| 1 |
/* |
| 2 |
* All Video Processing kernels |
| 3 |
* Copyright © <2010>, Intel Corporation. |
| 4 |
* |
| 5 |
* Permission is hereby granted, free of charge, to any person obtaining a |
| 6 |
* copy of this software and associated documentation files (the |
| 7 |
* "Software"), to deal in the Software without restriction, including |
| 8 |
* without limitation the rights to use, copy, modify, merge, publish, |
| 9 |
* distribute, sub license, and/or sell copies of the Software, and to |
| 10 |
* permit persons to whom the Software is furnished to do so, subject to |
| 11 |
* the following conditions: |
| 12 |
* |
| 13 |
* The above copyright notice and this permission notice (including the |
| 14 |
* next paragraph) shall be included in all copies or substantial portions |
| 15 |
* of the Software. |
| 16 |
* |
| 17 |
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS |
| 18 |
* OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF |
| 19 |
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. |
| 20 |
* IN NO EVENT SHALL PRECISION INSIGHT AND/OR ITS SUPPLIERS BE LIABLE FOR |
| 21 |
* ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, |
| 22 |
* TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE |
| 23 |
* SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. |
| 24 |
* |
| 25 |
* This file was originally licensed under the following license |
| 26 |
* |
| 27 |
* Licensed under the Apache License, Version 2.0 (the "License"); |
| 28 |
* you may not use this file except in compliance with the License. |
| 29 |
* You may obtain a copy of the License at |
| 30 |
* |
| 31 |
* http://www.apache.org/licenses/LICENSE-2.0 |
| 32 |
* |
| 33 |
* Unless required by applicable law or agreed to in writing, software |
| 34 |
* distributed under the License is distributed on an "AS IS" BASIS, |
| 35 |
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 36 |
* See the License for the specific language governing permissions and |
| 37 |
* limitations under the License. |
| 38 |
* |
| 39 |
* Authors: |
| 40 |
* Halley Zhao <[email protected]> |
| 41 |
*/ |
| 42 |
|
| 43 |
// Module name: YUVX_Save_RGBX_Fix.asm |
| 44 |
//---------------------------------------------------------------- |
| 45 |
|
| 46 |
#include "RGBX_Load_16x8.inc" |
| 47 |
|
| 48 |
#if (0) |
| 49 |
#define nTEMP0 34 // transformation coefficient |
| 50 |
#define nTEMP1 35 // one row of R (first half register is used) |
