1 /******************************************************************************
2 *
3 * Copyright (C) 2012 Ittiam Systems Pvt Ltd, Bangalore
4 *
5 * Licensed under the Apache License, Version 2.0 (the "License");
6 * you may not use this file except in compliance with the License.
7 * You may obtain a copy of the License at:
8 *
9 * http://www.apache.org/licenses/LICENSE-2.0
10 *
11 * Unless required by applicable law or agreed to in writing, software
12 * distributed under the License is distributed on an "AS IS" BASIS,
13 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
14 * See the License for the specific language governing permissions and
15 * limitations under the License.
16 *
17 ******************************************************************************/
18 /**
19  *******************************************************************************
20  * @file
21  *  ihevc_chroma_itrans_recon.c
22  *
23  * @brief
24  *  Contains function definitions for inverse transform  and reconstruction
25  * of chroma interleaved data.
26  *
27  * @author
28  *  100470
29  *
30  * @par List of Functions:
31  *  - ihevc_chroma_itrans_recon_4x4()
32  *
33  * @remarks
34  *  None
35  *
36  *******************************************************************************
37  */
38 
39 #include <stdio.h>
40 #include <string.h>
41 #include "ihevc_typedefs.h"
42 #include "ihevc_macros.h"
43 #include "ihevc_platform_macros.h"
44 #include "ihevc_defs.h"
45 #include "ihevc_trans_tables.h"
46 #include "ihevc_chroma_itrans_recon.h"
47 #include "ihevc_func_selector.h"
48 #include "ihevc_trans_macros.h"
49 
50 /* All the functions work one component(U or V) of interleaved data depending upon pointers passed to it */
51 /* Data visualization */
52 /* U V U V U V U V */
53 /* U V U V U V U V */
54 /* U V U V U V U V */
55 /* U V U V U V U V */
56 /* If the pointer points to first byte of above stream (U) , functions will operate on U component */
57 /* If the pointer points to second byte of above stream (V) , functions will operate on V component */
58 
59 /**
60  *******************************************************************************
61  *
62  * @brief
63  *  This function performs Inverse transform  and reconstruction for 4x4
64  * input block
65  *
66  * @par Description:
67  *  Performs inverse transform and adds the prediction  data and clips output
68  * to 8 bit
69  *
70  * @param[in] pi2_src
71  *  Input 4x4 coefficients
72  *
73  * @param[in] pi2_tmp
74  *  Temporary 4x4 buffer for storing inverse transform
75  *  1st stage output
76  *
77  * @param[in] pu1_pred
78  *  Prediction 4x4 block
79  *
80  * @param[out] pu1_dst
81  *  Output 4x4 block
82  *
83  * @param[in] src_strd
84  *  Input stride
85  *
86  * @param[in] pred_strd
87  *  Prediction stride
88  *
89  * @param[in] dst_strd
90  *  Output Stride
91  *
92  * @param[in] shift
93  *  Output shift
94  *
95  * @param[in] zero_cols
96  *  Zero columns in pi2_src
97  *
98  * @returns  Void
99  *
100  * @remarks
101  *  None
102  *
103  *******************************************************************************
104  */
105 
106 
ihevc_chroma_itrans_recon_4x4(WORD16 * pi2_src,WORD16 * pi2_tmp,UWORD8 * pu1_pred,UWORD8 * pu1_dst,WORD32 src_strd,WORD32 pred_strd,WORD32 dst_strd,WORD32 zero_cols,WORD32 zero_rows)107 void ihevc_chroma_itrans_recon_4x4(WORD16 *pi2_src,
108                                    WORD16 *pi2_tmp,
109                                    UWORD8 *pu1_pred,
110                                    UWORD8 *pu1_dst,
111                                    WORD32 src_strd,
112                                    WORD32 pred_strd,
113                                    WORD32 dst_strd,
