diff options
Diffstat (limited to 'libavcodec/hevcdsp_template.c')
-rw-r--r-- | libavcodec/hevcdsp_template.c | 448 |
1 files changed, 229 insertions, 219 deletions
diff --git a/libavcodec/hevcdsp_template.c b/libavcodec/hevcdsp_template.c index 027b77c7c7..b91ed3dfa9 100644 --- a/libavcodec/hevcdsp_template.c +++ b/libavcodec/hevcdsp_template.c @@ -1,37 +1,41 @@ /* - * HEVC video decoder + * HEVC video Decoder * * Copyright (C) 2012 - 2013 Guillaume Martres * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * 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. * - * Libav is distributed in the hope that it will be useful, + * 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 Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ #include "get_bits.h" +#include "bit_depth_template.c" +#include "hevcdsp.h" + #include "hevc.h" -#include "bit_depth_template.c" +#define SET(dst, x) (dst) = (x) +#define SCALE(dst, x) (dst) = av_clip_int16(((x) + add) >> shift) +#define ADD_AND_SCALE(dst, x) (dst) = av_clip_pixel((dst) + av_clip_int16(((x) + add) >> shift)) -static void FUNC(put_pcm)(uint8_t *_dst, ptrdiff_t stride, int size, +static void FUNC(put_pcm)(uint8_t *_dst, ptrdiff_t _stride, int size, GetBitContext *gb, int pcm_bit_depth) { int x, y; - pixel *dst = (pixel *)_dst; - - stride /= sizeof(pixel); + pixel *dst = (pixel*)_dst; + ptrdiff_t stride = _stride / sizeof(pixel); for (y = 0; y < size; y++) { for (x = 0; x < size; x++) @@ -40,13 +44,11 @@ static void FUNC(put_pcm)(uint8_t *_dst, ptrdiff_t stride, int size, } } -static void FUNC(transquant_bypass4x4)(uint8_t *_dst, int16_t *coeffs, - ptrdiff_t stride) +static void FUNC(transquant_bypass4x4)(uint8_t *_dst, int16_t *coeffs, ptrdiff_t _stride) { int x, y; - pixel *dst = (pixel *)_dst; - - stride /= sizeof(pixel); + pixel *dst = (pixel *)_dst; + ptrdiff_t stride = _stride / sizeof(pixel); for (y = 0; y < 4; y++) { for (x = 0; x < 4; x++) { @@ -55,15 +57,14 @@ static void FUNC(transquant_bypass4x4)(uint8_t *_dst, int16_t *coeffs, } dst += stride; } + } -static void FUNC(transquant_bypass8x8)(uint8_t *_dst, int16_t *coeffs, - ptrdiff_t stride) +static void FUNC(transquant_bypass8x8)(uint8_t *_dst, int16_t *coeffs, ptrdiff_t _stride) { int x, y; - pixel *dst = (pixel *)_dst; - - stride /= sizeof(pixel); + pixel *dst = (pixel*)_dst; + ptrdiff_t stride = _stride / sizeof(pixel); for (y = 0; y < 8; y++) { for (x = 0; x < 8; x++) { @@ -74,13 +75,11 @@ static void FUNC(transquant_bypass8x8)(uint8_t *_dst, int16_t *coeffs, } } -static void FUNC(transquant_bypass16x16)(uint8_t *_dst, int16_t *coeffs, - ptrdiff_t stride) +static void FUNC(transquant_bypass16x16)(uint8_t *_dst, int16_t *coeffs, ptrdiff_t _stride) { int x, y; - pixel *dst = (pixel *)_dst; - - stride /= sizeof(pixel); + pixel *dst = (pixel*)_dst; + ptrdiff_t stride = _stride / sizeof(pixel); for (y = 0; y < 16; y++) { for (x = 0; x < 16; x++) { @@ -89,15 +88,14 @@ static void FUNC(transquant_bypass16x16)(uint8_t *_dst, int16_t *coeffs, } dst += stride; } + } -static void FUNC(transquant_bypass32x32)(uint8_t *_dst, int16_t *coeffs, - ptrdiff_t stride) +static void FUNC(transquant_bypass32x32)(uint8_t *_dst, int16_t *coeffs, ptrdiff_t _stride) { int x, y; - pixel *dst = (pixel *)_dst; - - stride /= sizeof(pixel); + pixel *dst = (pixel*)_dst; + ptrdiff_t stride = _stride / sizeof(pixel); for (y = 0; y < 32; y++) { for (x = 0; x < 32; x++) { @@ -108,11 +106,11 @@ static void FUNC(transquant_bypass32x32)(uint8_t *_dst, int16_t *coeffs, } } -static void FUNC(transform_skip)(uint8_t *_dst, int16_t *coeffs, - ptrdiff_t stride) +static void FUNC(transform_skip)(uint8_t *_dst, int16_t *coeffs, ptrdiff_t _stride) { - pixel *dst = (pixel *)_dst; - int shift = 13 - BIT_DEPTH; + pixel *dst = (pixel*)_dst; + ptrdiff_t stride = _stride / sizeof(pixel); + int shift = 13 - BIT_DEPTH; #if BIT_DEPTH <= 13 int offset = 1 << (shift - 1); #else @@ -120,106 +118,96 @@ static void FUNC(transform_skip)(uint8_t *_dst, int16_t *coeffs, #endif int x, y; - stride /= sizeof(pixel); - - for (y = 0; y < 4 * 4; y += 4) { - for (x = 0; x < 4; x++) + for (y = 0; y < 4*4; y+=4) { + for (x = 0; x < 4; x++) { dst[x] = av_clip_pixel(dst[x] + ((coeffs[y + x] + offset) >> shift)); + } dst += stride; } } -#define SET(dst, x) (dst) = (x) -#define SCALE(dst, x) (dst) = av_clip_int16(((x) + add) >> shift) -#define ADD_AND_SCALE(dst, x) \ - (dst) = av_clip_pixel((dst) + av_clip_int16(((x) + add) >> shift)) - -#define TR_4x4_LUMA(dst, src, step, assign) \ - do { \ - int c0 = src[0 * step] + src[2 * step]; \ - int c1 = src[2 * step] + src[3 * step]; \ - int c2 = src[0 * step] - src[3 * step]; \ - int c3 = 74 * src[1 * step]; \ - \ - assign(dst[2 * step], 74 * (src[0 * step] - \ - src[2 * step] + \ - src[3 * step])); \ - assign(dst[0 * step], 29 * c0 + 55 * c1 + c3); \ - assign(dst[1 * step], 55 * c2 - 29 * c1 + c3); \ - assign(dst[3 * step], 55 * c0 + 29 * c2 - c3); \ +static void FUNC(transform_4x4_luma_add)(uint8_t *_dst, int16_t *coeffs, ptrdiff_t _stride) +{ +#define TR_4x4_LUMA(dst, src, step, assign) \ + do { \ + int c0 = src[0*step] + src[2*step]; \ + int c1 = src[2*step] + src[3*step]; \ + int c2 = src[0*step] - src[3*step]; \ + int c3 = 74 * src[1*step]; \ + \ + assign(dst[2*step], 74 * (src[0*step] - src[2*step] + src[3*step])); \ + assign(dst[0*step], 29 * c0 + 55 * c1 + c3); \ + assign(dst[1*step], 55 * c2 - 29 * c1 + c3); \ + assign(dst[3*step], 55 * c0 + 29 * c2 - c3); \ } while (0) -static void FUNC(transform_4x4_luma_add)(uint8_t *_dst, int16_t *coeffs, - ptrdiff_t stride) -{ int i; - pixel *dst = (pixel *)_dst; - int shift = 7; - int add = 1 << (shift - 1); + pixel *dst = (pixel*)_dst; + ptrdiff_t stride = _stride / sizeof(pixel); + int shift = 7; + int add = 1 << (shift - 1); int16_t *src = coeffs; - stride /= sizeof(pixel); - for (i = 0; i < 4; i++) { TR_4x4_LUMA(src, src, 4, SCALE); src++; } shift = 20 - BIT_DEPTH; - add = 1 << (shift - 1); + add = 1 << (shift - 1); for (i = 0; i < 4; i++) { TR_4x4_LUMA(dst, coeffs, 1, ADD_AND_SCALE); coeffs += 4; - dst += stride; + dst += stride; } -} #undef TR_4x4_LUMA +} -#define TR_4(dst, src, dstep, sstep, assign) \ - do { \ - const int e0 = transform[8 * 0][0] * src[0 * sstep] + \ - transform[8 * 2][0] * src[2 * sstep]; \ - const int e1 = transform[8 * 0][1] * src[0 * sstep] + \ - transform[8 * 2][1] * src[2 * sstep]; \ - const int o0 = transform[8 * 1][0] * src[1 * sstep] + \ - transform[8 * 3][0] * src[3 * sstep]; \ - const int o1 = transform[8 * 1][1] * src[1 * sstep] + \ - transform[8 * 3][1] * src[3 * sstep]; \ - \ - assign(dst[0 * dstep], e0 + o0); \ - assign(dst[1 * dstep], e1 + o1); \ - assign(dst[2 * dstep], e1 - o1); \ - assign(dst[3 * dstep], e0 - o0); \ +#define TR_4(dst, src, dstep, sstep, assign) \ + do { \ + const int e0 = transform[8*0][0] * src[0*sstep] + \ + transform[8*2][0] * src[2*sstep]; \ + const int e1 = transform[8*0][1] * src[0*sstep] + \ + transform[8*2][1] * src[2*sstep]; \ + const