1 /*
2 * Copyright (c) 2020, Alliance for Open Media. All rights reserved.
3 *
4 * This source code is subject to the terms of the BSD 2 Clause License and
5 * the Alliance for Open Media Patent License 1.0. If the BSD 2 Clause License
6 * was not distributed with this source code in the LICENSE file, you can
7 * obtain it at www.aomedia.org/license/software. If the Alliance for Open
8 * Media Patent License 1.0 was not distributed with this source code in the
9 * PATENTS file, you can obtain it at www.aomedia.org/license/patent.
10 */
11
12 #ifndef AOM_AV1_ENCODER_TRANSFORM_SEARCH_H_
13 #define AOM_AV1_ENCODER_TRANSFORM_SEARCH_H_
14
15 #include "av1/common/pred_common.h"
16 #include "av1/encoder/encoder.h"
17
18 #ifdef __cplusplus
19 extern "C" {
20 #endif
21
22 // Set this macro as 1 to collect data about tx size selection.
23 #define COLLECT_TX_SIZE_DATA 0
24
25 #if COLLECT_TX_SIZE_DATA
26 static const char av1_tx_size_data_output_file[] = "tx_size_data.txt";
27 #endif
28
29 enum {
30 FTXS_NONE = 0,
31 FTXS_DCT_AND_1D_DCT_ONLY = 1 << 0,
32 FTXS_DISABLE_TRELLIS_OPT = 1 << 1,
33 FTXS_USE_TRANSFORM_DOMAIN = 1 << 2
34 } UENUM1BYTE(FAST_TX_SEARCH_MODE);
35
tx_size_cost(const MACROBLOCK * const x,BLOCK_SIZE bsize,TX_SIZE tx_size)36 static inline int tx_size_cost(const MACROBLOCK *const x, BLOCK_SIZE bsize,
37 TX_SIZE tx_size) {
38 assert(bsize == x->e_mbd.mi[0]->bsize);
39 if (x->txfm_search_params.tx_mode_search_type != TX_MODE_SELECT ||
40 !block_signals_txsize(bsize))
41 return 0;
42
43 const int32_t tx_size_cat = bsize_to_tx_size_cat(bsize);
44 const int depth = tx_size_to_depth(tx_size, bsize);
45 const MACROBLOCKD *const xd = &x->e_mbd;
46 const int tx_size_ctx = get_tx_size_context(xd);
47 return x->mode_costs.tx_size_cost[tx_size_cat][tx_size_ctx][depth];
48 }
49
50 /*!\brief Compute the pixel domain distortion.
51 *
52 * \ingroup transform_search
53 * Compute the pixel domain distortion from diff on all visible 4x4s in the
54 * transform block.
55 *
56 * \param[in] x Pointer to structure holding the data for the
57 current encoding macroblock
58 * \param[in] plane Plane index
59 * \param[in] blk_row Block row index
60 * \param[in] blk_col Block col index
61 * \param[in] plane_bsize Current plane block size
62 * \param[in] tx_bsize Transform size
63 * \param[in] block_mse_q8 Block mse
64 * \return An int64_t value that is the block sse.
65 */
66 int64_t av1_pixel_diff_dist(const MACROBLOCK *x, int plane, int blk_row,
67 int blk_col, const BLOCK_SIZE plane_bsize,
68 const BLOCK_SIZE tx_bsize,
69 unsigned int *block_mse_q8);
70
71 int64_t av1_estimate_txfm_yrd(const AV1_COMP *const cpi, MACROBLOCK *x,
72 RD_STATS *rd_stats, int64_t ref_best_rd,
73 BLOCK_SIZE bs, TX_SIZE tx_size);
74
75 /*!\brief Recursive transform size and type search.
76 *
77 * \ingroup transform_search
78 * Search for best transform size and type for luma inter blocks. The transform
79 * block partitioning can be recursive resulting in non-uniform transform sizes.
80 * The best transform size and type, if found, will be saved in the MB_MODE_INFO
81 * structure, and the corresponding RD stats will be saved in rd_stats.
82 *
83 * \param[in] cpi Top-level encoder structure
84 * \param[in] x Pointer to structure holding the data for the
85 current encoding macroblock
86 * \param[in] rd_stats Pointer to struct to keep track of the RD stats
87 * \param[in] bsize Current macroblock size
88 * \param[in] ref_best_rd Best RD cost seen for this block so far
89 * \remark Nothing is returned. The selected transform size and type will
90 be saved in the MB_MODE_INFO structure
91 */
92 void av1_pick_recursive_tx_size_type_yrd(const AV1_COMP *cpi, MACROBLOCK *x,
93 RD_STATS *rd_stats, BLOCK_SIZE bsize,
94 int64_t ref_best_rd);
95
96 /*!\brief Uniform transform size and type search.
97 *
98 * \ingroup transform_search
99 * Search for the best transform size and type for current macroblock block,
100 * with the assumption that all the transform blocks have a uniform size
101 * (VP9 style). The selected transform size and type will be saved in the
102 * MB_MODE_INFO structure; the corresponding RD stats will be saved in rd_stats.