| 51 |
#define nTEMP2 36 // one row of G (first half register is used) |
| 52 |
#define nTEMP3 37 // one row of B (first half register is used) |
| 53 |
#define nTEMP4 38 // mul and add |
| 54 |
#define nTEMP5 39 // mul and add |
| 55 |
#define nTEMP6 40 // mul and add |
| 56 |
#define nTEMP7 41 // mul and add |
| 57 |
#define nTEMP8 42 // sum of mul |
| 58 |
#define nTEMP10 44 |
| 59 |
#define nTEMP10 44 // split ub pixel to word width 1st quarter |
| 60 |
#define nTEMP12 46 // split ub pixel to word width 2nd quarter |
| 61 |
#define nTEMP14 48 // split ub pixel to word width 3rd quarter |
| 62 |
#define nTEMP16 50 // split ub pixel to word width 4th quarter |
| 63 |
#define nTEMP17 51 |
| 64 |
#define nTEMP18 52 |
| 65 |
|
| 66 |
#define nTEMP24 58 // temp using for repeat U/V in NV12_Load_8x4.asm |
| 67 |
#endif |
| 68 |
|
| 69 |
#define ONE_ROW_DEBUG 0 |
| 70 |
|
| 71 |
#if (ONE_ROW_DEBUG) |
| 72 |
#define ROW_NUM 0 |
| 73 |
#define DBG_ROWNUM_BASE 1 |
| 74 |
CHANNEL_2 2 |
| 75 |
#else |
| 76 |
#define ROW_NUM %1 |
| 77 |
$for(0; <nY_NUM_OF_ROWS; 1) { |
| 78 |
#endif |
| 79 |
// C = Y' - 16 D = U - 128 E = V - 128 |
| 80 |
add (16) REG2(r,nTEMP10,0)<1>:w ubDEST_RGBX(0,ROW_NUM*64 )<0;16,1> bYUV_OFF<0;4,1>:b |
| 81 |
add (16) REG2(r,nTEMP12,0)<1>:w ubDEST_RGBX(0,ROW_NUM*64+16)<0;16,1> bYUV_OFF<0;4,1>:b |
| 82 |
add (16) REG2(r,nTEMP14,0)<1>:w ubDEST_RGBX(0,ROW_NUM*64+32)<0;16,1> bYUV_OFF<0;4,1>:b |
| 83 |
add (16) REG2(r,nTEMP16,0)<1>:w ubDEST_RGBX(0,ROW_NUM*64+48)<0;16,1> bYUV_OFF<0;4,1>:b |
| 84 |
|
| 85 |
#if (ONE_ROW_DEBUG) |
| 86 |
mov (16) ubDEST_RGBX(0,(DBG_ROWNUM_BASE)*64 )<1> REG2(r,nTEMP10, 0)<0;16,2>:ub |
| 87 |
mov (16) ubDEST_RGBX(0,(DBG_ROWNUM_BASE)*64+16)<1> REG2(r,nTEMP12, 0)<0;16,2>:ub |
| 88 |
mov (16) ubDEST_RGBX(0,(DBG_ROWNUM_BASE)*64+32)<1> REG2(r,nTEMP14, 0)<0;16,2>:ub |
| 89 |
mov (16) ubDEST_RGBX(0,(DBG_ROWNUM_BASE)*64+48)<1> REG2(r,nTEMP16, 0)<0;16,2>:ub |
| 90 |
#endif |
| 91 |
|
| 92 |
// |Y|U|V|X|==>|R|G|B|X| |
| 93 |
// ###### do one row for R |
| 94 |
// #### mul and add |
| 95 |
mul.sat (16) REG2(r, nTEMP4, 0)<1>:w REG2(r,nTEMP10,0)<0;16,1>:w wYUV_to_RGB_CH2_Coef_Fix<0;4,1>:w |
| 96 |
mul.sat (16) REG2(r, nTEMP5, 0)<1>:w REG2(r,nTEMP12,0)<0;16,1>:w wYUV_to_RGB_CH2_Coef_Fix<0;4,1>:w |
| 97 |
mul.sat (16) REG2(r, nTEMP6, 0)<1>:w REG2(r,nTEMP14,0)<0;16,1>:w wYUV_to_RGB_CH2_Coef_Fix<0;4,1>:w |
| 98 |
