1/*
2 * Copyright (c) 2003-2010 Michael Niedermayer <michaelni@gmx.at>
3 *
4 * This file is part of FFmpeg.
5 *
6 * FFmpeg is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Lesser General Public
8 * License as published by the Free Software Foundation; either
9 * version 2.1 of the License, or (at your option) any later version.
10 *
11 * FFmpeg is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
14 * Lesser General Public License for more details.
15 *
16 * You should have received a copy of the GNU Lesser General Public
17 * License along with FFmpeg; if not, write to the Free Software
18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19 */
20
21/**
22 * @file
23 * H.264 DSP functions.
24 * @author Michael Niedermayer <michaelni@gmx.at>
25 */
26
27#ifndef AVCODEC_H264DSP_H
28#define AVCODEC_H264DSP_H
29
30#include <stdint.h>
31
32typedef void (*h264_weight_func)(uint8_t *block, int stride, int height,
33                                 int log2_denom, int weight, int offset);
34typedef void (*h264_biweight_func)(uint8_t *dst, uint8_t *src,
35                                   int stride, int height, int log2_denom,
36                                   int weightd, int weights, int offset);
37
38/**
39 * Context for storing H.264 DSP functions
40 */
41typedef struct H264DSPContext {
42    /* weighted MC */
43    h264_weight_func weight_h264_pixels_tab[4];
44    h264_biweight_func biweight_h264_pixels_tab[4];
45
46    /* loop filter */
47    void (*h264_v_loop_filter_luma)(uint8_t *pix /*align 16*/, int stride,
48                                    int alpha, int beta, int8_t *tc0);
49    void (*h264_h_loop_filter_luma)(uint8_t *pix /*align 4 */, int stride,
50                                    int alpha, int beta, int8_t *tc0);
51    void (*h264_h_loop_filter_luma_mbaff)(uint8_t *pix /*align 16*/, int stride,
52                                          int alpha, int beta, int8_t *tc0);
53    /* v/h_loop_filter_luma_intra: align 16 */
54    void (*h264_v_loop_filter_luma_intra)(uint8_t *pix, int stride,
55                                          int alpha, int beta);
56    void (*h264_h_loop_filter_luma_intra)(uint8_t *pix, int stride,
57                                          int alpha, int beta);
58    void (*h264_h_loop_filter_luma_mbaff_intra)(uint8_t *pix /*align 16*/,
59                                                int stride, int alpha, int beta);
60    void (*h264_v_loop_filter_chroma)(uint8_t *pix /*align 8*/, int stride,
61                                      int alpha, int beta, int8_t *tc0);
62    void (*h264_h_loop_filter_chroma)(uint8_t *pix /*align 4*/, int stride,
63                                      int alpha, int beta, int8_t *tc0);
64    void (*h264_h_loop_filter_chroma_mbaff)(uint8_t *pix /*align 8*/,
65                                            int stride, int alpha, int beta,
66                                            int8_t *tc0);
67    void (*h264_v_loop_filter_chroma_intra)(uint8_t *pix /*align 8*/,
68                                            int stride, int alpha, int beta);
69    void (*h264_h_loop_filter_chroma_intra)(uint8_t *pix /*align 8*/,
70                                            int stride, int alpha, int beta);
71    void (*h264_h_loop_filter_chroma_mbaff_intra)(uint8_t *pix /*align 8*/,
72                                                  int stride, int alpha, int beta);
73    // h264_loop_filter_strength: simd only. the C version is inlined in h264.c
74    void (*h264_loop_filter_strength)(int16_t bS[2][4][4], uint8_t nnz[40],
75                                      int8_t ref[2][40], int16_t mv[2][40][2],
76                                      int bidir, int edges, int step,
77                                      int mask_mv0, int mask_mv1, int field);
78
79    /* IDCT */
80    void (*h264_idct_add)(uint8_t *dst /*align 4*/,
81                          int16_t *block /*align 16*/, int stride);
82    void (*h264_idct8_add)(uint8_t *dst /*align 8*/,
83                           int16_t *block /*align 16*/, int stride);
84    void (*h264_idct_dc_add)(uint8_t *dst /*align 4*/,
85                             int16_t *block /*align 16*/, int stride);
86    void (*h264_idct8_dc_add)(uint8_t *dst /*align 8*/,
87                              int16_t *block /*align 16*/, int stride);
88
89    void (*h264_idct_add16)(uint8_t *dst /*align 16*/, const int *blockoffset,
90                            int16_t *block /*align 16*/, int stride,
91                            const uint8_t nnzc[15 * 8]);
92    void (*h264_idct8_add4)(uint8_t *dst /*align 16*/, const int *blockoffset,
93                            int16_t *block /*align 16*/, int stride,
94                            const uint8_t nnzc[15 * 8]);
95    void (*h264_idct_add8)(uint8_t **dst /*align 16*/, const int *blockoffset,
96                           int16_t *block /*align 16*/, int stride,
97                           const uint8_t nnzc[15 * 8]);
98    void (*h264_idct_add16intra)(uint8_t *dst /*align 16*/, const int *blockoffset,
99                                 int16_t *block /*align 16*/,
100                                 int stride, const uint8_t nnzc[15 * 8]);
101    void (*h264_luma_dc_dequant_idct)(int16_t *output,
102                                      int16_t *input /*align 16*/, int qmul);
103    void (*h264_chroma_dc_dequant_idct)(int16_t *block, int qmul);
104
105    /* bypass-transform */
106    void (*h264_add_pixels8_clear)(uint8_t *dst, int16_t *block, int stride);
107    void (*h264_add_pixels4_clear)(uint8_t *dst, int16_t *block, int stride);
108
109    /**
110     * Search buf from the start for up to size bytes. Return the index
111     * of a zero byte, or >= size if not found. Ideally, use lookahead
112     * to filter out any zero bytes that are known to not be followed by
113     * one or more further zero bytes and a one byte. Better still, filter
114     * out any bytes that form the trailing_zero_8bits syntax element too.
115     */
116    int (*h264_find_start_code_candidate)(const uint8_t *buf, int size);
117} H264DSPContext;
118
119void ff_h264dsp_init(H264DSPContext *c, const int bit_depth,
120                     const int chroma_format_idc);
121void ff_h264dsp_init_aarch64(H264DSPContext *c, const int bit_depth,
122                             const int chroma_format_idc);
123void ff_h264dsp_init_arm(H264DSPContext *c, const int bit_depth,
124                         const int chroma_format_idc);
125void ff_h264dsp_init_ppc(H264DSPContext *c, const int bit_depth,
126                         const int chroma_format_idc);
127void ff_h264dsp_init_x86(H264DSPContext *c, const int bit_depth,
128                         const int chroma_format_idc);
129
130#endif /* AVCODEC_H264DSP_H */
131