int o0 = transform[8*1][0] * src[1*sstep] + \ + transform[8*3][0] * src[3*sstep]; \ + const int o1 = transform[8*1][1] * src[1*sstep] + \ + transform[8*3][1] * src[3*sstep]; \ + \ + assign(dst[0*dstep], e0 + o0); \ + assign(dst[1*dstep], e1 + o1); \ + assign(dst[2*dstep], e1 - o1); \ + assign(dst[3*dstep], e0 - o0); \ } while (0) +#define TR_4_1(dst, src) TR_4(dst, src, 4, 4, SCALE) +#define TR_4_2(dst, src) TR_4(dst, src, 1, 1, ADD_AND_SCALE) -static void FUNC(transform_4x4_add)(uint8_t *_dst, int16_t *coeffs, - ptrdiff_t stride) +static void FUNC(transform_4x4_add)(uint8_t *_dst, int16_t *coeffs, ptrdiff_t _stride) { int i; - pixel *dst = (pixel *)_dst; - int shift = 7; - int add = 1 << (shift - 1); + pixel *dst = (pixel*)_dst; + ptrdiff_t stride = _stride / sizeof(pixel); + int shift = 7; + int add = 1 << (shift - 1); int16_t *src = coeffs; - stride /= sizeof(pixel); - for (i = 0; i < 4; i++) { - TR_4(src, src, 4, 4, SCALE); + TR_4_1(src, src); src++; } shift = 20 - BIT_DEPTH; - add = 1 << (shift - 1); + add = 1 << (shift - 1); for (i = 0; i < 4; i++) { - TR_4(dst, coeffs, 1, 1, ADD_AND_SCALE); + TR_4_2(dst, coeffs); coeffs += 4; - dst += stride; + dst += stride; } } -#define TR_8(dst, src, dstep, sstep, assign) \ - do { \ +#define TR_8(dst, src, dstep, sstep, assign) \ + do { \ int i, j; \ int e_8[4]; \ int o_8[4] = { 0 }; \ @@ -266,118 +254,131 @@ static void FUNC(transform_4x4_add)(uint8_t *_dst, int16_t *coeffs, } \ } while (0) +#define TR_8_1(dst, src) TR_8(dst, src, 8, 8, SCALE) +#define TR_16_1(dst, src) TR_16(dst, src, 16, 16, SCALE) +#define TR_32_1(dst, src) TR_32(dst, src, 32, 32, SCALE) +#define TR_8_2(dst, src) TR_8(dst, src, 1, 1, ADD_AND_SCALE) +#define TR_16_2(dst, src) TR_16(dst, src, 1, 1, ADD_AND_SCALE) +#define TR_32_2(dst, src) TR_32(dst, src, 1, 1, ADD_AND_SCALE) -static void FUNC(transform_8x8_add)(uint8_t *_dst, int16_t *coeffs, - ptrdiff_t stride) +static void FUNC(transform_8x8_add)(uint8_t *_dst, int16_t *coeffs, ptrdiff_t _stride) { int i; - pixel *dst = (pixel *)_dst; - int shift = 7; - int add = 1 << (shift - 1); + pixel *dst = (pixel *)_dst; + ptrdiff_t stride = _stride / sizeof(pixel); + int shift = 7; + int add = 1 << (shift - 1); int16_t *src = coeffs; - stride /= sizeof(pixel); - for (i = 0; i < 8; i++) { - TR_8(src, src, 8, 8, SCALE); + TR_8_1(src, src); src++; } shift = 20 - BIT_DEPTH; - add = 1 << (shift - 1); + add = 1 << (shift - 1); for (i = 0; i < 8; i++) { - TR_8(dst, coeffs, 1, 1, ADD_AND_SCALE); + TR_8_2(dst, coeffs); coeffs += 8; - dst += stride; + dst += stride; } } -static void FUNC(transform_16x16_add)(uint8_t *_dst, int16_t *coeffs, - ptrdiff_t stride) +static void FUNC(transform_16x16_add)(uint8_t *_dst, int16_t *coeffs, ptrdiff_t _stride) { int i; - pixel *dst = (pixel *)_dst; - int shift = 7; - int add = 1 << (shift - 1); + pixel *dst = (pixel*)_dst; + ptrdiff_t stride = _stride / sizeof(pixel); + int shift = 7; + int add = 1 << (shift - 1); int16_t *src = coeffs; - stride /= sizeof(pixel); - for (i = 0; i < 16; i++) { - TR_16(src, src, 16, 16, SCALE); + TR_16_1(src, src); src++; } shift = 20 - BIT_DEPTH; - add = 1 << (shift - 1); + add = 1 << (shift - 1); for (i = 0; i < 16; i++) { - TR_16(dst, coeffs, 1, 1, ADD_AND_SCALE); + TR_16_2(dst, coeffs); coeffs += 16; - dst += stride; + dst += stride; } } -static void FUNC(transform_32x32_add)(uint8_t *_dst, int16_t *coeffs, - ptrdiff_t stride) +static void FUNC(transform_32x32_add)(uint8_t *_dst, int16_t *coeffs, ptrdiff_t _stride) { +#define IT32x32_even(i,w) ( src[ 0*w] * transform[ 0][i] ) + ( src[16*w] * transform[16][i] ) +#define IT32x32_odd(i,w) ( src[ 8*w] * transform[ 8][i] ) + ( src[24*w] * transform[24][i] ) +#define IT16x16(i,w) ( src[ 4*w] * transform[ 4][i] ) + ( src[12*w] * transform[12][i] ) + ( src[20*w] * transform[20][i] ) + ( src[28*w] * transform[28][i] ) +#define IT8x8(i,w) ( src[ 2*w] * transform[ 2][i] ) + ( src[ 6*w] * transform[ 6][i] ) + ( src[10*w] * transform[10][i] ) + ( src[14*w] * transform[14][i] ) + \ + ( src[18*w] * transform[18][i] ) + ( src[22*w] * transform[22][i] ) + ( src[26*w] * transform[26][i] ) + ( src[30*w] * transform[30][i] ) +#define IT4x4(i,w) ( src[ 1*w] * transform[ 1][i] ) + ( src[ 3*w] * transform[ 3][i] ) + ( src[ 5*w] * transform[ 5][i] ) + ( src[ 7*w] * transform[ 7][i] ) + \ + ( src[ 9*w] * transform[ 9][i] ) + ( src[11*w] * transform[11][i] ) + ( src[13*w] * transform[13][i] ) + ( src[15*w] * transform[15][i] ) + \ + ( src[17*w] * transform[17][i] ) + ( src[19*w] * transform[19][i] ) + ( src[21*w] * transform[21][i] ) + ( src[23*w] * transform[23][i] ) + \ + ( src[25*w] * transform[25][i] ) + ( src[27*w] * transform[27][i] ) + ( src[29*w] * transform[29][i] ) + ( src[31*w] * transform[31][i] ) int i; - pixel *dst = (pixel *)_dst; - int shift = 7; - int add = 1 << (shift - 1); + pixel *dst = (pixel*)_dst; + ptrdiff_t stride = _stride / sizeof(pixel); + int shift = 7; + int add = 1 << (shift - 1); int16_t *src = coeffs; - stride /= sizeof(pixel); - for (i = 0; i < 32; i++) { - TR_32(src, src, 32, 32, SCALE); + TR_32_1(src, src); src++; } src = coeffs; shift = 20 - BIT_DEPTH; add = 1 << (shift - 1); for (i = 0; i < 32; i++) { - TR_32(dst, coeffs, 1, 1, ADD_AND_SCALE); + TR_32_2(dst, coeffs); coeffs += 32; - dst += stride; + dst += stride; } +#undef IT32x32_even +#undef IT32x32_odd +#undef IT16x16 +#undef IT8x8 +#undef IT4x4 } static void FUNC(sao_band_filter)(uint8_t *_dst, uint8_t *_src, - ptrdiff_t stride, SAOParams *sao, - int *borders, int width, int height, + ptrdiff_t _stride, SAOParams *sao, + int *borders, int width, int height, int c_idx, int class) { - pixel *dst = (pixel *)_dst; - pixel *src = (pixel *)_src; + pixel *dst = (pixel*)_dst; + pixel *src = (pixel*)_src; + ptrdiff_t stride = _stride / sizeof(pixel); int offset_table[32] = { 0 }; int k, y, x; int chroma = !!c_idx; - int shift = BIT_DEPTH - 5; + int shift = BIT_DEPTH - 5; int *sao_offset_val = sao->offset_val[c_idx]; int sao_left_class = sao->band_position[c_idx]; int init_y = 0, init_x = 0; - stride /= sizeof(pixel); - switch (class) { case 0: if (!borders[2]) - width -= (8 >> chroma) + 2; + width -= ((8 >> chroma) + 2); if (!borders[3]) - height -= (4 >> chroma) + 2; + height -= ((4 >> chroma) + 2); break; case 1: init_y = -(4 >> chroma) - 2; if (!borders[2]) - width -= (8 >> chroma) + 2; + width -= ((8 >> chroma) + 2); height = (4 >> chroma) + 2; break; case 2: init_x = -(8 >> chroma) - 2; width = (8 >> chroma) + 2; if (!borders[3]) - height -= (4 >> chroma) + 2; + height -= ((4 >> chroma) + 2); break; case 3: init_y = -(4 >> chroma) - 2; @@ -422,32 +423,30 @@ static void FUNC(sao_band_filter_2)(uint8_t *dst, uint8_t *src, int *borders, int width, int height, int c_idx) { - FUNC(sao_band_filter)(dst, src, stride, sao, borders, - width, height, c_idx, 2); + FUNC(sao_band_filter)(dst, src, stride, sao, borders, width, height, c_idx, 2); } static void FUNC(sao_band_filter_3)(uint8_t *_dst, uint8_t *_src, - ptrdiff_t