1 /*****************************************************************************
2 * pixel.c: pixel metrics
3 *****************************************************************************
4 * Copyright (C) 2004-2016 x264 project
6 * Authors: Loren Merritt <lorenm@u.washington.edu>
7 * Fiona Glaser <fiona@x264.com>
8 Henrik Gramner <henrik@gramner.com>
10 * This program is free software; you can redistribute it and/or modify
11 * it under the terms of the GNU General Public License as published by
12 * the Free Software Foundation; either version 2 of the License, or
13 * (at your option) any later version.
15 * This program is distributed in the hope that it will be useful,
16 * but WITHOUT ANY WARRANTY; without even the implied warranty of
17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 * GNU General Public License for more details.
20 * You should have received a copy of the GNU General Public License
21 * along with this program; if not, write to the Free Software
22 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111, USA.
24 * This program is also available under a commercial proprietary license.
25 * For more information, contact us at licensing@x264.com.
26 *****************************************************************************/
31 // SSD assumes all args aligned
32 // other cmp functions assume first arg aligned
33 typedef int (*x264_pixel_cmp_t) ( pixel *, intptr_t, pixel *, intptr_t );
34 typedef void (*x264_pixel_cmp_x3_t) ( pixel *, pixel *, pixel *, pixel *, intptr_t, int[3] );
35 typedef void (*x264_pixel_cmp_x4_t) ( pixel *, pixel *, pixel *, pixel *, pixel *, intptr_t, int[4] );
47 /* Subsampled chroma only */
48 PIXEL_4x16 = 7, /* 4:2:2 */
50 PIXEL_2x8 = 9, /* 4:2:2 */
55 static const struct { uint8_t w, h; } x264_pixel_size[12] =
57 { 16, 16 }, { 16, 8 }, { 8, 16 }, { 8, 8 }, { 8, 4 }, { 4, 8 }, { 4, 4 },
58 { 4, 16 }, { 4, 2 }, { 2, 8 }, { 2, 4 }, { 2, 2 },
61 static const uint8_t x264_size2pixel[5][5] =
64 { 0, PIXEL_4x4, PIXEL_8x4, 0, 0 },
65 { 0, PIXEL_4x8, PIXEL_8x8, 0, PIXEL_16x8 },
67 { 0, 0, PIXEL_8x16, 0, PIXEL_16x16 }
70 static const uint8_t x264_luma2chroma_pixel[4][7] =
73 { PIXEL_8x8, PIXEL_8x4, PIXEL_4x8, PIXEL_4x4, PIXEL_4x2, PIXEL_2x4, PIXEL_2x2 }, /* 4:2:0 */
74 { PIXEL_8x16, PIXEL_8x8, PIXEL_4x16, PIXEL_4x8, PIXEL_4x4, PIXEL_2x8, PIXEL_2x4 }, /* 4:2:2 */
75 { PIXEL_16x16, PIXEL_16x8, PIXEL_8x16, PIXEL_8x8, PIXEL_8x4, PIXEL_4x8, PIXEL_4x4 }, /* 4:4:4 */
80 x264_pixel_cmp_t sad[8];
81 x264_pixel_cmp_t ssd[8];
82 x264_pixel_cmp_t satd[8];
83 x264_pixel_cmp_t ssim[7];
84 x264_pixel_cmp_t sa8d[4];
85 x264_pixel_cmp_t mbcmp[8]; /* either satd or sad for subpel refine and mode decision */
86 x264_pixel_cmp_t mbcmp_unaligned[8]; /* unaligned mbcmp for subpel */
87 x264_pixel_cmp_t fpelcmp[8]; /* either satd or sad for fullpel motion search */
88 x264_pixel_cmp_x3_t fpelcmp_x3[7];
89 x264_pixel_cmp_x4_t fpelcmp_x4[7];
90 x264_pixel_cmp_t sad_aligned[8]; /* Aligned SAD for mbcmp */
91 int (*vsad)( pixel *, intptr_t, int );
92 int (*asd8)( pixel *pix1, intptr_t stride1, pixel *pix2, intptr_t stride2, int height );
93 uint64_t (*sa8d_satd[1])( pixel *pix1, intptr_t stride1, pixel *pix2, intptr_t stride2 );
95 uint64_t (*var[4])( pixel *pix, intptr_t stride );
96 int (*var2[4])( pixel *pix1, intptr_t stride1,
97 pixel *pix2, intptr_t stride2, int *ssd );
98 uint64_t (*hadamard_ac[4])( pixel *pix, intptr_t stride );
100 void (*ssd_nv12_core)( pixel *pixuv1, intptr_t stride1,
101 pixel *pixuv2, intptr_t stride2, int width, int height,
102 uint64_t *ssd_u, uint64_t *ssd_v );
103 void (*ssim_4x4x2_core)( const pixel *pix1, intptr_t stride1,
104 const pixel *pix2, intptr_t stride2, int sums[2][4] );
105 float (*ssim_end4)( int sum0[5][4], int sum1[5][4], int width );
107 /* multiple parallel calls to cmp. */
108 x264_pixel_cmp_x3_t sad_x3[7];
109 x264_pixel_cmp_x4_t sad_x4[7];
110 x264_pixel_cmp_x3_t satd_x3[7];
111 x264_pixel_cmp_x4_t satd_x4[7];
113 /* abs-diff-sum for successive elimination.
