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#include <stddef.h> 32 33typedef void (*h264_weight_func)(uint8_t *block, ptrdiff_t stride, int height, 34 int log2_denom, int weight, int offset); 35typedef void (*h264_biweight_func)(uint8_t *dst, uint8_t *src, 36 ptrdiff_t stride, int height, int log2_denom, 37 int weightd, int weights, int offset); 38 39/** 40 * Context for storing H.264 DSP functions 41 */ 42typedef struct H264DSPContext { 43 /* weighted MC */ 44 h264_weight_func weight_h264_pixels_tab[4]; 45 h264_biweight_func biweight_h264_pixels_tab[4]; 46 47 /* loop filter */ 48 void (*h264_v_loop_filter_luma)(uint8_t *pix /*align 16*/, ptrdiff_t stride, 49 int alpha, int beta, int8_t *tc0); 50 void (*h264_h_loop_filter_luma)(uint8_t *pix /*align 4 */, ptrdiff_t stride, 51 int alpha, int beta, int8_t *tc0); 52 void (*h264_h_loop_filter_luma_mbaff)(uint8_t *pix /*align 16*/, ptrdiff_t stride, 53 int alpha, int beta, int8_t *tc0); 54 /* v/h_loop_filter_luma_intra: align 16 */ 55 void (*h264_v_loop_filter_luma_intra)(uint8_t *pix, ptrdiff_t stride, 56 int alpha, int beta); 57 void (*h264_h_loop_filter_luma_intra)(uint8_t *pix, ptrdiff_t stride, 58 int alpha, int beta); 59 void (*h264_h_loop_filter_luma_mbaff_intra)(uint8_t *pix /*align 16*/, 60 ptrdiff_t stride, int alpha, int beta); 61 void (*h264_v_loop_filter_chroma)(uint8_t *pix /*align 8*/, ptrdiff_t stride, 62 int alpha, int beta, int8_t *tc0); 63 void (*h264_h_loop_filter_chroma)(uint8_t *pix /*align 4*/, ptrdiff_t stride, 64 int alpha, int beta, int8_t *tc0); 65 void (*h264_h_loop_filter_chroma_mbaff)(uint8_t *pix /*align 8*/, 66 ptrdiff_t stride, int alpha, int beta, 67 int8_t *tc0); 68 void (*h264_v_loop_filter_chroma_intra)(uint8_t *pix /*align 8*/, 69 ptrdiff_t stride, int alpha, int beta); 70 void (*h264_h_loop_filter_chroma_intra)(uint8_t *pix /*align 8*/, 71 ptrdiff_t stride, int alpha, int beta); 72 void (*h264_h_loop_filter_chroma_mbaff_intra)(uint8_t *pix /*align 8*/, 73 ptrdiff_t stride, int alpha, int beta); 74 // h264_loop_filter_strength: simd only. the C version is inlined in h264_loopfilter.c 75 void (*h264_loop_filter_strength)(int16_t bS[2][4][4], uint8_t nnz[40], 76 int8_t ref[2][40], int16_t mv[2][40][2], 77 int bidir, int edges, int step, 78 int mask_mv0, int mask_mv1, int field); 79 80 /* IDCT */ 81 void (*h264_idct_add)(uint8_t *dst /*align 4*/, 82 int16_t *block /*align 16*/, int stride); 83 void (*h264_idct8_add)(uint8_t *dst /*align 8*/, 84 int16_t *block /*align 16*/, int stride); 85 void (*h264_idct_dc_add)(uint8_t *dst /*align 4*/, 86 int16_t *block /*align 16*/, int stride); 87 void (*h264_idct8_dc_add)(uint8_t *dst /*align 8*/, 88 int16_t *block /*align 16*/, int stride); 89 90 void (*h264_idct_add16)(uint8_t *dst /*align 16*/, const int *blockoffset, 91 int16_t *block /*align 16*/, int stride, 92 const uint8_t nnzc[5 * 8]); 93 void (*h264_idct8_add4)(uint8_t *dst /*align 16*/, const int *blockoffset, 94 int16_t *block /*align 16*/, int stride, 95 const uint8_t nnzc[5 * 8]); 96 void (*h264_idct_add8)(uint8_t **dst /*align 16*/, const int *blockoffset, 97 int16_t *block /*align 16*/, int stride, 98 const uint8_t nnzc[15 * 8]); 99 void (*h264_idct_add16intra)(uint8_t *dst /*align 16*/, const int *blockoffset, 100 int16_t *block /*align 16*/, 101 int stride, const uint8_t nnzc[5 * 8]); 102 void (*h264_luma_dc_dequant_idct)(int16_t *output, 103 int16_t *input /*align 16*/, int qmul); 104 void (*h264_chroma_dc_dequant_idct)(int16_t *block, int qmul); 105 106 /* bypass-transform */ 107 void (*h264_add_pixels8_clear)(uint8_t *dst, int16_t *block, int stride); 108 void (*h264_add_pixels4_clear)(uint8_t *dst, int16_t *block, int stride); 109 110 /** 111 * Search buf from the start for up to size bytes. Return the index 112 * of a zero byte, or >= size if not found. Ideally, use lookahead 113 * to filter out any zero bytes that are known to not be followed by 114 * one or more further zero bytes and a one byte. Better still, filter 115 * out any bytes that form the trailing_zero_8bits syntax element too. 116 */ 117 int (*startcode_find_candidate)(const uint8_t *buf, int size); 118} H264DSPContext; 119 120void ff_h264dsp_init(H264DSPContext *c, const int bit_depth, 121 const int chroma_format_idc); 122void ff_h264dsp_init_aarch64(H264DSPContext *c, const int bit_depth, 123 const int chroma_format_idc); 124void ff_h264dsp_init_arm(H264DSPContext *c, const int bit_depth, 125 const int chroma_format_idc); 126void ff_h264dsp_init_ppc(H264DSPContext *c, const int bit_depth, 127 const int chroma_format_idc); 128void ff_h264dsp_init_x86(H264DSPContext *c, const int bit_depth, 129 const int chroma_format_idc); 130void ff_h264dsp_init_mips(H264DSPContext *c, const int bit_depth, 131 const int chroma_format_idc); 132void ff_h264dsp_init_loongarch(H264DSPContext *c, const int bit_depth, 133 const int chroma_format_idc); 134 135#endif /* AVCODEC_H264DSP_H */ 136