yading@10: /*
yading@10:  * This file is part of FFmpeg.
yading@10:  *
yading@10:  * FFmpeg is free software; you can redistribute it and/or
yading@10:  * modify it under the terms of the GNU Lesser General Public
yading@10:  * License as published by the Free Software Foundation; either
yading@10:  * version 2.1 of the License, or (at your option) any later version.
yading@10:  *
yading@10:  * FFmpeg is distributed in the hope that it will be useful,
yading@10:  * but WITHOUT ANY WARRANTY; without even the implied warranty of
yading@10:  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
yading@10:  * Lesser General Public License for more details.
yading@10:  *
yading@10:  * You should have received a copy of the GNU Lesser General Public
yading@10:  * License along with FFmpeg; if not, write to the Free Software
yading@10:  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
yading@10:  */
yading@10: 
yading@10: #include "libavutil/attributes.h"
yading@10: #include "libavutil/common.h"
yading@10: #include "avcodec.h"
yading@10: #include "dsputil.h"
yading@10: #include "wmv2dsp.h"
yading@10: 
yading@10: #define W0 2048
yading@10: #define W1 2841 /* 2048*sqrt (2)*cos (1*pi/16) */
yading@10: #define W2 2676 /* 2048*sqrt (2)*cos (2*pi/16) */
yading@10: #define W3 2408 /* 2048*sqrt (2)*cos (3*pi/16) */
yading@10: #define W4 2048 /* 2048*sqrt (2)*cos (4*pi/16) */
yading@10: #define W5 1609 /* 2048*sqrt (2)*cos (5*pi/16) */
yading@10: #define W6 1108 /* 2048*sqrt (2)*cos (6*pi/16) */
yading@10: #define W7 565  /* 2048*sqrt (2)*cos (7*pi/16) */
yading@10: 
yading@10: static void wmv2_idct_row(short * b)
yading@10: {
yading@10:     int s1, s2;
yading@10:     int a0, a1, a2, a3, a4, a5, a6, a7;
yading@10: 
yading@10:     /* step 1 */
yading@10:     a1 = W1 * b[1] + W7 * b[7];
yading@10:     a7 = W7 * b[1] - W1 * b[7];
yading@10:     a5 = W5 * b[5] + W3 * b[3];
yading@10:     a3 = W3 * b[5] - W5 * b[3];
yading@10:     a2 = W2 * b[2] + W6 * b[6];
yading@10:     a6 = W6 * b[2] - W2 * b[6];
yading@10:     a0 = W0 * b[0] + W0 * b[4];
yading@10:     a4 = W0 * b[0] - W0 * b[4];
yading@10: 
yading@10:     /* step 2 */
yading@10:     s1 = (181 * (a1 - a5 + a7 - a3) + 128) >> 8; // 1, 3, 5, 7
yading@10:     s2 = (181 * (a1 - a5 - a7 + a3) + 128) >> 8;
yading@10: 
yading@10:     /* step 3 */
yading@10:     b[0] = (a0 + a2 + a1 + a5 + (1 << 7)) >> 8;
yading@10:     b[1] = (a4 + a6 + s1      + (1 << 7)) >> 8;
yading@10:     b[2] = (a4 - a6 + s2      + (1 << 7)) >> 8;
yading@10:     b[3] = (a0 - a2 + a7 + a3 + (1 << 7)) >> 8;
yading@10:     b[4] = (a0 - a2 - a7 - a3 + (1 << 7)) >> 8;
yading@10:     b[5] = (a4 - a6 - s2      + (1 << 7)) >> 8;
yading@10:     b[6] = (a4 + a6 - s1      + (1 << 7)) >> 8;
yading@10:     b[7] = (a0 + a2 - a1 - a5 + (1 << 7)) >> 8;
yading@10: }
yading@10: 
yading@10: static void wmv2_idct_col(short * b)
yading@10: {
yading@10:     int s1, s2;
yading@10:     int a0, a1, a2, a3, a4, a5, a6, a7;
yading@10: 
yading@10:     /* step 1, with extended precision */
yading@10:     a1 = (W1 * b[8 * 1] + W7 * b[8 * 7] + 4) >> 3;
yading@10:     a7 = (W7 * b[8 * 1] - W1 * b[8 * 7] + 4) >> 3;
yading@10:     a5 = (W5 * b[8 * 5] + W3 * b[8 * 3] + 4) >> 3;
