dwt_decode_step1(opj_v4_t * w,OPJ_INT32 count,const OPJ_FLOAT32 c)725 void opj_v4dwt_decode_step1(opj_v4_t* w, OPJ_INT32 count, const OPJ_FLOAT32 c)
726 {
727 	OPJ_FLOAT32* restrict fw = (OPJ_FLOAT32*) w;
728 	OPJ_INT32 i;
729 	for(i = 0; i < count; ++i){
730 		OPJ_FLOAT32 tmp1 = fw[i*8    ];
731 		OPJ_FLOAT32 tmp2 = fw[i*8 + 1];
732 		OPJ_FLOAT32 tmp3 = fw[i*8 + 2];
733 		OPJ_FLOAT32 tmp4 = fw[i*8 + 3];
734 		fw[i*8    ] = tmp1 * c;
735 		fw[i*8 + 1] = tmp2 * c;
736 		fw[i*8 + 2] = tmp3 * c;
737 		fw[i*8 + 3] = tmp4 * c;
738 	}
739 }
740 
opj_v4dwt_decode_step2(opj_v4_t * l,opj_v4_t * w,OPJ_INT32 k,OPJ_INT32 m,OPJ_FLOAT32 c)741 void opj_v4dwt_decode_step2(opj_v4_t* l, opj_v4_t* w, OPJ_INT32 k, OPJ_INT32 m, OPJ_FLOAT32 c)
742 {
743 	OPJ_FLOAT32* restrict fl = (OPJ_FLOAT32*) l;
744 	OPJ_FLOAT32* restrict fw = (OPJ_FLOAT32*) w;
745 	OPJ_INT32 i;
746 	for(i = 0; i < m; ++i){
747 		OPJ_FLOAT32 tmp1_1 = fl[0];
748 		OPJ_FLOAT32 tmp1_2 = fl[1];
749 		OPJ_FLOAT32 tmp1_3 = fl[2];
750 		OPJ_FLOAT32 tmp1_4 = fl[3];
751 		OPJ_FLOAT32 tmp2_1 = fw[-4];
752 		OPJ_FLOAT32 tmp2_2 = fw[-3];
753 		OPJ_FLOAT32 tmp2_3 = fw[-2];
754 		OPJ_FLOAT32 tmp2_4 = fw[-1];
755 		OPJ_FLOAT32 tmp3_1 = fw[0];
756 		OPJ_FLOAT32 tmp3_2 = fw[1];
757 		OPJ_FLOAT32 tmp3_3 = fw[2];
758 		OPJ_FLOAT32 tmp3_4 = fw[3];
759 		fw[-4] = tmp2_1 + ((tmp1_1 + tmp3_1) * c);
760 		fw[-3] = tmp2_2 + ((tmp1_2 + tmp3_2) * c);
761 		fw[-2] = tmp2_3 + ((tmp1_3 + tmp3_3) * c);
762 		fw[-1] = tmp2_4 + ((tmp1_4 + tmp3_4) * c);
763 		fl = fw;
764 		fw += 8;
765 	}
766 	if(m < k){
767 		OPJ_FLOAT32 c1;
768 		OPJ_FLOAT32 c2;
769 		OPJ_FLOAT32 c3;
770 		OPJ_FLOAT32 c4;
771 		c += c;
772 		c1 = fl[0] * c;
773 		c2 = fl[1] * c;
774 		c3 = fl[2] * c;
775 		c4 = fl[3] * c;
776 		for(; m < k; ++m){
777 			OPJ_FLOAT32 tmp1 = fw[-4];
778 			OPJ_FLOAT32 tmp2 = fw[-3];
779 			OPJ_FLOAT32 tmp3 = fw[-2];
780 			OPJ_FLOAT32 tmp4 = fw[-1];
781 			fw[-4] = tmp1 + c1;
782 			fw[-3] = tmp2 + c2;
783 			fw[-2] = tmp3 + c3;
784 			fw[-1] = tmp4 + c4;
785 			fw += 8;
786 		}
787 	}
788 }
789 
790 #endif
791 
792 /* <summary>                             */
793 /* Inverse 9-7 wavelet transform in 1-D. */
794 /* </summary>                            */
opj_v4dwt_decode(opj_v4dwt_t * restrict dwt)795 void opj_v4dwt_decode(opj_v4dwt_t* restrict dwt)
796 {
797 	OPJ_INT32 a, b;
