1/* 2 * VP9 compatible video decoder 3 * 4 * Copyright (C) 2013 Ronald S. Bultje <rsbultje gmail com> 5 * Copyright (C) 2013 Clément Bœsch <u pkh me> 6 * 7 * This file is part of FFmpeg. 8 * 9 * FFmpeg is free software; you can redistribute it and/or 10 * modify it under the terms of the GNU Lesser General Public 11 * License as published by the Free Software Foundation; either 12 * version 2.1 of the License, or (at your option) any later version. 13 * 14 * FFmpeg is distributed in the hope that it will be useful, 15 * but WITHOUT ANY WARRANTY; without even the implied warranty of 16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU 17 * Lesser General Public License for more details. 18 * 19 * You should have received a copy of the GNU Lesser General Public 20 * License along with FFmpeg; if not, write to the Free Software 21 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA 22 */ 23 24#include "vp9dec.h" 25 26static av_always_inline void filter_plane_cols(VP9Context *s, int col, int ss_h, int ss_v, 27 uint8_t *lvl, uint8_t (*mask)[4], 28 uint8_t *dst, ptrdiff_t ls) 29{ 30 int y, x, bytesperpixel = s->bytesperpixel; 31 32 // filter edges between columns (e.g. block1 | block2) 33 for (y = 0; y < 8; y += 2 << ss_v, dst += 16 * ls, lvl += 16 << ss_v) { 34 uint8_t *ptr = dst, *l = lvl, *hmask1 = mask[y], *hmask2 = mask[y + 1 + ss_v]; 35 unsigned hm1 = hmask1[0] | hmask1[1] | hmask1[2], hm13 = hmask1[3]; 36 unsigned hm2 = hmask2[1] | hmask2[2], hm23 = hmask2[3]; 37 unsigned hm = hm1 | hm2 | hm13 | hm23; 38 39 for (x = 1; hm & ~(x - 1); x <<= 1, ptr += 8 * bytesperpixel >> ss_h) { 40 if (col || x > 1) { 41 if (hm1 & x) { 42 int L = *l, H = L >> 4; 43 int E = s->filter_lut.mblim_lut[L], I = s->filter_lut.lim_lut[L]; 44 45 if (hmask1[0] & x) { 46 if (hmask2[0] & x) { 47 av_assert2(l[8 << ss_v] == L); 48 s->dsp.loop_filter_16[0](ptr, ls, E, I, H); 49 } else { 50 s->dsp.loop_filter_8[2][0](ptr, ls, E, I, H); 51 } 52 } else if (hm2 & x) { 53 L = l[8 << ss_v]; 54 H |= (L >> 4) << 8; 55 E |= s->filter_lut.mblim_lut[L] << 8; 56 I |= s->filter_lut.lim_lut[L] << 8; 57 s->dsp.loop_filter_mix2[!!(hmask1[1] & x)] 58 [!!(hmask2[1] & x)] 59 [0](ptr, ls, E, I, H); 60 } else { 61 s->dsp.loop_filter_8[!!(hmask1[1] & x)] 62 [0](ptr, ls, E, I, H); 63 } 64 } else if (hm2 & x) { 65 int L = l[8 << ss_v], H = L >> 4; 66 int E = s->filter_lut.mblim_lut[L], I = s->filter_lut.lim_lut[L]; 67 68 s->dsp.loop_filter_8[!!(hmask2[1] & x)] 69 [0](ptr + 8 * ls, ls, E, I, H); 70 } 71 } 72 if (ss_h) { 73 if (x & 0xAA) 74 l += 2; 75 } else { 76 if (hm13 & x) { 77 int L = *l, H = L >> 4; 78 int E = s->filter_lut.mblim_lut[L], I = s->filter_lut.lim_lut[L]; 79 80 if (hm23 & x) { 81 L = l[8 << ss_v]; 82 H |= (L >> 4) << 8; 83 E |= s->filter_lut.mblim_lut[L] << 8; 84 I |= s->filter_lut.lim_lut[L] << 8; 85 s->dsp.loop_filter_mix2[0][0][0](ptr + 4 * bytesperpixel, ls, E, I, H); 86 } else { 87 s->dsp.loop_filter_8[0][0](ptr + 4 * bytesperpixel, ls, E, I, H); 88 } 89 } else if (hm23 & x) { 90 int L = l[8 << ss_v], H = L >> 4; 91 int E = s->filter_lut.mblim_lut[L], I = s->filter_lut.lim_lut[L]; 92 93 s->dsp.loop_filter_8[0][0](ptr + 8 * ls + 4 * bytesperpixel, ls, E, I, H); 94 } 95 l++; 96 } 97 } 98 } 99} 100 101static av_always_inline void filter_plane_rows(VP9Context *s, int row, int ss_h, int ss_v, 102 uint8_t *lvl, uint8_t (*mask)[4], 103 uint8_t *dst, ptrdiff_t ls) 104{ 105 int y, x, bytesperpixel = s->bytesperpixel; 106 107 // block1 108 // filter edges between rows (e.g. ------) 109 // block2 110 for (y = 0; y < 8; y++, dst += 8 * ls >> ss_v) { 111 uint8_t *ptr = dst, *l = lvl, *vmask = mask[y]; 112 unsigned vm = vmask[0] | vmask[1] | vmask[2], vm3 = vmask[3]; 113 114 for (x = 1; vm & ~(x - 1); x <<= (2 << ss_h), ptr += 16 * bytesperpixel, l += 2 << ss_h) { 115 if (row || y) { 116 if (vm & x) { 117 int L = *l, H = L >> 4; 118 int E = s->filter_lut.mblim_lut[L], I = s->filter_lut.lim_lut[L]; 119 120 if (vmask[0] & x) { 121 if (vmask[0] & (x << (1 + ss_h))) { 122 av_assert2(l[1 + ss_h] == L); 123 s->dsp.loop_filter_16[1](ptr, ls, E, I, H); 124 } else { 125 s->dsp.loop_filter_8[2][1](ptr, ls, E, I, H); 126 } 127 } else if (vm & (x << (1 + ss_h))) { 128 L = l[1 + ss_h]; 129 H |= (L >> 4) << 8; 130 E |= s->filter_lut.mblim_lut[L] << 8; 131 I |= s->filter_lut.lim_lut[L] << 8; 132 s->dsp.loop_filter_mix2[!!(vmask[1] & x)] 133 [!!(vmask[1] & (x << (1 + ss_h)))] 134 [1](ptr, ls, E, I, H); 135 } else { 136 s->dsp.loop_filter_8[!!(vmask[1] & x)] 137 [1](ptr, ls, E, I, H); 138 } 139 } else if (vm & (x << (1 + ss_h))) { 140 int L = l[1 + ss_h], H = L >> 4; 141 int E = s->filter_lut.mblim_lut[L], I = s->filter_lut.lim_lut[L]; 142 143 s->dsp.loop_filter_8[!!(vmask[1] & (x << (1 + ss_h)))] 144 [1](ptr + 8 * bytesperpixel, ls, E, I, H); 145 } 146 } 147 if (!ss_v) { 148 if (vm3 & x) { 149 int L = *l, H = L >> 4; 150 int E = s->filter_lut.mblim_lut[L], I = s->filter_lut.lim_lut[L]; 151 152 if (vm3 & (x << (1 + ss_h))) { 153 L = l[1 + ss_h]; 154 H |= (L >> 4) << 8; 155 E |= s->filter_lut.mblim_lut[L] << 8; 156 I |= s->filter_lut.lim_lut[L] << 8; 157 s->dsp.loop_filter_mix2[0][0][1](ptr + ls * 4, ls, E, I, H); 158 } else { 159 s->dsp.loop_filter_8[0][1](ptr + ls * 4, ls, E, I, H); 160 } 161 } else if (vm3 & (x << (1 + ss_h))) { 162 int L = l[1 + ss_h], H = L >> 4; 163 int E = s->filter_lut.mblim_lut[L], I = s->filter_lut.lim_lut[L]; 164 165 s->dsp.loop_filter_8[0][1](ptr + ls * 4 + 8 * bytesperpixel, ls, E, I, H); 166 } 167 } 168 } 169 if (ss_v) { 170 if (y & 1) 171 lvl += 16; 172 } else { 173 lvl += 8; 174 } 175 } 176} 177 178void ff_vp9_loopfilter_sb(AVCodecContext *avctx, VP9Filter *lflvl, 179 int row, int col, ptrdiff_t yoff, ptrdiff_t uvoff) 180{ 181 VP9Context *s = avctx->priv_data; 182 AVFrame *f = s->s.frames[CUR_FRAME].tf.f; 183 uint8_t *dst = f->data[0] + yoff; 184 ptrdiff_t ls_y = f->linesize[0], ls_uv = f->linesize[1]; 185 uint8_t (*uv_masks)[8][4] = lflvl->mask[s->ss_h | s->ss_v]; 186 int p; 187 188 /* FIXME: In how far can we interleave the v/h loopfilter calls? E.g. 189 * if you think of them as acting on a 8x8 block max, we can interleave 190 * each v/h within the single x loop, but that only works if we work on 191 * 8 pixel blocks, and we won't always do that (we want at least 16px 192 * to use SSE2 optimizations, perhaps 32 for AVX2) */ 193 194 filter_plane_cols(s, col, 0, 0, lflvl->level, lflvl->mask[0][0], dst, ls_y); 195 filter_plane_rows(s, row, 0, 0, lflvl->level, lflvl->mask[0][1], dst, ls_y); 196 197 for (p = 0; p < 2; p++) { 198 dst = f->data[1 + p] + uvoff; 199 filter_plane_cols(s, col, s->ss_h, s->ss_v, lflvl->level, uv_masks[0], dst, ls_uv); 200 filter_plane_rows(s, row, s->ss_h, s->ss_v, lflvl->level, uv_masks[1], dst, ls_uv); 201 } 202} 203