Parent Directory
| 
Revision Log
Revision 1.4 - (view) (download)
| 1 : | edgomez | 1.2 | /***************************************************************************** |
| 2 : | * | ||
| 3 : | * XVID MPEG-4 VIDEO CODEC | ||
| 4 : | * Reduced-Resolution utilities | ||
| 5 : | * | ||
| 6 : | * Copyright(C) 2002 Pascal Massimino <skal@planet-d.net> | ||
| 7 : | * | ||
| 8 : | * XviD is free software; you can redistribute it and/or modify it | ||
| 9 : | * under the terms of the GNU General Public License as published by | ||
| 10 : | * the Free Software Foundation; either version 2 of the License, or | ||
| 11 : | * (at your option) any later version. | ||
| 12 : | * | ||
| 13 : | * This program is distributed in the hope that it will be useful, | ||
| 14 : | * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
| 15 : | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | ||
| 16 : | * GNU General Public License for more details. | ||
| 17 : | * | ||
| 18 : | * You should have received a copy of the GNU General Public License | ||
| 19 : | * along with this program; if not, write to the Free Software | ||
| 20 : | * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA | ||
| 21 : | * | ||
| 22 : | edgomez | 1.3 | * $Id$ |
| 23 : | edgomez | 1.2 | * |
| 24 : | ****************************************************************************/ | ||
| 25 : | |||
| 26 : | #include "../portab.h" | ||
| 27 : | #include "../global.h" | ||
| 28 : | #include "reduced.h" | ||
| 29 : | |||
| 30 : | edgomez | 1.3 | /* function pointers */ |
| 31 : | edgomez | 1.2 | COPY_UPSAMPLED_8X8_16TO8 * copy_upsampled_8x8_16to8; |
| 32 : | ADD_UPSAMPLED_8X8_16TO8 * add_upsampled_8x8_16to8; | ||
| 33 : | VFILTER_31 * vfilter_31; | ||
| 34 : | HFILTER_31 * hfilter_31; | ||
| 35 : | FILTER_18X18_TO_8X8 * filter_18x18_to_8x8; | ||
| 36 : | FILTER_DIFF_18X18_TO_8X8 * filter_diff_18x18_to_8x8; | ||
| 37 : | |||
| 38 : | edgomez | 1.3 | /*---------------------------------------------------------------------------- |
| 39 : | * Upsampling (1/3/3/1) filter | ||
| 40 : | *--------------------------------------------------------------------------*/ | ||
| 41 : | edgomez | 1.2 | |
| 42 : | #define ADD(dst,src) (dst) = CLIP((dst)+(src), 0, 255) | ||
| 43 : | |||
| 44 : | static __inline void Filter_31(uint8_t *Dst1, uint8_t *Dst2, | ||
| 45 : | const int16_t *Src1, const int16_t *Src2) | ||
| 46 : | { | ||
| 47 : | /* Src[] is assumed to be >=0. So we can use ">>2" instead of "/2" */ | ||
| 48 : | int16_t a = (3*Src1[0]+ Src2[0]+2) >> 2; | ||
| 49 : | int16_t b = ( Src1[0]+3*Src2[0]+2) >> 2; | ||
| 50 : | Dst1[0] = CLIP(a, 0, 255); | ||
| 51 : | Dst2[0] = CLIP(b, 0, 255); | ||
| 52 : | } | ||
| 53 : | |||
| 54 : | static __inline void Filter_9331(uint8_t *Dst1, uint8_t *Dst2, | ||
| 55 : | const int16_t *Src1, const int16_t *Src2) | ||
| 56 : | { | ||
| 57 : | /* Src[] is assumed to be >=0. So we can use ">>4" instead of "/16" */ | ||
| 58 : | int16_t a = (9*Src1[0]+ 3*Src1[1]+ 3*Src2[0] + 1*Src2[1] + 8) >> 4; | ||
| 59 : | int16_t b = (3*Src1[0]+ 9*Src1[1]+ 1*Src2[0] + 3*Src2[1] + 8) >> 4; | ||
| 60 : | int16_t c = (3*Src1[0]+ 1*Src1[1]+ 9*Src2[0] + 3*Src2[1] + 8) >> 4; | ||
