85373336c87392bfc8111a6288efc1677ce8423f
[libav.git] / libavcodec / dsputil.h
1 /*
2 * DSP utils
3 * Copyright (c) 2000, 2001, 2002 Fabrice Bellard.
4 *
5 * This library is free software; you can redistribute it and/or
6 * modify it under the terms of the GNU Lesser General Public
7 * License as published by the Free Software Foundation; either
8 * version 2 of the License, or (at your option) any later version.
9 *
10 * This library is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 * Lesser General Public License for more details.
14 *
15 * You should have received a copy of the GNU Lesser General Public
16 * License along with this library; if not, write to the Free Software
17 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
18 */
19 #ifndef DSPUTIL_H
20 #define DSPUTIL_H
21
22 #include "common.h"
23 #include "avcodec.h"
24
25 //#define DEBUG
26 /* dct code */
27 typedef short DCTELEM;
28
29 void fdct_ifast (DCTELEM *data);
30
31 void j_rev_dct (DCTELEM *data);
32
33 void fdct_mmx(DCTELEM *block);
34
35 extern void (*av_fdct)(DCTELEM *block);
36
37 /* encoding scans */
38 extern UINT8 ff_alternate_horizontal_scan[64];
39 extern UINT8 ff_alternate_vertical_scan[64];
40 extern UINT8 zigzag_direct[64];
41
42 /* permutation table */
43 extern UINT8 permutation[64];
44
45 /* pixel operations */
46 #define MAX_NEG_CROP 384
47
48 /* temporary */
49 extern UINT32 squareTbl[512];
50 extern UINT8 cropTbl[256 + 2 * MAX_NEG_CROP];
51
52 void dsputil_init(void);
53
54 /* pixel ops : interface with DCT */
55
56 extern void (*ff_idct)(DCTELEM *block);
57 extern void (*ff_idct_put)(UINT8 *dest, int line_size, DCTELEM *block);
58 extern void (*ff_idct_add)(UINT8 *dest, int line_size, DCTELEM *block);
59 extern void (*get_pixels)(DCTELEM *block, const UINT8 *pixels, int line_size);
60 extern void (*diff_pixels)(DCTELEM *block, const UINT8 *s1, const UINT8 *s2, int stride);
61 extern void (*put_pixels_clamped)(const DCTELEM *block, UINT8 *pixels, int line_size);
62 extern void (*add_pixels_clamped)(const DCTELEM *block, UINT8 *pixels, int line_size);
63 extern void (*gmc1)(UINT8 *dst, UINT8 *src, int srcStride, int h, int x16, int y16, int rounder);
64 extern void (*clear_blocks)(DCTELEM *blocks);
65
66
67 void get_pixels_c(DCTELEM *block, const UINT8 *pixels, int line_size);
68 void diff_pixels_c(DCTELEM *block, const UINT8 *s1, const UINT8 *s2, int stride);
69 void put_pixels_clamped_c(const DCTELEM *block, UINT8 *pixels, int line_size);
70 void add_pixels_clamped_c(const DCTELEM *block, UINT8 *pixels, int line_size);
71 void clear_blocks_c(DCTELEM *blocks);
72
73 /* add and put pixel (decoding) */
74 typedef void (*op_pixels_func)(UINT8 *block, const UINT8 *pixels, int line_size, int h);
75 typedef void (*qpel_mc_func)(UINT8 *dst, UINT8 *src, int dstStride, int srcStride, int mx, int my);
76
77 extern op_pixels_func put_pixels_tab[4];
78 extern op_pixels_func avg_pixels_tab[4];
79 extern op_pixels_func put_no_rnd_pixels_tab[4];
80 extern op_pixels_func avg_no_rnd_pixels_tab[4];
81 extern qpel_mc_func qpel_mc_rnd_tab[16];
82 extern qpel_mc_func qpel_mc_no_rnd_tab[16];
83
84 /* motion estimation */
85
86 typedef int (*op_pixels_abs_func)(UINT8 *blk1, UINT8 *blk2, int line_size);
87
88 extern op_pixels_abs_func pix_abs16x16;
89 extern op_pixels_abs_func pix_abs16x16_x2;
90 extern op_pixels_abs_func pix_abs16x16_y2;
91 extern op_pixels_abs_func pix_abs16x16_xy2;
92 extern op_pixels_abs_func pix_abs8x8;
93 extern op_pixels_abs_func pix_abs8x8_x2;
94 extern op_pixels_abs_func pix_abs8x8_y2;
95 extern op_pixels_abs_func pix_abs8x8_xy2;
96
97 int pix_abs16x16_c(UINT8 *blk1, UINT8 *blk2, int lx);
98 int pix_abs16x16_x2_c(UINT8 *blk1, UINT8 *blk2, int lx);
99 int pix_abs16x16_y2_c(UINT8 *blk1, UINT8 *blk2, int lx);
100 int pix_abs16x16_xy2_c(UINT8 *blk1, UINT8 *blk2, int lx);
101
102 static inline int block_permute_op(int j)
103 {
104 return permutation[j];
105 }
106
107 void block_permute(INT16 *block);
108
109 #if defined(HAVE_MMX)
110
111 #define MM_MMX 0x0001 /* standard MMX */
112 #define MM_3DNOW 0x0004 /* AMD 3DNOW */
113 #define MM_MMXEXT 0x0002 /* SSE integer functions or AMD MMX ext */
114 #define MM_SSE 0x0008 /* SSE functions */
115 #define MM_SSE2 0x0010 /* PIV SSE2 functions */
116
117 extern int mm_flags;
118
119 int mm_support(void);
120
121 static inline void emms(void)
122 {
123 __asm __volatile ("emms;":::"memory");
124 }
125
126 #define emms_c() \
127 {\
128 if (mm_flags & MM_MMX)\
129 emms();\
130 }
131
132 #define __align8 __attribute__ ((aligned (8)))
133
134 void dsputil_init_mmx(void);
135 void dsputil_set_bit_exact_mmx(void);
136
137 #elif defined(ARCH_ARMV4L)
138
139 #define emms_c()
140
141 /* This is to use 4 bytes read to the IDCT pointers for some 'zero'
142 line ptimizations */
143 #define __align8 __attribute__ ((aligned (4)))
144
145 void dsputil_init_armv4l(void);
146
147 #elif defined(HAVE_MLIB)
148
149 #define emms_c()
150
151 /* SPARC/VIS IDCT needs 8-byte aligned DCT blocks */
152 #define __align8 __attribute__ ((aligned (8)))
153
154 void dsputil_init_mlib(void);
155
156 #elif defined(ARCH_ALPHA)
157
158 #define emms_c()
159 #define __align8 __attribute__ ((aligned (8)))
160
161 void dsputil_init_alpha(void);
162
163 #else
164
165 #define emms_c()
166
167 #define __align8
168
169 #endif
170
171 /* PSNR */
172 void get_psnr(UINT8 *orig_image[3], UINT8 *coded_image[3],
173 int orig_linesize[3], int coded_linesize,
174 AVCodecContext *avctx);
175
176 #endif