2 * This file is part of FFmpeg.
4 * FFmpeg is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU Lesser General Public
6 * License as published by the Free Software Foundation; either
7 * version 2.1 of the License, or (at your option) any later version.
9 * FFmpeg is distributed in the hope that it will be useful,
10 * but WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 * Lesser General Public License for more details.
14 * You should have received a copy of the GNU Lesser General Public
15 * License along with FFmpeg; if not, write to the Free Software
16 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
24 #include "libavutil/avassert.h"
25 #include "libavutil/attributes.h"
26 #include "libavutil/imgutils.h"
29 #include "mpegvideoencdsp.h"
31 static int try_8x8basis_c(int16_t rem[64], int16_t weight[64],
32 int16_t basis[64], int scale)
37 for (i = 0; i < 8 * 8; i++) {
38 int b = rem[i] + ((basis[i] * scale +
39 (1 << (BASIS_SHIFT - RECON_SHIFT - 1))) >>
40 (BASIS_SHIFT - RECON_SHIFT));
43 av_assert2(-512 < b && b < 512);
45 sum += (w * b) * (w * b) >> 4;
50 static void add_8x8basis_c(int16_t rem[64], int16_t basis[64], int scale)
54 for (i = 0; i < 8 * 8; i++)
55 rem[i] += (basis[i] * scale +
56 (1 << (BASIS_SHIFT - RECON_SHIFT - 1))) >>
57 (BASIS_SHIFT - RECON_SHIFT);
60 static int pix_sum_c(uint8_t *pix, int line_size)
64 for (i = 0; i < 16; i++) {
65 for (j = 0; j < 16; j += 8) {
76 pix += line_size - 16;
81 static int pix_norm1_c(uint8_t *pix, int line_size)
84 const uint32_t *sq = ff_square_tab + 256;
86 for (i = 0; i < 16; i++) {
87 for (j = 0; j < 16; j += 8) {
89 register uint64_t x = *(uint64_t *) pix;
91 s += sq[(x >> 8) & 0xff];
92 s += sq[(x >> 16) & 0xff];
93 s += sq[(x >> 24) & 0xff];
94 s += sq[(x >> 32) & 0xff];
95 s += sq[(x >> 40) & 0xff];
96 s += sq[(x >> 48) & 0xff];
97 s += sq[(x >> 56) & 0xff];
99 register uint32_t x = *(uint32_t *) pix;
101 s += sq[(x >> 8) & 0xff];
102 s += sq[(x >> 16) & 0xff];
103 s += sq[(x >> 24) & 0xff];
104 x = *(uint32_t *) (pix + 4);
106 s += sq[(x >> 8) & 0xff];
107 s += sq[(x >> 16) & 0xff];
108 s += sq[(x >> 24) & 0xff];
112 pix += line_size - 16;
117 /* draw the edges of width 'w' of an image of size width, height */
118 // FIXME: Check that this is OK for MPEG-4 interlaced.
119 static void draw_edges_8_c(uint8_t *buf, int wrap, int width, int height,
120 int w, int h, int sides)
122 uint8_t *ptr = buf, *last_line;
126 for (i = 0; i < height; i++) {
127 memset(ptr - w, ptr[0], w);
128 memset(ptr + width, ptr[width - 1], w);
132 /* top and bottom + corners */
134 last_line = buf + (height - 1) * wrap;
135 if (sides & EDGE_TOP)
136 for (i = 0; i < h; i++)
138 memcpy(buf - (i + 1) * wrap, buf, width + w + w);
139 if (sides & EDGE_BOTTOM)
140 for (i = 0; i < h; i++)
142 memcpy(last_line + (i + 1) * wrap, last_line, width + w + w);
146 static void shrink22(uint8_t *dst, int dst_wrap,
147 const uint8_t *src, int src_wrap,
148 int width, int height)
151 const uint8_t *s1, *s2;
154 for (; height > 0; height--) {
158 for (w = width; w >= 4; w -= 4) {
159 d[0] = (s1[0] + s1[1] + s2[0] + s2[1] + 2) >> 2;
160 d[1] = (s1[2] + s1[3] + s2[2] + s2[3] + 2) >> 2;
161 d[2] = (s1[4] + s1[5] + s2[4] + s2[5] + 2) >> 2;
162 d[3] = (s1[6] + s1[7] + s2[6] + s2[7] + 2) >> 2;
168 d[0] = (s1[0] + s1[1] + s2[0] + s2[1] + 2) >> 2;
179 static void shrink44(uint8_t *dst, int dst_wrap,
180 const uint8_t *src, int src_wrap,
181 int width, int height)
184 const uint8_t *s1, *s2, *s3, *s4;
187 for (; height > 0; height--) {
193 for (w = width; w > 0; w--) {
194 d[0] = (s1[0] + s1[1] + s1[2] + s1[3] +
195 s2[0] + s2[1] + s2[2] + s2[3] +
196 s3[0] + s3[1] + s3[2] + s3[3] +
197 s4[0] + s4[1] + s4[2] + s4[3] + 8) >> 4;
210 static void shrink88(uint8_t *dst, int dst_wrap,
211 const uint8_t *src, int src_wrap,
212 int width, int height)
216 for (; height > 0; height--) {
217 for(w = width;w > 0; w--) {
219 for (i = 0; i < 8; i++) {
220 tmp += src[0] + src[1] + src[2] + src[3] +
221 src[4] + src[5] + src[6] + src[7];
224 *(dst++) = (tmp + 32) >> 6;
225 src += 8 - 8 * src_wrap;
227 src += 8 * src_wrap - 8 * width;
228 dst += dst_wrap - width;
232 av_cold void ff_mpegvideoencdsp_init(MpegvideoEncDSPContext *c,
233 AVCodecContext *avctx)
235 c->try_8x8basis = try_8x8basis_c;
236 c->add_8x8basis = add_8x8basis_c;
238 c->shrink[0] = av_image_copy_plane;
239 c->shrink[1] = shrink22;
240 c->shrink[2] = shrink44;
241 c->shrink[3] = shrink88;
243 c->pix_sum = pix_sum_c;
244 c->pix_norm1 = pix_norm1_c;
246 c->draw_edges = draw_edges_8_c;
249 ff_mpegvideoencdsp_init_arm(c, avctx);
251 ff_mpegvideoencdsp_init_ppc(c, avctx);
253 ff_mpegvideoencdsp_init_x86(c, avctx);
255 ff_mpegvideoencdsp_init_mips(c, avctx);