798 	if(dwt->cas == 0) {
799 		if(!((dwt->dn > 0) || (dwt->sn > 1))){
800 			return;
801 		}
802 		a = 0;
803 		b = 1;
804 	}else{
805 		if(!((dwt->sn > 0) || (dwt->dn > 1))) {
806 			return;
807 		}
808 		a = 1;
809 		b = 0;
810 	}
811 #ifdef __SSE__
812 	opj_v4dwt_decode_step1_sse(dwt->wavelet+a, dwt->sn, _mm_set1_ps(opj_K));
813 	opj_v4dwt_decode_step1_sse(dwt->wavelet+b, dwt->dn, _mm_set1_ps(opj_c13318));
814 	opj_v4dwt_decode_step2_sse(dwt->wavelet+b, dwt->wavelet+a+1, dwt->sn, opj_int_min(dwt->sn, dwt->dn-a), _mm_set1_ps(opj_dwt_delta));
815 	opj_v4dwt_decode_step2_sse(dwt->wavelet+a, dwt->wavelet+b+1, dwt->dn, opj_int_min(dwt->dn, dwt->sn-b), _mm_set1_ps(opj_dwt_gamma));
816 	opj_v4dwt_decode_step2_sse(dwt->wavelet+b, dwt->wavelet+a+1, dwt->sn, opj_int_min(dwt->sn, dwt->dn-a), _mm_set1_ps(opj_dwt_beta));
817 	opj_v4dwt_decode_step2_sse(dwt->wavelet+a, dwt->wavelet+b+1, dwt->dn, opj_int_min(dwt->dn, dwt->sn-b), _mm_set1_ps(opj_dwt_alpha));
818 #else
819 	opj_v4dwt_decode_step1(dwt->wavelet+a, dwt->sn, opj_K);
820 	opj_v4dwt_decode_step1(dwt->wavelet+b, dwt->dn, opj_c13318);
821 	opj_v4dwt_decode_step2(dwt->wavelet+b, dwt->wavelet+a+1, dwt->sn, opj_int_min(dwt->sn, dwt->dn-a), opj_dwt_delta);
822 	opj_v4dwt_decode_step2(dwt->wavelet+a, dwt->wavelet+b+1, dwt->dn, opj_int_min(dwt->dn, dwt->sn-b), opj_dwt_gamma);
823 	opj_v4dwt_decode_step2(dwt->wavelet+b, dwt->wavelet+a+1, dwt->sn, opj_int_min(dwt->sn, dwt->dn-a), opj_dwt_beta);
824 	opj_v4dwt_decode_step2(dwt->wavelet+a, dwt->wavelet+b+1, dwt->dn, opj_int_min(dwt->dn, dwt->sn-b), opj_dwt_alpha);
825 #endif
826 }
827 
828 
829 /* <summary>                             */
830 /* Inverse 9-7 wavelet transform in 2-D. */
831 /* </summary>                            */
opj_dwt_decode_real(opj_tcd_tilecomp_t * restrict tilec,OPJ_UINT32 numres)832 OPJ_BOOL opj_dwt_decode_real(opj_tcd_tilecomp_t* restrict tilec, OPJ_UINT32 numres)
833 {
834 	opj_v4dwt_t h;
835 	opj_v4dwt_t v;
836 
837 	opj_tcd_resolution_t* res = tilec->resolutions;
838 
839 	OPJ_UINT32 rw = (OPJ_UINT32)(res->x1 - res->x0);	/* width of the resolution level computed */
840 	OPJ_UINT32 rh = (OPJ_UINT32)(res->y1 - res->y0);	/* height of the resolution level computed */
841 
842 	OPJ_UINT32 w = (OPJ_UINT32)(tilec->x1 - tilec->x0);
843 
844 	h.wavelet = (opj_v4_t*) opj_aligned_malloc((opj_dwt_max_resolution(res, numres)+5) * sizeof(opj_v4_t));
