/* * VP9 compatible video decoder * * Copyright (C) 2013 Ronald S. Bultje <rsbultje gmail com> * Copyright (C) 2013 Clément Bœsch <u pkh me> * * This file is part of FFmpeg. * * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include"avcodec.h" #include"vp9dec.h"
static av_always_inline void filter_plane_cols(VP9Context *s, int col, int ss_h, int ss_v,
uint8_t *lvl, uint8_t (*mask)[4],
uint8_t *dst, ptrdiff_t ls)
{ int y, x, bytesperpixel = s->bytesperpixel;
for (x = 1; hm & ~(x - 1); x <<= 1, ptr += 8 * bytesperpixel >> ss_h) { if (col || x > 1) { if (hm1 & x) { int L = *l, H = L >> 4; int E = s->filter_lut.mblim_lut[L], I = s->filter_lut.lim_lut[L];
if (hmask1[0] & x) { if (hmask2[0] & x) {
av_assert2(l[8 << ss_v] == L);
s->dsp.loop_filter_16[0](ptr, ls, E, I, H);
} else {
s->dsp.loop_filter_8[2][0](ptr, ls, E, I, H);
}
} elseif (hm2 & x) {
L = l[8 << ss_v];
H |= (L >> 4) << 8;
E |= s->filter_lut.mblim_lut[L] << 8;
I |= s->filter_lut.lim_lut[L] << 8;
s->dsp.loop_filter_mix2[!!(hmask1[1] & x)]
[!!(hmask2[1] & x)]
[0](ptr, ls, E, I, H);
} else {
s->dsp.loop_filter_8[!!(hmask1[1] & x)]
[0](ptr, ls, E, I, H);
}
} elseif (hm2 & x) { int L = l[8 << ss_v], H = L >> 4; int E = s->filter_lut.mblim_lut[L], I = s->filter_lut.lim_lut[L];
s->dsp.loop_filter_8[!!(hmask2[1] & x)]
[0](ptr + 8 * ls, ls, E, I, H);
}
} if (ss_h) { if (x & 0xAA)
l += 2;
} else { if (hm13 & x) { int L = *l, H = L >> 4; int E = s->filter_lut.mblim_lut[L], I = s->filter_lut.lim_lut[L];
if (hm23 & x) {
L = l[8 << ss_v];
H |= (L >> 4) << 8;
E |= s->filter_lut.mblim_lut[L] << 8;
I |= s->filter_lut.lim_lut[L] << 8;
s->dsp.loop_filter_mix2[0][0][0](ptr + 4 * bytesperpixel, ls, E, I, H);
} else {
s->dsp.loop_filter_8[0][0](ptr + 4 * bytesperpixel, ls, E, I, H);
}
} elseif (hm23 & x) { int L = l[8 << ss_v], H = L >> 4; int E = s->filter_lut.mblim_lut[L], I = s->filter_lut.lim_lut[L];
s->dsp.loop_filter_8[0][0](ptr + 8 * ls + 4 * bytesperpixel, ls, E, I, H);
}
l++;
}
}
}
}
static av_always_inline void filter_plane_rows(VP9Context *s, int row, int ss_h, int ss_v,
uint8_t *lvl, uint8_t (*mask)[4],
uint8_t *dst, ptrdiff_t ls)
{ int y, x, bytesperpixel = s->bytesperpixel;
// block1 // filter edges between rows (e.g. ------) // block2 for (y = 0; y < 8; y++, dst += 8 * ls >> ss_v) {
uint8_t *ptr = dst, *l = lvl, *vmask = mask[y]; unsigned vm = vmask[0] | vmask[1] | vmask[2], vm3 = vmask[3];
for (x = 1; vm & ~(x - 1); x <<= (2 << ss_h), ptr += 16 * bytesperpixel, l += 2 << ss_h) { if (row || y) { if (vm & x) { int L = *l, H = L >> 4; int E = s->filter_lut.mblim_lut[L], I = s->filter_lut.lim_lut[L];
if (vmask[0] & x) { if (vmask[0] & (x << (1 + ss_h))) {
av_assert2(l[1 + ss_h] == L);
s->dsp.loop_filter_16[1](ptr, ls, E, I, H);
} else {
s->dsp.loop_filter_8[2][1](ptr, ls, E, I, H);
}
} elseif (vm & (x << (1 + ss_h))) {
L = l[1 + ss_h];
H |= (L >> 4) << 8;
E |= s->filter_lut.mblim_lut[L] << 8;
I |= s->filter_lut.lim_lut[L] << 8;
s->dsp.loop_filter_mix2[!!(vmask[1] & x)]
[!!(vmask[1] & (x << (1 + ss_h)))]
[1](ptr, ls, E, I, H);
} else {
s->dsp.loop_filter_8[!!(vmask[1] & x)]
[1](ptr, ls, E, I, H);
}
} elseif (vm & (x << (1 + ss_h))) { int L = l[1 + ss_h], H = L >> 4; int E = s->filter_lut.mblim_lut[L], I = s->filter_lut.lim_lut[L];
s->dsp.loop_filter_8[!!(vmask[1] & (x << (1 + ss_h)))]
[1](ptr + 8 * bytesperpixel, ls, E, I, H);
}
} if (!ss_v) { if (vm3 & x) { int L = *l, H = L >> 4; int E = s->filter_lut.mblim_lut[L], I = s->filter_lut.lim_lut[L];
if (vm3 & (x << (1 + ss_h))) {
L = l[1 + ss_h];
H |= (L >> 4) << 8;
E |= s->filter_lut.mblim_lut[L] << 8;
I |= s->filter_lut.lim_lut[L] << 8;
s->dsp.loop_filter_mix2[0][0][1](ptr + ls * 4, ls, E, I, H);
} else {
s->dsp.loop_filter_8[0][1](ptr + ls * 4, ls, E, I, H);
}
} elseif (vm3 & (x << (1 + ss_h))) { int L = l[1 + ss_h], H = L >> 4; int E = s->filter_lut.mblim_lut[L], I = s->filter_lut.lim_lut[L];
s->dsp.loop_filter_8[0][1](ptr + ls * 4 + 8 * bytesperpixel, ls, E, I, H);
}
}
} if (ss_v) { if (y & 1)
lvl += 16;
} else {
lvl += 8;
}
}
}
/* FIXME: In how far can we interleave the v/h loopfilter calls? E.g. * if you think of them as acting on a 8x8 block max, we can interleave * each v/h within the single x loop, but that only works if we work on * 8 pixel blocks, and we won't always do that (we want at least 16px
* to use SSE2 optimizations, perhaps 32 for AVX2) */
Die Informationen auf dieser Webseite wurden
nach bestem Wissen sorgfältig zusammengestellt. Es wird jedoch weder Vollständigkeit, noch Richtigkeit,
noch Qualität der bereit gestellten Informationen zugesichert.
Bemerkung:
Die farbliche Syntaxdarstellung und die Messung sind noch experimentell.