103 * This function may be used for both intra and inter predicted blocks.
104 *
105 * \param[in] cpi Top-level encoder structure
106 * \param[in] x Pointer to structure holding the data for the
107 current encoding macroblock
108 * \param[in] rd_stats Pointer to struct to keep track of the RD stats
109 * \param[in] bs Current macroblock size
110 * \param[in] ref_best_rd Best RD cost seen for this block so far
111 * \remark Nothing is returned. The selected transform size and type will
112 be saved in the MB_MODE_INFO structure
113 */
114 void av1_pick_uniform_tx_size_type_yrd(const AV1_COMP *const cpi, MACROBLOCK *x,
115 RD_STATS *rd_stats, BLOCK_SIZE bs,
116 int64_t ref_best_rd);
117
118 /*!\brief Chroma block transform search.
119 *
120 * \ingroup transform_search
121 * Calculate the transform coefficient RD cost for the given chroma macroblock
122 * If the current mode is intra, then this function will compute the predictor.
123 *
124 * \param[in] cpi Top-level encoder structure
125 * \param[in] x Pointer to structure holding the data for the
126 current encoding macroblock
127 * \param[in] rd_stats Pointer to struct to keep track of the RD stats
128 * \param[in] bsize Current macroblock size
129 * \param[in] ref_best_rd Best RD cost seen for this block so far
130 * \return An integer value is returned. 0: early termination triggered,
131 no valid rd cost available; 1: rd cost values are valid.
132 */
133 int av1_txfm_uvrd(const AV1_COMP *const cpi, MACROBLOCK *x, RD_STATS *rd_stats,
134 BLOCK_SIZE bsize, int64_t ref_best_rd);
135
136 /*!\brief Transform type search with fixed transform size.
137 *
138 * \ingroup transform_search
139 * Search for the best transform type and calculate the transform coefficients
140 * RD cost of the current transform block with the specified (uniform) transform
141 * size and plane. The RD results will be saved in rd_stats.
142 *
143 * \param[in] x Pointer to structure holding the data for the
144 current encoding macroblock
145 * \param[in] cpi Top-level encoder structure
146 * \param[in] rd_stats Pointer to struct to keep track of the RD stats
147 * \param[in] ref_best_rd Best RD cost seen for this block so far
148 * \param[in] current_rd Current RD cost for this block so far
149 * \param[in] plane Plane index
150 * \param[in] plane_bsize Size of the current macroblock considering
151 sup-sampling
152 * \param[in] tx_size The given transform size
153 * \param[in] ftxs_mode Transform search mode specifying desired speed
154 and quality tradeoff
155 * \param[in] skip_trellis Binary flag indicating if trellis optimization
156 should be skipped
157 *
158 * \remark Nothing is returned. The RD results will be saved in rd_stats.
159 */
160 void av1_txfm_rd_in_plane(MACROBLOCK *x, const AV1_COMP *cpi,
161 RD_STATS *rd_stats, int64_t ref_best_rd,
162 int64_t current_rd, int plane, BLOCK_SIZE plane_bsize,
163 TX_SIZE tx_size, FAST_TX_SEARCH_MODE ftxs_mode,
164 int skip_trellis);
165
166 /*!\brief Recursive transform size and type search.
167 *
168 * \ingroup transform_search
169 * This function combines y and uv planes' transform search processes together
170 * for inter-predicted blocks (including IntraBC), when the prediction is
171 * already generated. It first does subtraction to obtain the prediction error.
172 * Then it calls
173 * av1_pick_recursive_tx_size_type_yrd/av1_pick_uniform_tx_size_type_yrd and
174 * av1_txfm_uvrd sequentially and handles possible early terminations.
175 * The RD metrics are calculated and stored in rd_stats/_y/_uv.
176 *
177 * \param[in] cpi Top-level encoder structure
178 * \param[in] x Pointer to structure holding the data for the
179 current encoding macroblock
180 * \param[in] bsize Current macroblock size
181 * \param[in] rd_stats Pointer to struct to keep track of the overal RD
182 stats
183 * \param[in] rd_stats_y Pointer to struct to keep track of the RD
184 stats for the luma plane
185 * \param[in] rd_stats_uv Pointer to struct to keep track of the RD
186 stats for the chroma planes
187 * \param[in] mode_rate Rate cost to encode the prediction mode info. of
188 the current macroblock
189 * \param[in] ref_best_rd Best RD cost seen for this block so far
190 *
191 * \return An integer value is returned indicating if a valid transform
192 candidate is found (1) or not (0).
193 */
194 int av1_txfm_search(const AV1_COMP *cpi, MACROBLOCK *x, BLOCK_SIZE bsize,
195 RD_STATS *rd_stats, RD_STATS *rd_stats_y,
196 RD_STATS *rd_stats_uv, int mode_rate, int64_t ref_best_rd);
197
198 #ifdef __cplusplus
199 } // extern "C"
200 #endif
201
202 #endif // AOM_AV1_ENCODER_TRANSFORM_SEARCH_H_
203