4 * Copyright (c) 2001-2003 BERO <bero@geocities.co.jp>
6 * This file is part of FFmpeg.
8 * FFmpeg is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License as published by the Free Software Foundation; either
11 * version 2.1 of the License, or (at your option) any later version.
13 * FFmpeg is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * Lesser General Public License for more details.
18 * You should have received a copy of the GNU Lesser General Public
19 * License along with FFmpeg; if not, write to the Free Software
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
23 #include "libavcodec/dsputil.h"
24 #include "dsputil_sh4.h"
27 #define c1 1.38703984532214752434 /* sqrt(2)*cos(1*pi/16) */
28 #define c2 1.30656296487637657577 /* sqrt(2)*cos(2*pi/16) */
29 #define c3 1.17587560241935884520 /* sqrt(2)*cos(3*pi/16) */
30 #define c4 1.00000000000000000000 /* sqrt(2)*cos(4*pi/16) */
31 #define c5 0.78569495838710234903 /* sqrt(2)*cos(5*pi/16) */
32 #define c6 0.54119610014619712324 /* sqrt(2)*cos(6*pi/16) */
33 #define c7 0.27589937928294311353 /* sqrt(2)*cos(7*pi/16) */
35 static const float even_table[] __attribute__ ((aligned(8))) = {
42 static const float odd_table[] __attribute__ ((aligned(8))) = {
57 #define load_matrix(table) \
59 const float *t = table; \
76 __asm__ volatile("ftrv xmtrx,fv0" \
77 : "+f"(fr0),"+f"(fr1),"+f"(fr2),"+f"(fr3));
80 register float fr0 __asm__("fr0"); \
81 register float fr1 __asm__("fr1"); \
82 register float fr2 __asm__("fr2"); \
83 register float fr3 __asm__("fr3")
85 #define DESCALE(x,n) (x)*(1.0f/(1<<(n)))
87 /* this code work worse on gcc cvs. 3.2.3 work fine */
92 void idct_sh4(DCTELEM *block)
97 float tblock[8*8],*fblock;
101 fp_single_enter(fpscr);
106 load_matrix(even_table);
126 load_matrix(odd_table);
143 *--fblock = t0 - fr0;
144 *--fblock = t1 - fr1;
145 *--fblock = t2 - fr2;
146 *--fblock = t3 - fr3;
147 *--fblock = t3 + fr3;
148 *--fblock = t2 + fr2;
149 *--fblock = t1 + fr1;
150 *--fblock = t0 + fr0;
159 load_matrix(even_table);
161 ofs1 = sizeof(float)*2*8;
162 ofs2 = sizeof(float)*4*8;
163 ofs3 = sizeof(float)*6*8;
167 #define OA(fblock,ofs) *(float*)((char*)fblock + ofs)
171 fr1 = OA(fblock,ofs1);
172 fr2 = OA(fblock,ofs2);
173 fr3 = OA(fblock,ofs3);
176 OA(fblock,ofs1) = fr1;
177 OA(fblock,ofs2) = fr2;
178 OA(fblock,ofs3) = fr3;
183 load_matrix(odd_table);
188 t0 = OA(fblock, 0); /* [8*0] */
189 t1 = OA(fblock,ofs1); /* [8*2] */
190 t2 = OA(fblock,ofs2); /* [8*4] */
191 t3 = OA(fblock,ofs3); /* [8*6] */
193 fr0 = OA(fblock, 0); /* [8*1] */
194 fr1 = OA(fblock,ofs1); /* [8*3] */
195 fr2 = OA(fblock,ofs2); /* [8*5] */
196 fr3 = OA(fblock,ofs3); /* [8*7] */
199 block[8*0] = DESCALE(t0 + fr0,3);
200 block[8*7] = DESCALE(t0 - fr0,3);
201 block[8*1] = DESCALE(t1 + fr1,3);
202 block[8*6] = DESCALE(t1 - fr1,3);
203 block[8*2] = DESCALE(t2 + fr2,3);
204 block[8*5] = DESCALE(t2 - fr2,3);
205 block[8*3] = DESCALE(t3 + fr3,3);
206 block[8*4] = DESCALE(t3 - fr3,3);
210 fp_single_leave(fpscr);