stride, SAOParams *sao, + ptrdiff_t _stride, SAOParams *sao, int *borders, int width, int height, int c_idx) { - FUNC(sao_band_filter)(_dst, _src, stride, sao, borders, - width, height, c_idx, 3); + FUNC(sao_band_filter)(_dst, _src, _stride, sao, borders, width, height, c_idx, 3); } static void FUNC(sao_edge_filter_0)(uint8_t *_dst, uint8_t *_src, - ptrdiff_t stride, SAOParams *sao, + ptrdiff_t _stride, SAOParams *sao, int *borders, int _width, int _height, - int c_idx, uint8_t vert_edge, - uint8_t horiz_edge, uint8_t diag_edge) + int c_idx, + uint8_t vert_edge, uint8_t horiz_edge, uint8_t diag_edge) { int x, y; - pixel *dst = (pixel *)_dst; - pixel *src = (pixel *)_src; + pixel *dst = (pixel*)_dst; + pixel *src = (pixel*)_src; + ptrdiff_t stride = _stride / sizeof(pixel); int chroma = !!c_idx; int *sao_offset_val = sao->offset_val[c_idx]; - int sao_eo_class = sao->eo_class[c_idx]; - int init_x = 0, init_y = 0, width = _width, height = _height; + int sao_eo_class = sao->eo_class[c_idx]; static const int8_t pos[4][2][2] = { { { -1, 0 }, { 1, 0 } }, // horizontal @@ -457,9 +456,9 @@ static void FUNC(sao_edge_filter_0)(uint8_t *_dst, uint8_t *_src, }; static const uint8_t edge_idx[] = { 1, 2, 0, 3, 4 }; -#define CMP(a, b) ((a) > (b) ? 1 : ((a) == (b) ? 0 : -1)) + int init_x = 0, init_y = 0, width = _width, height = _height; - stride /= sizeof(pixel); +#define CMP(a, b) ((a) > (b) ? 1 : ((a) == (b) ? 0 : -1)) if (!borders[2]) width -= (8 >> chroma) + 2; @@ -492,15 +491,17 @@ static void FUNC(sao_edge_filter_0)(uint8_t *_dst, uint8_t *_src, if (sao_eo_class != SAO_EO_HORIZ) { if (borders[1]) { int offset_val = sao_offset_val[0]; - for (x = init_x; x < width; x++) + for (x = init_x; x < width; x++) { dst[x] = av_clip_pixel(src[x] + offset_val); + } init_y = 1; } if (borders[3]) { int offset_val = sao_offset_val[0]; int y_stride = stride * (height - 1); - for (x = init_x; x < width; x++) + for (x = init_x; x < width; x++) { dst[x + y_stride] = av_clip_pixel(src[x + y_stride] + offset_val); + } height--; } } @@ -543,7 +544,7 @@ static void FUNC(sao_edge_filter_0)(uint8_t *_dst, uint8_t *_src, } static void FUNC(sao_edge_filter_1)(uint8_t *_dst, uint8_t *_src, - ptrdiff_t stride, SAOParams *sao, + ptrdiff_t _stride, SAOParams *sao, int *borders, int _width, int _height, int c_idx, uint8_t vert_edge, uint8_t horiz_edge, uint8_t diag_edge) @@ -551,10 +552,10 @@ static void FUNC(sao_edge_filter_1)(uint8_t *_dst, uint8_t *_src, int x, y; pixel *dst = (pixel *)_dst; pixel *src = (pixel *)_src; + ptrdiff_t stride = _stride / sizeof(pixel); int chroma = !!c_idx; int *sao_offset_val = sao->offset_val[c_idx]; int sao_eo_class = sao->eo_class[c_idx]; - int init_x = 0, init_y = 0, width = _width, height = _height; static const int8_t pos[4][2][2] = { { { -1, 0 }, { 1, 0 } }, // horizontal @@ -564,9 +565,9 @@ static void FUNC(sao_edge_filter_1)(uint8_t *_dst, uint8_t *_src, }; static const uint8_t edge_idx[] = { 1, 2, 0, 3, 4 }; -#define CMP(a, b) ((a) > (b) ? 1 : ((a) == (b) ? 0 : -1)) + int init_x = 0, init_y = 0, width = _width, height = _height; - stride /= sizeof(pixel); +#define CMP(a, b) ((a) > (b) ? 1 : ((a) == (b) ? 