114                                    WORD32 zero_cols,
115                                    WORD32 zero_rows)
116 {
117     WORD32 j;
118     WORD32 e[2], o[2];
119     WORD32 add;
120     WORD32 shift;
121     WORD16 *pi2_tmp_orig;
122     WORD32 trans_size;
123     UNUSED(zero_rows);
124     trans_size = TRANS_SIZE_4;
125 
126     pi2_tmp_orig = pi2_tmp;
127 
128     /* Inverse Transform 1st stage */
129     shift = IT_SHIFT_STAGE_1;
130     add = 1 << (shift - 1);
131 
132     for(j = 0; j < trans_size; j++)
133     {
134         /* Checking for Zero Cols */
135         if((zero_cols & 1) == 1)
136         {
137             memset(pi2_tmp, 0, trans_size * sizeof(WORD16));
138         }
139         else
140         {
141 
142             /* Utilizing symmetry properties to the maximum to minimize the number of multiplications */
143             o[0] = g_ai2_ihevc_trans_4[1][0] * pi2_src[src_strd]
144                             + g_ai2_ihevc_trans_4[3][0] * pi2_src[3 * src_strd];
145             o[1] = g_ai2_ihevc_trans_4[1][1] * pi2_src[src_strd]
146                             + g_ai2_ihevc_trans_4[3][1] * pi2_src[3 * src_strd];
147             e[0] = g_ai2_ihevc_trans_4[0][0] * pi2_src[0]
148                             + g_ai2_ihevc_trans_4[2][0] * pi2_src[2 * src_strd];
149             e[1] = g_ai2_ihevc_trans_4[0][1] * pi2_src[0]
150                             + g_ai2_ihevc_trans_4[2][1] * pi2_src[2 * src_strd];
151 
152             pi2_tmp[0] =
153                             CLIP_S16(((e[0] + o[0] + add) >> shift));
154             pi2_tmp[1] =
155                             CLIP_S16(((e[1] + o[1] + add) >> shift));
156             pi2_tmp[2] =
157                             CLIP_S16(((e[1] - o[1] + add) >> shift));
158             pi2_tmp[3] =
159                             CLIP_S16(((e[0] - o[0] + add) >> shift));
160 
161         }
162         pi2_src++;
163         pi2_tmp += trans_size;
164         zero_cols = zero_cols >> 1;
165     }
166 
167     pi2_tmp = pi2_tmp_orig;
168 
169     /* Inverse Transform 2nd stage */
170     shift = IT_SHIFT_STAGE_2;
171     add = 1 << (shift - 1);
172 
173     for(j = 0; j < trans_size; j++)
174     {
175         WORD32 itrans_out;
176         /* Utilizing symmetry properties to the maximum to minimize the number of multiplications */
177         o[0] = g_ai2_ihevc_trans_4[1][0] * pi2_tmp[trans_size]
178                         + g_ai2_ihevc_trans_4[3][0] * pi2_tmp[3 * trans_size];
179         o[1] = g_ai2_ihevc_trans_4[1][1] * pi2_tmp[trans_size]
180                         + g_ai2_ihevc_trans_4[3][1] * pi2_tmp[3 * trans_size];
181         e[0] = g_ai2_ihevc_trans_4[0][0] * pi2_tmp[0]
182                         + g_ai2_ihevc_trans_4[2][0] * pi2_tmp[2 * trans_size];
183         e[1] = g_ai2_ihevc_trans_4[0][1] * pi2_tmp[0]
184                         + g_ai2_ihevc_trans_4[2][1] * pi2_tmp[2 * trans_size];
185 
186         itrans_out =
187                         CLIP_S16(((e[0] + o[0] + add) >> shift));
188         pu1_dst[0 * 2] = CLIP_U8((itrans_out + pu1_pred[0 * 2]));
189         itrans_out =
190                         CLIP_S16(((e[1] + o[1] + add) >> shift));
191         pu1_dst[1 * 2] = CLIP_U8((itrans_out + pu1_pred[1 * 2]));
192         itrans_out =
193                         CLIP_S16(((e[1] - o[1] + add) >> shift));
194         pu1_dst[2 * 2] = CLIP_U8((itrans_out + pu1_pred[2 * 2]));
195         itrans_out =
196                         CLIP_S16(((e[0] - o[0] + add) >> shift));
197         pu1_dst[3 * 2] = CLIP_U8((itrans_out + pu1_pred[3 * 2]));
198 
199         pi2_tmp++;
200         pu1_pred += pred_strd;
201         pu1_dst += dst_strd;
202 
203     }
204 }
205 
206