mul.sat (16) REG2(r, nTEMP7, 0)<1>:w REG2(r,nTEMP16,0)<0;16,1>:w wYUV_to_RGB_CH2_Coef_Fix<0;4,1>:w |
| 99 |
|
| 100 |
#if (ONE_ROW_DEBUG) |
| 101 |
mov (8) ubDEST_RGBX(0,(DBG_ROWNUM_BASE+1)*64+CHANNEL_2 )<4> bYUV_to_RGB_CH2_Coef_Fix<0;8,1>:ub |
| 102 |
mov (8) ubDEST_RGBX(0,(DBG_ROWNUM_BASE+1)*64+CHANNEL_2+32)<4> bYUV_to_RGB_CH2_Coef_Fix<0;8,1>:ub |
| 103 |
|
| 104 |
mov (8) ubDEST_RGBX(0,(DBG_ROWNUM_BASE+2)*64+CHANNEL_2 )<4> REG2(r,nTEMP4, 0)<0;8,1>:ub |
| 105 |
mov (8) ubDEST_RGBX(0,(DBG_ROWNUM_BASE+2)*64+CHANNEL_2+32)<4> REG2(r,nTEMP4, 8)<0;8,1>:ub |
| 106 |
#endif |
| 107 |
|
| 108 |
add.sat (4) REG2(r, nTEMP4, 0)<4>:uw REG2(r, nTEMP4, 0)<0;4,4>:w REG2(r, nTEMP4, 1)<0;4,4>:w |
| 109 |
add.sat (4) REG2(r, nTEMP5, 0)<4>:uw REG2(r, nTEMP5, 0)<0;4,4>:w REG2(r, nTEMP5, 1)<0;4,4>:w |
| 110 |
add.sat (4) REG2(r, nTEMP6, 0)<4>:uw REG2(r, nTEMP6, 0)<0;4,4>:w REG2(r, nTEMP6, 1)<0;4,4>:w |
| 111 |
add.sat (4) REG2(r, nTEMP7, 0)<4>:uw REG2(r, nTEMP7, 0)<0;4,4>:w REG2(r, nTEMP7, 1)<0;4,4>:w |
| 112 |
add.sat (4) REG2(r, nTEMP4, 0)<4>:uw REG2(r, nTEMP4, 0)<0;4,4>:uw REG2(r, nTEMP4, 2)<0;4,4>:w |
| 113 |
add.sat (4) REG2(r, nTEMP5, 0)<4>:uw REG2(r, nTEMP5, 0)<0;4,4>:uw REG2(r, nTEMP5, 2)<0;4,4>:w |
| 114 |
add.sat (4) REG2(r, nTEMP6, 0)<4>:uw REG2(r, nTEMP6, 0)<0;4,4>:uw REG2(r, nTEMP6, 2)<0;4,4>:w |
| 115 |
add.sat (4) REG2(r, nTEMP7, 0)<4>:uw REG2(r, nTEMP7, 0)<0;4,4>:uw REG2(r, nTEMP7, 2)<0;4,4>:w |
| 116 |
|
| 117 |
// #### write one row of R to rnTEMP1 |
| 118 |
mov (4) REG2(r, nTEMP8, 0)<1>:uw REG2(r, nTEMP4, 0)<0; 4, 4>:uw |
| 119 |
mov (4) REG2(r, nTEMP8, 4)<1>:uw REG2(r, nTEMP5, 0)<0; 4, 4>:uw |
| 120 |
mov (4) REG2(r, nTEMP8, 8)<1>:uw REG2(r, nTEMP6, 0)<0; 4, 4>:uw |
| 121 |
mov (4) REG2(r, nTEMP8, 12)<1>:uw REG2(r, nTEMP7, 0)<0; 4, 4>:uw |
| 122 |
|
| 123 |
#if (ONE_ROW_DEBUG) |
| 124 |
mov (8) ubDEST_RGBX(0,(DBG_ROWNUM_BASE+3)*64+CHANNEL_2 )<4> REG2(r,nTEMP8, 0)<0;8,1>:ub |
| 125 |
mov (8) ubDEST_RGBX(0,(DBG_ROWNUM_BASE+3)*64+CHANNEL_2+32)<4> REG2(r,nTEMP8, 8)<0;8,1>:ub |
| 126 |
#endif |
| 127 |
|
| 128 |
add.sat (16) REG2(r, nTEMP8, 0)<1>:uw REG2(r, nTEMP8, 0)<0; 16, 1>:uw 0x80:uw |
| 129 |
shl.sat (16) REG2(r, nTEMP8, 0)<1>:uw REG2(r, nTEMP8, 0)<0; 16, 1>:uw 1:w |
| 130 |
mov (16) REG2(r, nTEMP1, 0)<1>:ub REG2(r, nTEMP8, 1)<0; 16, 2>:ub |
| 131 |
|
| 132 |
#if (ONE_ROW_DEBUG) |
| 133 |
mov (8) ubDEST_RGBX(0,(DBG_ROWNUM_BASE+4)*64+CHANNEL_2 )<4> REG2(r,nTEMP8, 0)<0;8,1>:ub |