yading@10:     a3 = (W3 * b[8 * 5] - W5 * b[8 * 3] + 4) >> 3;
yading@10:     a2 = (W2 * b[8 * 2] + W6 * b[8 * 6] + 4) >> 3;
yading@10:     a6 = (W6 * b[8 * 2] - W2 * b[8 * 6] + 4) >> 3;
yading@10:     a0 = (W0 * b[8 * 0] + W0 * b[8 * 4]    ) >> 3;
yading@10:     a4 = (W0 * b[8 * 0] - W0 * b[8 * 4]    ) >> 3;
yading@10: 
yading@10:     /* step 2 */
yading@10:     s1 = (181 * (a1 - a5 + a7 - a3) + 128) >> 8;
yading@10:     s2 = (181 * (a1 - a5 - a7 + a3) + 128) >> 8;
yading@10: 
yading@10:     /* step 3 */
yading@10:     b[8 * 0] = (a0 + a2 + a1 + a5 + (1 << 13)) >> 14;
yading@10:     b[8 * 1] = (a4 + a6 + s1      + (1 << 13)) >> 14;
yading@10:     b[8 * 2] = (a4 - a6 + s2      + (1 << 13)) >> 14;
yading@10:     b[8 * 3] = (a0 - a2 + a7 + a3 + (1 << 13)) >> 14;
yading@10: 
yading@10:     b[8 * 4] = (a0 - a2 - a7 - a3 + (1 << 13)) >> 14;
yading@10:     b[8 * 5] = (a4 - a6 - s2      + (1 << 13)) >> 14;
yading@10:     b[8 * 6] = (a4 + a6 - s1      + (1 << 13)) >> 14;
yading@10:     b[8 * 7] = (a0 + a2 - a1 - a5 + (1 << 13)) >> 14;
yading@10: }
yading@10: 
yading@10: static void wmv2_idct_add_c(uint8_t *dest, int line_size, int16_t *block)
yading@10: {
yading@10:     int i;
yading@10: 
yading@10:     for (i = 0; i < 64; i += 8)
yading@10:         wmv2_idct_row(block + i);
yading@10:     for (i = 0; i < 8; i++)
yading@10:         wmv2_idct_col(block + i);
yading@10: 
yading@10:     for (i = 0; i < 8; i++) {
yading@10:         dest[0] = av_clip_uint8(dest[0] + block[0]);
yading@10:         dest[1] = av_clip_uint8(dest[1] + block[1]);
yading@10:         dest[2] = av_clip_uint8(dest[2] + block[2]);
yading@10:         dest[3] = av_clip_uint8(dest[3] + block[3]);
yading@10:         dest[4] = av_clip_uint8(dest[4] + block[4]);
yading@10:         dest[5] = av_clip_uint8(dest[5] + block[5]);
yading@10:         dest[6] = av_clip_uint8(dest[6] + block[6]);
yading@10:         dest[7] = av_clip_uint8(dest[7] + block[7]);
yading@10:         dest += line_size;
yading@10:         block += 8;
yading@10:     }
yading@10: }
yading@10: 
yading@10: static void wmv2_idct_put_c(uint8_t *dest, int line_size, int16_t *block)
yading@10: {
yading@10:     int i;
yading@10: 
yading@10:     for (i = 0; i < 64; i += 8)
yading@10:         wmv2_idct_row(block + i);
yading@10:     for (i = 0; i < 8; i++)
yading@10:         wmv2_idct_col(block + i);
yading@10: 
yading@10:     for (i = 0; i < 8; i++) {
yading@10:         dest[0] = av_clip_uint8(block[0]);
yading@10:         dest[1] = av_clip_uint8(block[1]);
yading@10:         dest[2] = av_clip_uint8(block[2]);
yading@10:         dest[3] = av_clip_uint8(block[3]);
yading@10:         dest[4] = av_clip_uint8(block[4]);
yading@10:         dest[5] = av_clip_uint8(block[5]);
yading@10:         dest[6] = av_clip_uint8(block[6]);
yading@10:         dest[7] = av_clip_uint8(block[7]);
yading@10:         dest += line_size;
yading@10:         block += 8;
yading@10:     }
yading@10: }
yading@10: 
yading@10: av_cold void ff_wmv2dsp_init(WMV2DSPContext *c)
yading@10: {
yading@10:     c->idct_add  = wmv2_idct_add_c;
yading@10:     c->idct_put  = wmv2_idct_put_c;
yading@10:     c->idct_perm = FF_NO_IDCT_PERM;
yading@10: }