| 61 : | int16_t d = (1*Src1[0]+ 3*Src1[1]+ 3*Src2[0] + 9*Src2[1] + 8) >> 4; | ||
| 62 : | Dst1[0] = CLIP(a, 0, 255); | ||
| 63 : | Dst1[1] = CLIP(b, 0, 255); | ||
| 64 : | Dst2[0] = CLIP(c, 0, 255); | ||
| 65 : | Dst2[1] = CLIP(d, 0, 255); | ||
| 66 : | } | ||
| 67 : | |||
| 68 : | void xvid_Copy_Upsampled_8x8_16To8_C(uint8_t *Dst, const int16_t *Src, const int BpS) | ||
| 69 : | { | ||
| 70 : | int x, y; | ||
| 71 : | |||
| 72 : | Dst[0] = CLIP(Src[0], 0, 255); | ||
| 73 : | for(x=0; x<7; ++x) Filter_31(Dst+2*x+1, Dst+2*x+2, Src+x, Src+x+1); | ||
| 74 : | Dst[15] = CLIP(Src[7], 0, 255); | ||
| 75 : | Dst += BpS; | ||
| 76 : | for(y=0; y<7; ++y) { | ||
| 77 : | uint8_t *const Dst2 = Dst + BpS; | ||
| 78 : | Filter_31(Dst, Dst2, Src, Src+8); | ||
| 79 : | for(x=0; x<7; ++x) | ||
| 80 : | Filter_9331(Dst+2*x+1, Dst2+2*x+1, Src+x, Src+x+8); | ||
| 81 : | Filter_31(Dst+15, Dst2+15, Src+7, Src+7+8); | ||
| 82 : | edgomez | 1.3 | Src += 8; |
| 83 : | edgomez | 1.2 | Dst += 2*BpS; |
| 84 : | } | ||
| 85 : | Dst[0] = CLIP(Src[0], 0, 255); | ||
| 86 : | for(x=0; x<7; ++x) Filter_31(Dst+2*x+1, Dst+2*x+2, Src+x, Src+x+1); | ||
| 87 : | Dst[15] = CLIP(Src[7], 0, 255); | ||
| 88 : | } | ||
| 89 : | |||
| 90 : | static __inline void Filter_Add_31(uint8_t *Dst1, uint8_t *Dst2, | ||
| 91 : | const int16_t *Src1, const int16_t *Src2) | ||
| 92 : | { | ||
| 93 : | /* Here, we must use "/4", since Src[] is in [-256, 255] */ | ||
| 94 : | int16_t a = (3*Src1[0]+ Src2[0] + 2) / 4; | ||
| 95 : | int16_t b = ( Src1[0]+3*Src2[0] + 2) / 4; | ||
| 96 : | ADD(Dst1[0], a); | ||
| 97 : | ADD(Dst2[0], b); | ||
| 98 : | } | ||
| 99 : | |||
| 100 : | static __inline void Filter_Add_9331(uint8_t *Dst1, uint8_t *Dst2, | ||
| 101 : | const int16_t *Src1, const int16_t *Src2) | ||
| 102 : | { | ||
| 103 : | int16_t a = (9*Src1[0]+ 3*Src1[1]+ 3*Src2[0] + 1*Src2[1] + 8) / 16; | ||
| 104 : | int16_t b = (3*Src1[0]+ 9*Src1[1]+ 1*Src2[0] + 3*Src2[1] + 8) / 16; | ||
| 105 : | int16_t c = (3*Src1[0]+ 1*Src1[1]+ 9*Src2[0] + 3*Src2[1] + 8) / 16; | ||
| 106 : | int16_t d = (1*Src1[0]+ 3*Src1[1]+ 3*Src2[0] + 9*Src2[1] + 8) / 16; | ||
| 107 : | ADD(Dst1[0], a); | ||
| 108 : | ADD(Dst1[1], b); | ||
| 109 : | ADD(Dst2[0], c); | ||
| 110 : | ADD(Dst2[1], d); | ||
| 111 : | } | ||
| 112 : | |||
| 113 : | void xvid_Add_Upsampled_8x8_16To8_C(uint8_t *Dst, const int16_t *Src, const int BpS) | ||
| 114 : | { | ||
| 115 : | int x, y; | ||
| 116 : | |||
| 117 : | ADD(Dst[0], Src[0]); | ||
| 118 : | for(x=0; x<7; ++x) Filter_Add_31(Dst+2*x+1, Dst+2*x+2, Src+x, Src+x+1); | ||
| 119 : | ADD(Dst[15], Src[7]); | ||
| 120 : | Dst += BpS; | ||
| 121 : | for(y=0; y<7; ++y) { | ||
| 122 : | uint8_t *const Dst2 = Dst + BpS; | ||
| 123 : | Filter_Add_31(Dst, Dst2, Src, Src+8); | ||
| 124 : | for(x=0; x<7; ++x) | ||
| 125 : | Filter_Add_9331(Dst+2*x+1, Dst2+2*x+1, Src+x, Src+x+8); | ||
| 126 : | Filter_Add_31(Dst+15, Dst2+15, Src+7, Src+7+8); | ||
| 127 : | edgomez | 1.3 | Src += 8; |
| 128 : | edgomez | 1.2 | Dst += 2*BpS; |
| 129 : | } | ||
| 130 : | ADD(Dst[0], Src[0]); | ||
| 131 : | for(x=0; x<7; ++x) Filter_Add_31(Dst+2*x+1, Dst+2*x+2, Src+x, Src+x+1); | ||