| 134 |
mov (8) ubDEST_RGBX(0,(DBG_ROWNUM_BASE+4)*64+CHANNEL_2+32)<4> REG2(r,nTEMP8, 8)<0;8,1>:ub |
| 135 |
#endif |
| 136 |
// ###### do one row for G |
| 137 |
// #### mul and add |
| 138 |
mul.sat (16) REG2(r, nTEMP4, 0)<1>:w REG2(r,nTEMP10,0)<0;16,1>:w wYUV_to_RGB_CH1_Coef_Fix<0;4,1>:w |
| 139 |
mul.sat (16) REG2(r, nTEMP5, 0)<1>:w REG2(r,nTEMP12,0)<0;16,1>:w wYUV_to_RGB_CH1_Coef_Fix<0;4,1>:w |
| 140 |
mul.sat (16) REG2(r, nTEMP6, 0)<1>:w REG2(r,nTEMP14,0)<0;16,1>:w wYUV_to_RGB_CH1_Coef_Fix<0;4,1>:w |
| 141 |
mul.sat (16) REG2(r, nTEMP7, 0)<1>:w REG2(r,nTEMP16,0)<0;16,1>:w wYUV_to_RGB_CH1_Coef_Fix<0;4,1>:w |
| 142 |
|
| 143 |
add.sat (4) REG2(r, nTEMP4, 0)<4>:uw REG2(r, nTEMP4, 0)<0;4,4>:w REG2(r, nTEMP4, 1)<0;4,4>:w |
| 144 |
add.sat (4) REG2(r, nTEMP5, 0)<4>:uw REG2(r, nTEMP5, 0)<0;4,4>:w REG2(r, nTEMP5, 1)<0;4,4>:w |
| 145 |
add.sat (4) REG2(r, nTEMP6, 0)<4>:uw REG2(r, nTEMP6, 0)<0;4,4>:w REG2(r, nTEMP6, 1)<0;4,4>:w |
| 146 |
add.sat (4) REG2(r, nTEMP7, 0)<4>:uw REG2(r, nTEMP7, 0)<0;4,4>:w REG2(r, nTEMP7, 1)<0;4,4>:w |
| 147 |
add.sat (4) REG2(r, nTEMP4, 0)<4>:uw REG2(r, nTEMP4, 0)<0;4,4>:uw REG2(r, nTEMP4, 2)<0;4,4>:w |
| 148 |
add.sat (4) REG2(r, nTEMP5, 0)<4>:uw REG2(r, nTEMP5, 0)<0;4,4>:uw REG2(r, nTEMP5, 2)<0;4,4>:w |
| 149 |
add.sat (4) REG2(r, nTEMP6, 0)<4>:uw REG2(r, nTEMP6, 0)<0;4,4>:uw REG2(r, nTEMP6, 2)<0;4,4>:w |
| 150 |
add.sat (4) REG2(r, nTEMP7, 0)<4>:uw REG2(r, nTEMP7, 0)<0;4,4>:uw REG2(r, nTEMP7, 2)<0;4,4>:w |
| 151 |
|
| 152 |
// #### write one row of G to rnTEMP2 |
| 153 |
mov (4) REG2(r, nTEMP8, 0)<1>:uw REG2(r, nTEMP4, 0)<0; 4, 4>:uw |
| 154 |
mov (4) REG2(r, nTEMP8, 4)<1>:uw REG2(r, nTEMP5, 0)<0; 4, 4>:uw |
| 155 |
mov (4) REG2(r, nTEMP8, 8)<1>:uw REG2(r, nTEMP6, 0)<0; 4, 4>:uw |
| 156 |
mov (4) REG2(r, nTEMP8, 12)<1>:uw REG2(r, nTEMP7, 0)<0; 4, 4>:uw |
| 157 |
|
| 158 |
add (16) REG2(r, nTEMP8, 0)<1>:uw REG2(r, nTEMP8, 0)<0; 16, 1>:uw 0x80:uw // saturation |
| 159 |
shl.sat (16) REG2(r, nTEMP8, 0)<1>:uw REG2(r, nTEMP8, 0)<0; 16, 1>:uw 1:w |
| 160 |
mov (16) REG2(r, nTEMP2, 0)<1>:ub REG2(r, nTEMP8, 1)<0; 16, 2>:ub |
| 161 |
|
| 162 |
// ###### do one row for B |
| 163 |
// #### mul and add |
| 164 |
mul.sat (16) REG2(r, nTEMP4, 0)<1>:w REG2(r,nTEMP10,0)<0;16,1>:w wYUV_to_RGB_CH0_Coef_Fix<0;4,1>:w |
| 165 |
mul.sat (16) REG2(r, nTEMP5, 0)<1>:w REG2(r,nTEMP12,0)<0;16,1>:w wYUV_to_RGB_CH0_Coef_Fix<0;4,1>:w |
| 166 |