845 	v.wavelet = h.wavelet;
846 
847 	while( --numres) {
848 		OPJ_FLOAT32 * restrict aj = (OPJ_FLOAT32*) tilec->data;
849 		OPJ_UINT32 bufsize = (OPJ_UINT32)((tilec->x1 - tilec->x0) * (tilec->y1 - tilec->y0));
850 		OPJ_INT32 j;
851 
852 		h.sn = (OPJ_INT32)rw;
853 		v.sn = (OPJ_INT32)rh;
854 
855 		++res;
856 
857 		rw = (OPJ_UINT32)(res->x1 - res->x0);	/* width of the resolution level computed */
858 		rh = (OPJ_UINT32)(res->y1 - res->y0);	/* height of the resolution level computed */
859 
860 		h.dn = (OPJ_INT32)(rw - (OPJ_UINT32)h.sn);
861 		h.cas = res->x0 % 2;
862 
863 		for(j = (OPJ_INT32)rh; j > 3; j -= 4) {
864 			OPJ_INT32 k;
865 			opj_v4dwt_interleave_h(&h, aj, (OPJ_INT32)w, (OPJ_INT32)bufsize);
866 			opj_v4dwt_decode(&h);
867 
868 			for(k = (OPJ_INT32)rw; --k >= 0;){
869 				aj[k               ] = h.wavelet[k].f[0];
870 				aj[k+(OPJ_INT32)w  ] = h.wavelet[k].f[1];
871 				aj[k+(OPJ_INT32)w*2] = h.wavelet[k].f[2];
872 				aj[k+(OPJ_INT32)w*3] = h.wavelet[k].f[3];
873 			}
874 
875 			aj += w*4;
876 			bufsize -= w*4;
877 		}
878 
879 		if (rh & 0x03) {
880 			OPJ_INT32 k;
881 			j = rh & 0x03;
882 			opj_v4dwt_interleave_h(&h, aj, (OPJ_INT32)w, (OPJ_INT32)bufsize);
883 			opj_v4dwt_decode(&h);
884 			for(k = (OPJ_INT32)rw; --k >= 0;){
885 				switch(j) {
886 					case 3: aj[k+(OPJ_INT32)w*2] = h.wavelet[k].f[2];
887 					case 2: aj[k+(OPJ_INT32)w  ] = h.wavelet[k].f[1];
888 					case 1: aj[k               ] = h.wavelet[k].f[0];
889 				}
890 			}
891 		}
892 
893 		v.dn = (OPJ_INT32)(rh - (OPJ_UINT32)v.sn);
894 		v.cas = res->y0 % 2;
895 
896 		aj = (OPJ_FLOAT32*) tilec->data;
897 		for(j = (OPJ_INT32)rw; j > 3; j -= 4){
898 			OPJ_UINT32 k;
899 
900 			opj_v4dwt_interleave_v(&v, aj, (OPJ_INT32)w, 4);
901 			opj_v4dwt_decode(&v);
902 
903 			for(k = 0; k < rh; ++k){
904 				memcpy(&aj[k*w], &v.wavelet[k], 4 * sizeof(OPJ_FLOAT32));
905 			}
906 			aj += 4;
907 		}
908 
909 		if (rw & 0x03){
910 			OPJ_UINT32 k;
911 
912 			j = rw & 0x03;
913 
914 			opj_v4dwt_interleave_v(&v, aj, (OPJ_INT32)w, j);
915 			opj_v4dwt_decode(&v);
916 
917 			for(k = 0; k < rh; ++k){
918 				memcpy(&aj[k*w], &v.wavelet[k], (size_t)j * sizeof(OPJ_FLOAT32));
919 			}
920 		}
921 	}
922 
923 	opj_aligned_free(h.wavelet);
924 	return OPJ_TRUE;
925 }
926