0 : -1)) init_y = -(4 >> chroma) - 2; if (!borders[2]) @@ -635,7 +636,7 @@ static void FUNC(sao_edge_filter_1)(uint8_t *_dst, uint8_t *_src, } static void FUNC(sao_edge_filter_2)(uint8_t *_dst, uint8_t *_src, - ptrdiff_t stride, SAOParams *sao, + ptrdiff_t _stride, SAOParams *sao, int *borders, int _width, int _height, int c_idx, uint8_t vert_edge, uint8_t horiz_edge, uint8_t diag_edge) @@ -643,10 +644,10 @@ static void FUNC(sao_edge_filter_2)(uint8_t *_dst, uint8_t *_src, int x, y; pixel *dst = (pixel *)_dst; pixel *src = (pixel *)_src; + ptrdiff_t stride = _stride / sizeof(pixel); int chroma = !!c_idx; int *sao_offset_val = sao->offset_val[c_idx]; int sao_eo_class = sao->eo_class[c_idx]; - int init_x = 0, init_y = 0, width = _width, height = _height; static const int8_t pos[4][2][2] = { { { -1, 0 }, { 1, 0 } }, // horizontal @@ -656,9 +657,9 @@ static void FUNC(sao_edge_filter_2)(uint8_t *_dst, uint8_t *_src, }; static const uint8_t edge_idx[] = { 1, 2, 0, 3, 4 }; -#define CMP(a, b) ((a) > (b) ? 1 : ((a) == (b) ? 0 : -1)) + int init_x = 0, init_y = 0, width = _width, height = _height; - stride /= sizeof(pixel); +#define CMP(a, b) ((a) > (b) ? 1 : ((a) == (b) ? 0 : -1)) init_x = -(8 >> chroma) - 2; width = (8 >> chroma) + 2; @@ -671,15 +672,17 @@ static void FUNC(sao_edge_filter_2)(uint8_t *_dst, uint8_t *_src, if (sao_eo_class != SAO_EO_HORIZ) { if (borders[1]) { int offset_val = sao_offset_val[0]; - for (x = init_x; x < width; x++) + for (x = init_x; x < width; x++) { dst[x] = av_clip_pixel(src[x] + offset_val); + } init_y = 1; } if (borders[3]) { int offset_val = sao_offset_val[0]; int y_stride = stride * (height - 1); - for (x = init_x; x < width; x++) + for (x = init_x; x < width; x++) { dst[x + y_stride] = av_clip_pixel(src[x + y_stride] + offset_val); + } height--; } } @@ -721,7 +724,7 @@ static void FUNC(sao_edge_filter_2)(uint8_t *_dst, uint8_t *_src, } static void FUNC(sao_edge_filter_3)(uint8_t *_dst, uint8_t *_src, - ptrdiff_t stride, SAOParams *sao, + ptrdiff_t _stride, SAOParams *sao, int *borders, int _width, int _height, int c_idx, uint8_t vert_edge, uint8_t horiz_edge, uint8_t diag_edge) @@ -729,10 +732,10 @@ static void FUNC(sao_edge_filter_3)(uint8_t *_dst, uint8_t *_src, int x, y; pixel *dst = (pixel *)_dst; pixel *src = (pixel *)_src; + ptrdiff_t stride = _stride / sizeof(pixel); int chroma = !!c_idx; int *sao_offset_val = sao->offset_val[c_idx]; int sao_eo_class = sao->eo_class[c_idx]; - int init_x = 0, init_y = 0, width = _width, height = _height; static const int8_t pos[4][2][2] = { { { -1, 0 }, { 1, 0 } }, // horizontal @@ -742,9 +745,9 @@ static void FUNC(sao_edge_filter_3)(uint8_t *_dst, uint8_t *_src, }; static const uint8_t edge_idx[] = { 1, 2, 0, 3, 4 }; -#define CMP(a, b) ((a) > (b) ? 1 : ((a) == (b) ? 0 : -1)) + int init_x = 0, init_y = 0, width = _width, height = _height; - stride /= sizeof(pixel); +#define CMP(a, b) ((a) > (b) ? 1 : ((a) == (b) ? 0 : -1)) init_y = -(4 >> chroma) - 2; init_x = -(8 >> chroma) - 2; @@ -798,9 +801,17 @@ static void FUNC(sao_edge_filter_3)(uint8_t *_dst, uint8_t *_src, #undef SCALE #undef ADD_AND_SCALE #undef TR_4 +#undef TR_4_1 +#undef TR_4_2 #undef TR_8 +#undef TR_8_1 +#undef TR_8_2 #undef TR_16 +#undef TR_16_1 +#undef TR_16_2 #undef TR_32 +#undef TR_32_1 +#undef TR_32_2 static void FUNC(put_hevc_qpel_pixels)(int16_t *dst, ptrdiff_t dststride, uint8_t *_src, ptrdiff_t _srcstride, @@ -819,32 +830,16 @@ static void FUNC(put_hevc_qpel_pixels)(int16_t *dst, ptrdiff_t dststride, } #define QPEL_FILTER_1(src, stride) \ - (1 * -src[x - 3 * stride] + \ - 4 * src[x - 2 * stride] - \ - 10 * src[x - stride] + \ - 58 * src[x] + \ - 17 * src[x + stride] - \ - 5 * src[x + 2 * stride] + \ - 1 * src[x + 3 * stride]) + (-src[x - 3 * stride] + 4 * src[x - 2 * stride] - 10 * src[x - stride] + 58 * src[x] + \ + 17 * src[x + stride] - 5 * src[x + 2 * stride] + 1 * src[x + 3 * stride]) #define QPEL_FILTER_2(src, stride) \ - (1 * -src[x - 3 * stride] + \ - 4 * src[x - 2 * stride] - \ - 11 * src[x - stride] + \ - 40 * src[x] + \ - 40 * src[x + stride] - \ - 11 * src[x + 2 * stride] + \ - 4 * src[x + 3 * stride] - \ - 1 * src[x + 4 * stride]) + (-src[x - 3 * stride] + 4 * src[x - 2 * stride] - 11 * src[x - stride] + 40 * src[x] + \ + 40 * src[x + stride] - 11 * src[x + 2 * stride] + 4 * src[x + 3 * stride] - src[x + 4 * stride]) #define QPEL_FILTER_3(src, stride) \ - (1 * src[x - 2 * stride] - \ - 5 * src[x - stride] + \ - 17 * src[x] + \ - 58 * src[x + stride] - \ - 10 * src[x + 2 * stride] + \ - 4 * src[x + 3 * stride] - \ - 1 * src[x + 4 * stride]) + (src[x - 2 * stride] - 5 * src[x - stride] + 17 * src[x] + 58 * src[x + stride] \ + - 10 * src[x + 2 * stride] + 4 * src[x + 3 * stride] - src[x + 4 * stride]) #define PUT_HEVC_QPEL_H(H) \ @@ -884,10 +879,8 @@ static void FUNC(put_hevc_qpel_v ## V)(int16_t *dst, ptrdiff_t dststride, \ } #define PUT_HEVC_QPEL_HV(H, V) \ -static void FUNC(put_hevc_qpel_h ## H ## v ## V)(int16_t *dst, \ - ptrdiff_t dststride, \ - uint8_t *_src, \ - ptrdiff_t _srcstride, \ +static void FUNC(put_hevc_qpel_h ## H ## v ## V)(int16_t *dst, ptrdiff_t dststride, \ + uint8_t *_src, ptrdiff_t _srcstride,\ int width, int height, \ int16_t* mcbuffer) \ { \ @@ -943,18 +936,16 @@ static void FUNC(put_hevc_epel_pixels)(int16_t *dst, ptrdiff_t dststride, ptrdiff_t srcstride = _srcstride / sizeof(pixel); for (y = 0; y < height; y++) { - for (x = 0; x < width; x++) + for (x = 0; x < width; x++) { dst[x] = src[x] << (14 - BIT_DEPTH); + } src += srcstride; dst += dststride; } } #define EPEL_FILTER(src, stride) \ - (filter_0 * src[x - stride] + \ - filter_1 * src[x] + \ - filter_2 * src[x + stride] + \ - filter_3 * src[x + 2 * stride]) + (filter_0 * src[x - stride] + filter_1 * src[x] + filter_2 * src[x + stride] + filter_3 * src[x + 2 * stride]) static void FUNC(put_hevc_epel_h)(int16_t *dst, ptrdiff_t dststride, uint8_t *_src, ptrdiff_t _srcstride, @@ -970,8 +961,9 @@ static void FUNC(put_hevc_epel_h)(int16_t *dst, ptrdiff_t dststride, int8_t filter_2 = filter[2]; int8_t filter_3 = filter[3]; for (y = 0; y < height; y++) { - for (x = 0; x < width; x++) + for (x = 0; x < width; x++) { dst[x] = EPEL_FILTER(src, 1) >> (BIT_DEPTH - 8); + } src += srcstride; dst += dststride; } @@ -992,8 +984,9 @@ static void FUNC(put_hevc_epel_v)(int16_t *dst, ptrdiff_t dststride, int8_t filter_3 = filter[3]; for (y = 0; y < height; y++) { - for (x = 0; x < width; x++) + for (x = 0; x < width; x++) { dst[x] = EPEL_FILTER(src, srcstride) >> (BIT_DEPTH - 8); + } src += srcstride; dst += dststride; } @@ -1019,8 +1012,9 @@ static void FUNC(put_hevc_epel_hv)(int16_t *dst, ptrdiff_t dststride, src -= EPEL_EXTRA_BEFORE * srcstride; for (y = 0; y < height + EPEL_EXTRA; y++) { - for (x = 0; x < width; x++) + for (x = 0; x < width; x++) { tmp[x] = EPEL_FILTER(src, 1) >> (BIT_DEPTH - 8); + } src += srcstride; tmp += MAX_PB_SIZE; } @@ -1031,8 +1025,9 @@ static void FUNC(put_hevc_epel_hv)(int16_t *dst, ptrdiff_t dststride, filter_2 = filter_v[2]; filter_3 = filter_v[3]; for (y = 0; y < height; y++) { - for (x = 0; x < width; x++) + for (x = 0; x < width; x++) { dst[x] = EPEL_FILTER(tmp, MAX_PB_SIZE) >> 6; + } tmp += MAX_PB_SIZE; dst += dststride; } @@ -1053,8 +1048,9 @@ static void FUNC(put_unweighted_pred)(uint8_t *_dst, ptrdiff_t _dststride, int offset = 0; #endif for (y = 0; y < height; y++) { - for (x = 0; x < width; x++) + for (x = 0; x < width; x++) { dst[x] = av_clip_pixel((src[x] + offset) >> shift); + } dst += dststride; src += srcstride; } @@ -1077,8 +1073,9 @@ static void FUNC(put_weighted_pred_avg)(uint8_t *_dst, ptrdiff_t _dststride, #endif for (y = 0; y < height; y++) { - for (x = 0; x < width; x++) + for (x = 0; x < width; x++) { dst[x] = av_clip_pixel((src1[x] + src2[x] + offset) >> shift); + } dst += dststride; src1 += srcstride; src2 += srcstride; @@ -1090,7 +1087,12 @@ static void FUNC(weighted_pred)(uint8_t denom, int16_t wlxFlag, int16_t olxFlag, int16_t *src, ptrdiff_t srcstride, int width, int height) { - int shift, log2Wd, wx, ox, x, y, offset; + int shift; + int log2Wd; + int wx; + int ox; + int x , y; + int offset; pixel *dst = (pixel *)_dst; ptrdiff_t dststride = _dststride / sizeof(pixel); @@ -1121,7 +1123,13 @@ static void FUNC(weighted_pred_avg)(uint8_t denom, ptrdiff_t srcstride, int width, int height) { - int shift, log2Wd, w0, w1, o0, o1, x, y; + int shift; + int log2Wd; + int w0; + int w1; + int o0; + int o1; + int x , y; pixel *dst = (pixel *)_dst; ptrdiff_t dststride = _dststride / sizeof(pixel); @@ -1129,13 +1137,14 @@ static void FUNC(weighted_pred_avg)(uint8_t denom, log2Wd = denom + shift; w0 = wl0Flag; w1 = wl1Flag; - o0 = ol0Flag * (1 << (BIT_DEPTH - 8)); - o1 = ol1Flag * (1 << (BIT_DEPTH - 8)); + o0 = (ol0Flag) * (1 << (BIT_DEPTH - 8)); + o1 = (ol1Flag) * (1 << (BIT_DEPTH - 8)); for (y = 0; y < height; y++) { - for (x = 0; x < width; x++) + for (x = 0; x < width; x++) { dst[x] = av_clip_pixel((src1[x] * w0 + src2[x] * w1 + ((o0 + o1 + 1) << log2Wd)) >> (log2Wd + 1)); + } dst += dststride; src1 += srcstride; src2 += srcstride; @@ -1146,9 +1155,9 @@ static void FUNC(weighted_pred_avg)(uint8_t denom, #define P3 pix[-4 * xstride] #define P2 pix[-3 * xstride] #define P1 pix[-2 * xstride] -#define P0 pix[-1 * xstride] -#define Q0 pix[0 * xstride] -#define Q1 pix[1 * xstride] +#define P0 pix[-xstride] +#define Q0 pix[0] +#define Q1 pix[xstride] #define Q2 pix[2 * xstride] #define Q3 pix[3 * xstride] @@ -1156,9 +1165,9 @@ static void FUNC(weighted_pred_avg)(uint8_t denom, #define TP3 pix[-4 * xstride + 3 * ystride] #define TP2 pix[-3 * xstride + 3 * ystride] #define TP1 pix[-2 * xstride + 3 * ystride] -#define TP0 pix[-1 * xstride + 3 * ystride] -#define TQ0 pix[0 * xstride + 3 * ystride] -#define TQ1 pix[1 * xstride + 3 * ystride] +#define TP0 pix[-xstride+3 * ystride] +#define TQ0 pix[3 * ystride] +#define TQ1 pix[xstride+3 * ystride] #define TQ2 pix[2 * xstride + 3 * ystride] #define TQ3 pix[3 * xstride + 3 * ystride] @@ -1179,12 +1188,12 @@ static void FUNC(hevc_loop_filter_luma)(uint8_t *_pix, const int dq3 = abs(TQ2 - 2 * TQ1 + TQ0); const int d0 = dp0 + dq0; const int d3 = dp3 + dq3; - const int beta = _beta[j] << (BIT_DEPTH - 8); + int beta = _beta[j] << (BIT_DEPTH - 8); const int tc = _tc[j] << (BIT_DEPTH - 8); const int no_p = _no_p[j]; const int no_q = _no_q[j]; - if (d0 + d3 >= beta) { + if (d0 + d3 >= beta /*|| tc <= 0*/) { pix += 4 * ystride; continue; } else { @@ -1261,7 +1270,8 @@ static void FUNC(hevc_loop_filter_chroma)(uint8_t *_pix, ptrdiff_t _xstride, ptrdiff_t _ystride, int *_tc, uint8_t *_no_p, uint8_t *_no_q) { - int d, j, no_p, no_q; + int d, j; + int no_p, no_q; pixel *pix = (pixel *)_pix; ptrdiff_t xstride = _xstride / sizeof(pixel); ptrdiff_t ystride = _ystride / sizeof(pixel); |