mul.sat (16) REG2(r, nTEMP6, 0)<1>:w REG2(r,nTEMP14,0)<0;16,1>:w wYUV_to_RGB_CH0_Coef_Fix<0;4,1>:w |
| 167 |
mul.sat (16) REG2(r, nTEMP7, 0)<1>:w REG2(r,nTEMP16,0)<0;16,1>:w wYUV_to_RGB_CH0_Coef_Fix<0;4,1>:w |
| 168 |
|
| 169 |
// I had reduced the following add because U coef is zero for B; but in order to support BGR/RGB at the same time, I have to add it back. |
| 170 |
add.sat (4) REG2(r, nTEMP4, 0)<4>:uw REG2(r, nTEMP4, 0)<0;4,4>:w REG2(r, nTEMP4, 1)<0;4,4>:w |
| 171 |
add.sat (4) REG2(r, nTEMP5, 0)<4>:uw REG2(r, nTEMP5, 0)<0;4,4>:w REG2(r, nTEMP5, 1)<0;4,4>:w |
| 172 |
add.sat (4) REG2(r, nTEMP6, 0)<4>:uw REG2(r, nTEMP6, 0)<0;4,4>:w REG2(r, nTEMP6, 1)<0;4,4>:w |
| 173 |
add.sat (4) REG2(r, nTEMP7, 0)<4>:uw REG2(r, nTEMP7, 0)<0;4,4>:w REG2(r, nTEMP7, 1)<0;4,4>:w |
| 174 |
add.sat (4) REG2(r, nTEMP4, 0)<4>:uw REG2(r, nTEMP4, 0)<0;4,4>:uw REG2(r, nTEMP4, 2)<0;4,4>:w |
| 175 |
add.sat (4) REG2(r, nTEMP5, 0)<4>:uw REG2(r, nTEMP5, 0)<0;4,4>:uw REG2(r, nTEMP5, 2)<0;4,4>:w |
| 176 |
add.sat (4) REG2(r, nTEMP6, 0)<4>:uw REG2(r, nTEMP6, 0)<0;4,4>:uw REG2(r, nTEMP6, 2)<0;4,4>:w |
| 177 |
add.sat (4) REG2(r, nTEMP7, 0)<4>:uw REG2(r, nTEMP7, 0)<0;4,4>:uw REG2(r, nTEMP7, 2)<0;4,4>:w |
| 178 |
|
| 179 |
// #### write one row of B to rnTEMP3 |
| 180 |
mov (4) REG2(r, nTEMP8, 0)<1>:uw REG2(r, nTEMP4, 0)<0; 4, 4>:uw |
| 181 |
mov (4) REG2(r, nTEMP8, 4)<1>:uw REG2(r, nTEMP5, 0)<0; 4, 4>:uw |
| 182 |
mov (4) REG2(r, nTEMP8, 8)<1>:uw REG2(r, nTEMP6, 0)<0; 4, 4>:uw |
| 183 |
mov (4) REG2(r, nTEMP8, 12)<1>:uw REG2(r, nTEMP7, 0)<0; 4, 4>:uw |
| 184 |
|
| 185 |
add.sat (16) REG2(r, nTEMP8, 0)<1>:uw REG2(r, nTEMP8, 0)<0; 16, 1>:uw 0x80:uw // saturation |
| 186 |
shl.sat (16) REG2(r, nTEMP8, 0)<1>:uw REG2(r, nTEMP8, 0)<0; 16, 1>:uw 1:w |
| 187 |
mov (16) REG2(r, nTEMP3, 0)<1>:ub REG2(r, nTEMP8, 1)<0; 16, 2>:ub |
| 188 |
|
| 189 |
// B |
| 190 |
mov (8) ubDEST_RGBX(0,ROW_NUM*64 )<4> REG2(r,nTEMP3, 0)<0;8,1>:ub |
| 191 |
mov (8) ubDEST_RGBX(0,ROW_NUM*64+32)<4> REG2(r,nTEMP3, 8)<0;8,1>:ub |
| 192 |
// G |
| 193 |
mov (8) ubDEST_RGBX(0,ROW_NUM*64+1 )<4> REG2(r,nTEMP2, 0)<0;8,1>:ub |
| 194 |
mov (8) ubDEST_RGBX(0,ROW_NUM*64+1+32)<4> REG2(r,nTEMP2, 8)<0;8,1>:ub |
| 195 |
// R |
| 196 |
mov (8) ubDEST_RGBX(0,ROW_NUM*64+2 )<4> REG2(r,nTEMP1, 0)<0;8,1>:ub |
| 197 |
mov (8) ubDEST_RGBX(0,ROW_NUM*64+2+32)<4> REG2(r,nTEMP1, 8)<0;8,1>:ub |
| 198 |
#if (!ONE_ROW_DEBUG) |
| 199 |
} |
| 200 |
#endif |