| 132 : | ADD(Dst[15], Src[7]); | ||
| 133 : | } | ||
| 134 : | #undef ADD | ||
| 135 : | |||
| 136 : | edgomez | 1.3 | /*---------------------------------------------------------------------------- |
| 137 : | * horizontal and vertical deblocking | ||
| 138 : | *--------------------------------------------------------------------------*/ | ||
| 139 : | edgomez | 1.2 | |
| 140 : | void xvid_HFilter_31_C(uint8_t *Src1, uint8_t *Src2, int Nb_Blks) | ||
| 141 : | { | ||
| 142 : | Nb_Blks *= 8; | ||
| 143 : | while(Nb_Blks-->0) { | ||
| 144 : | uint8_t a = ( 3*Src1[0] + 1*Src2[0] + 2 ) >> 2; | ||
| 145 : | uint8_t b = ( 1*Src1[0] + 3*Src2[0] + 2 ) >> 2; | ||
| 146 : | *Src1++ = a; | ||
| 147 : | *Src2++ = b; | ||
| 148 : | } | ||
| 149 : | } | ||
| 150 : | |||
| 151 : | void xvid_VFilter_31_C(uint8_t *Src1, uint8_t *Src2, const int BpS, int Nb_Blks) | ||
| 152 : | { | ||
| 153 : | Nb_Blks *= 8; | ||
| 154 : | while(Nb_Blks-->0) { | ||
| 155 : | uint8_t a = ( 3*Src1[0] + 1*Src2[0] + 2 ) >> 2; | ||
| 156 : | uint8_t b = ( 1*Src1[0] + 3*Src2[0] + 2 ) >> 2; | ||
| 157 : | *Src1 = a; | ||
| 158 : | *Src2 = b; | ||
| 159 : | Src1 += BpS; | ||
| 160 : | Src2 += BpS; | ||
| 161 : | } | ||
| 162 : | } | ||
| 163 : | |||
| 164 : | edgomez | 1.3 | /*---------------------------------------------------------------------------- |
| 165 : | * 16x16 -> 8x8 (1/3/3/1) downsampling | ||
| 166 : | * | ||
| 167 : | * Warning! These read 1 pixel outside of the input 16x16 block! | ||
| 168 : | *--------------------------------------------------------------------------*/ | ||
| 169 : | edgomez | 1.2 | |
| 170 : | void xvid_Filter_18x18_To_8x8_C(int16_t *Dst, const uint8_t *Src, const int BpS) | ||
| 171 : | { | ||
| 172 : | int16_t *T, Tmp[18*8]; | ||
| 173 : | int i, j; | ||
| 174 : | |||
| 175 : | T = Tmp; | ||
| 176 : | Src -= BpS; | ||
| 177 : | for(j=-1; j<17; j++) { | ||
| 178 : | for(i=0; i<8; ++i) | ||
| 179 : | T[i] = Src[2*i-1] + 3*Src[2*i+0] + 3*Src[2*i+1] + Src[2*i+2]; | ||
| 180 : | T += 8; | ||
| 181 : | Src += BpS; | ||
| 182 : | } | ||
| 183 : | T = Tmp + 8; | ||
| 184 : | for(j=0; j<8; j++) { | ||
| 185 : | for(i=0; i<8; ++i) | ||
| 186 : | Dst[i] = ( T[-8+i] + 3*T[0+i] + 3*T[8+i] + T[16+i] + 32 ) / 64; | ||
| 187 : | Dst += 8; | ||
| 188 : | T += 16; | ||
| 189 : | } | ||
| 190 : | } | ||
| 191 : | |||
| 192 : | void xvid_Filter_Diff_18x18_To_8x8_C(int16_t *Dst, const uint8_t *Src, const int BpS) | ||
| 193 : | { | ||
| 194 : | int16_t *T, Tmp[18*8]; | ||
| 195 : | int i, j; | ||
| 196 : | |||
| 197 : | T = Tmp; | ||
| 198 : | Src -= BpS; | ||
| 199 : | for(j=-1; j<17; j++) { | ||
| 200 : | for(i=0; i<8; ++i) | ||
| 201 : | T[i] = Src[2*i-1] + 3*Src[2*i+0] + 3*Src[2*i+1] + Src[2*i+2]; | ||
| 202 : | T += 8; | ||
| 203 : | Src += BpS; | ||
| 204 : | } | ||
| 205 : | T = Tmp; | ||
| 206 : | for(j=0; j<8; j++) { | ||
| 207 : | for(i=0; i<8; ++i) | ||
| 208 : | Dst[i] -= ( T[i] + 3*T[8+i] + 3*T[16+i] + T[24+i] + 32 ) / 64; | ||
| 209 : | Dst += 8; | ||
| 210 : | T += 16; | ||
| 211 : | } | ||
| 212 : | } |
| No admin address has been configured | ViewVC Help |
| Powered by ViewVC 1.0.4 |