114 * may round width up to a multiple of 16. */
115 int (*ads[7])( int enc_dc[4], uint16_t *sums, int delta,
116 uint16_t *cost_mvx, int16_t *mvs, int width, int thresh );
118 /* calculate satd or sad of V, H, and DC modes. */
119 void (*intra_mbcmp_x3_16x16)( pixel *fenc, pixel *fdec, int res[3] );
120 void (*intra_satd_x3_16x16) ( pixel *fenc, pixel *fdec, int res[3] );
121 void (*intra_sad_x3_16x16) ( pixel *fenc, pixel *fdec, int res[3] );
122 void (*intra_mbcmp_x3_4x4) ( pixel *fenc, pixel *fdec, int res[3] );
123 void (*intra_satd_x3_4x4) ( pixel *fenc, pixel *fdec, int res[3] );
124 void (*intra_sad_x3_4x4) ( pixel *fenc, pixel *fdec, int res[3] );
125 void (*intra_mbcmp_x3_chroma)( pixel *fenc, pixel *fdec, int res[3] );
126 void (*intra_satd_x3_chroma) ( pixel *fenc, pixel *fdec, int res[3] );
127 void (*intra_sad_x3_chroma) ( pixel *fenc, pixel *fdec, int res[3] );
128 void (*intra_mbcmp_x3_8x16c) ( pixel *fenc, pixel *fdec, int res[3] );
129 void (*intra_satd_x3_8x16c) ( pixel *fenc, pixel *fdec, int res[3] );
130 void (*intra_sad_x3_8x16c) ( pixel *fenc, pixel *fdec, int res[3] );
131 void (*intra_mbcmp_x3_8x8c) ( pixel *fenc, pixel *fdec, int res[3] );
132 void (*intra_satd_x3_8x8c) ( pixel *fenc, pixel *fdec, int res[3] );
133 void (*intra_sad_x3_8x8c) ( pixel *fenc, pixel *fdec, int res[3] );
134 void (*intra_mbcmp_x3_8x8) ( pixel *fenc, pixel edge[36], int res[3] );
135 void (*intra_sa8d_x3_8x8) ( pixel *fenc, pixel edge[36], int res[3] );
136 void (*intra_sad_x3_8x8) ( pixel *fenc, pixel edge[36], int res[3] );
137 /* find minimum satd or sad of all modes, and set fdec.
138 * may be NULL, in which case just use pred+satd instead. */
139 int (*intra_mbcmp_x9_4x4)( pixel *fenc, pixel *fdec, uint16_t *bitcosts );
140 int (*intra_satd_x9_4x4) ( pixel *fenc, pixel *fdec, uint16_t *bitcosts );
141 int (*intra_sad_x9_4x4) ( pixel *fenc, pixel *fdec, uint16_t *bitcosts );
142 int (*intra_mbcmp_x9_8x8)( pixel *fenc, pixel *fdec, pixel edge[36], uint16_t *bitcosts, uint16_t *satds );
143 int (*intra_sa8d_x9_8x8) ( pixel *fenc, pixel *fdec, pixel edge[36], uint16_t *bitcosts, uint16_t *satds );
144 int (*intra_sad_x9_8x8) ( pixel *fenc, pixel *fdec, pixel edge[36], uint16_t *bitcosts, uint16_t *satds );
145 } x264_pixel_function_t;
147 void x264_pixel_init( int cpu, x264_pixel_function_t *pixf );
148 void x264_pixel_ssd_nv12 ( x264_pixel_function_t *pf, pixel *pix1, intptr_t i_pix1, pixel *pix2, intptr_t i_pix2,
149 int i_width, int i_height, uint64_t *ssd_u, uint64_t *ssd_v );
150 uint64_t x264_pixel_ssd_wxh( x264_pixel_function_t *pf, pixel *pix1, intptr_t i_pix1, pixel *pix2, intptr_t i_pix2,
151 int i_width, int i_height );
152 float x264_pixel_ssim_wxh ( x264_pixel_function_t *pf, pixel *pix1, intptr_t i_pix1, pixel *pix2, intptr_t i_pix2,
153 int i_width, int i_height, void *buf, int *cnt );
154 int x264_field_vsad( x264_t *h, int mb_x, int mb_y );