]> git.sesse.net Git - ffmpeg/blob - libavcodec/arm/h264dsp_init_arm.c
sgidec: make compiler optimize away memcpy call in inner loop.
[ffmpeg] / libavcodec / arm / h264dsp_init_arm.c
1 /*
2  * Copyright (c) 2010 Mans Rullgard <mans@mansr.com>
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 #include <stdint.h>
22
23 #include "libavcodec/dsputil.h"
24 #include "libavcodec/h264dsp.h"
25
26 void ff_h264_v_loop_filter_luma_neon(uint8_t *pix, int stride, int alpha,
27                                      int beta, int8_t *tc0);
28 void ff_h264_h_loop_filter_luma_neon(uint8_t *pix, int stride, int alpha,
29                                      int beta, int8_t *tc0);
30 void ff_h264_v_loop_filter_chroma_neon(uint8_t *pix, int stride, int alpha,
31                                        int beta, int8_t *tc0);
32 void ff_h264_h_loop_filter_chroma_neon(uint8_t *pix, int stride, int alpha,
33                                        int beta, int8_t *tc0);
34
35 void ff_weight_h264_pixels_16_neon(uint8_t *dst, int stride, int height,
36                                    int log2_den, int weight, int offset);
37 void ff_weight_h264_pixels_8_neon(uint8_t *dst, int stride, int height,
38                                   int log2_den, int weight, int offset);
39 void ff_weight_h264_pixels_4_neon(uint8_t *dst, int stride, int height,
40                                   int log2_den, int weight, int offset);
41
42 void ff_biweight_h264_pixels_16_neon(uint8_t *dst, uint8_t *src, int stride,
43                                      int height, int log2_den, int weightd,
44                                      int weights, int offset);
45 void ff_biweight_h264_pixels_8_neon(uint8_t *dst, uint8_t *src, int stride,
46                                     int height, int log2_den, int weightd,
47                                     int weights, int offset);
48 void ff_biweight_h264_pixels_4_neon(uint8_t *dst, uint8_t *src, int stride,
49                                     int height, int log2_den, int weightd,
50                                     int weights, int offset);
51
52 void ff_h264_idct_add_neon(uint8_t *dst, DCTELEM *block, int stride);
53 void ff_h264_idct_dc_add_neon(uint8_t *dst, DCTELEM *block, int stride);
54 void ff_h264_idct_add16_neon(uint8_t *dst, const int *block_offset,
55                              DCTELEM *block, int stride,
56                              const uint8_t nnzc[6*8]);
57 void ff_h264_idct_add16intra_neon(uint8_t *dst, const int *block_offset,
58                                   DCTELEM *block, int stride,
59                                   const uint8_t nnzc[6*8]);
60 void ff_h264_idct_add8_neon(uint8_t **dest, const int *block_offset,
61                             DCTELEM *block, int stride,
62                             const uint8_t nnzc[6*8]);
63
64 void ff_h264_idct8_add_neon(uint8_t *dst, DCTELEM *block, int stride);
65 void ff_h264_idct8_dc_add_neon(uint8_t *dst, DCTELEM *block, int stride);
66 void ff_h264_idct8_add4_neon(uint8_t *dst, const int *block_offset,
67                              DCTELEM *block, int stride,
68                              const uint8_t nnzc[6*8]);
69
70 static void ff_h264dsp_init_neon(H264DSPContext *c, const int bit_depth, const int chroma_format_idc)
71 {
72     if (bit_depth == 8) {
73     c->h264_v_loop_filter_luma   = ff_h264_v_loop_filter_luma_neon;
74     c->h264_h_loop_filter_luma   = ff_h264_h_loop_filter_luma_neon;
75     if(chroma_format_idc == 1){
76     c->h264_v_loop_filter_chroma = ff_h264_v_loop_filter_chroma_neon;
77     c->h264_h_loop_filter_chroma = ff_h264_h_loop_filter_chroma_neon;
78     }
79
80     c->weight_h264_pixels_tab[0] = ff_weight_h264_pixels_16_neon;
81     c->weight_h264_pixels_tab[1] = ff_weight_h264_pixels_8_neon;
82     c->weight_h264_pixels_tab[2] = ff_weight_h264_pixels_4_neon;
83
84     c->biweight_h264_pixels_tab[0] = ff_biweight_h264_pixels_16_neon;
85     c->biweight_h264_pixels_tab[1] = ff_biweight_h264_pixels_8_neon;
86     c->biweight_h264_pixels_tab[2] = ff_biweight_h264_pixels_4_neon;
87
88     c->h264_idct_add        = ff_h264_idct_add_neon;
89     c->h264_idct_dc_add     = ff_h264_idct_dc_add_neon;
90     c->h264_idct_add16      = ff_h264_idct_add16_neon;
91     c->h264_idct_add16intra = ff_h264_idct_add16intra_neon;
92     if (chroma_format_idc == 1)
93         c->h264_idct_add8   = ff_h264_idct_add8_neon;
94     c->h264_idct8_add       = ff_h264_idct8_add_neon;
95     c->h264_idct8_dc_add    = ff_h264_idct8_dc_add_neon;
96     c->h264_idct8_add4      = ff_h264_idct8_add4_neon;
97     }
98 }
99
100 void ff_h264dsp_init_arm(H264DSPContext *c, const int bit_depth, const int chroma_format_idc)
101 {
102     if (HAVE_NEON) ff_h264dsp_init_neon(c, bit_depth, chroma_format_idc);
103 }