/****************************************************************************** * * Copyright (C) 2012 Ittiam Systems Pvt Ltd, Bangalore * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at: * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. * ******************************************************************************/ /** ******************************************************************************* * @file * ihevc_itrans_recon_32x32.c * * @brief * Contains function definitions for inverse transform and reconstruction 32x32 * * * @author * 100470 * * @par List of Functions: * - ihevc_itrans_recon_32x32() * * @remarks * None * ******************************************************************************* */ #include #include #include "ihevc_typedefs.h" #include "ihevc_macros.h" #include "ihevc_platform_macros.h" #include "ihevc_defs.h" #include "ihevc_trans_tables.h" #include "ihevc_itrans_recon.h" #include "ihevc_func_selector.h" #include "ihevc_trans_macros.h" /** ******************************************************************************* * * @brief * This function performs Inverse transform and reconstruction for 32x32 * input block * * @par Description: * Performs inverse transform and adds the prediction data and clips output * to 8 bit * * @param[in] pi2_src * Input 32x32 coefficients * * @param[in] pi2_tmp * Temporary 32x32 buffer for storing inverse * * transform * 1st stage output * * @param[in] pu1_pred * Prediction 32x32 block * * @param[out] pu1_dst * Output 32x32 block * * @param[in] src_strd * Input stride * * @param[in] pred_strd * Prediction stride * * @param[in] dst_strd * Output Stride * * @param[in] shift * Output shift * * @param[in] zero_cols * Zero columns in pi2_src * * @returns Void * * @remarks * None * ******************************************************************************* */ void ihevc_itrans_recon_32x32(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) { WORD32 j, k; WORD32 e[16], o[16]; WORD32 ee[8], eo[8]; WORD32 eee[4], eeo[4]; WORD32 eeee[2], eeeo[2]; WORD32 add; WORD32 shift; WORD16 *pi2_tmp_orig; WORD32 trans_size; WORD32 zero_rows_2nd_stage = zero_cols; WORD32 row_limit_2nd_stage; trans_size = TRANS_SIZE_32; pi2_tmp_orig = pi2_tmp; if((zero_cols & 0xFFFFFFF0) == 0xFFFFFFF0) row_limit_2nd_stage = 4; else if((zero_cols & 0xFFFFFF00) == 0xFFFFFF00) row_limit_2nd_stage = 8; else row_limit_2nd_stage = TRANS_SIZE_32; if((zero_rows & 0xFFFFFFF0) == 0xFFFFFFF0) /* First 4 rows of input are non-zero */ { /************************************************************************************************/ /**********************************START - IT_RECON_32x32****************************************/ /************************************************************************************************/ /* Inverse Transform 1st stage */ shift = IT_SHIFT_STAGE_1; add = 1 << (shift - 1); for(j = 0; j < row_limit_2nd_stage; j++) { /* Checking for Zero Cols */ if((zero_cols & 1) == 1) { memset(pi2_tmp, 0, trans_size * sizeof(WORD16)); } else { /* Utilizing symmetry properties to the maximum to minimize the number of multiplications */ for(k = 0; k < 16; k++) { o[k] = g_ai2_ihevc_trans_32[1][k] * pi2_src[src_strd] + g_ai2_ihevc_trans_32[3][k] * pi2_src[3 * src_strd]; } for(k = 0; k < 8; k++) { eo[k] = g_ai2_ihevc_trans_32[2][k] * pi2_src[2 * src_strd]; } // for(k = 0; k < 4; k++) { eeo[0] = 0; eeo[1] = 0; eeo[2] = 0; eeo[3] = 0; } eeeo[0] = 0; eeeo[1] = 0; eeee[0] = g_ai2_ihevc_trans_32[0][0] * pi2_src[0]; eeee[1] = g_ai2_ihevc_trans_32[0][1] * pi2_src[0]; /* Combining e and o terms at each hierarchy levels to calculate the final spatial domain vector */ eee[0] = eeee[0] + eeeo[0]; eee[3] = eeee[0] - eeeo[0]; eee[1] = eeee[1] + eeeo[1]; eee[2] = eeee[1] - eeeo[1]; for(k = 0; k < 4; k++) { ee[k] = eee[k] + eeo[k]; ee[k + 4] = eee[3 - k] - eeo[3 - k]; } for(k = 0; k < 8; k++) { e[k] = ee[k] + eo[k]; e[k + 8] = ee[7 - k] - eo[7 - k]; } for(k = 0; k < 16; k++) { pi2_tmp[k] = CLIP_S16(((e[k] + o[k] + add) >> shift)); pi2_tmp[k + 16] = CLIP_S16(((e[15 - k] - o[15 - k] + add) >> shift)); } } pi2_src++; pi2_tmp += trans_size; zero_cols = zero_cols >> 1; } pi2_tmp = pi2_tmp_orig; /* Inverse Transform 2nd stage */ shift = IT_SHIFT_STAGE_2; add = 1 << (shift - 1); if((zero_rows_2nd_stage & 0xFFFFFFF0) == 0xFFFFFFF0) /* First 4 rows of output of 1st stage are non-zero */ { for(j = 0; j < trans_size; j++) { /* Utilizing symmetry properties to the maximum to minimize the number of multiplications */ for(k = 0; k < 16; k++) { o[k] = g_ai2_ihevc_trans_32[1][k] * pi2_tmp[trans_size] + g_ai2_ihevc_trans_32[3][k] * pi2_tmp[3 * trans_size]; } for(k = 0; k < 8; k++) { eo[k] = g_ai2_ihevc_trans_32[2][k] * pi2_tmp[2 * trans_size]; } // for(k = 0; k < 4; k++) { eeo[0] = 0; eeo[1] = 0; eeo[2] = 0; eeo[3] = 0; } eeeo[0] = 0; eeeo[1] = 0; eeee[0] = g_ai2_ihevc_trans_32[0][0] * pi2_tmp[0]; eeee[1] = g_ai2_ihevc_trans_32[0][1] * pi2_tmp[0]; /* Combining e and o terms at each hierarchy levels to calculate the final spatial domain vector */ eee[0] = eeee[0] + eeeo[0]; eee[3] = eeee[0] - eeeo[0]; eee[1] = eeee[1] + eeeo[1]; eee[2] = eeee[1] - eeeo[1]; for(k = 0; k < 4; k++) { ee[k] = eee[k] + eeo[k]; ee[k + 4] = eee[3 - k] - eeo[3 - k]; } for(k = 0; k < 8; k++) { e[k] = ee[k] + eo[k]; e[k + 8] = ee[7 - k] - eo[7 - k]; } for(k = 0; k < 16; k++) { WORD32 itrans_out; itrans_out = CLIP_S16(((e[k] + o[k] + add) >> shift)); pu1_dst[k] = CLIP_U8((itrans_out + pu1_pred[k])); itrans_out = CLIP_S16(((e[15 - k] - o[15 - k] + add) >> shift)); pu1_dst[k + 16] = CLIP_U8((itrans_out + pu1_pred[k + 16])); } pi2_tmp++; pu1_pred += pred_strd; pu1_dst += dst_strd; } } else if((zero_rows_2nd_stage & 0xFFFFFF00) == 0xFFFFFF00) /* First 8 rows of output of 1st stage are non-zero */ { for(j = 0; j < trans_size; j++) { /* Utilizing symmetry properties to the maximum to minimize the number of multiplications */ for(k = 0; k < 16; k++) { o[k] = g_ai2_ihevc_trans_32[1][k] * pi2_tmp[trans_size] + g_ai2_ihevc_trans_32[3][k] * pi2_tmp[3 * trans_size] + g_ai2_ihevc_trans_32[5][k] * pi2_tmp[5 * trans_size] + g_ai2_ihevc_trans_32[7][k] * pi2_tmp[7 * trans_size]; } for(k = 0; k < 8; k++) { eo[k] = g_ai2_ihevc_trans_32[2][k] * pi2_tmp[2 * trans_size] + g_ai2_ihevc_trans_32[6][k] * pi2_tmp[6 * trans_size]; } for(k = 0; k < 4; k++) { eeo[k] = g_ai2_ihevc_trans_32[4][k] * pi2_tmp[4 * trans_size]; } eeeo[0] = 0; eeeo[1] = 0; eeee[0] = g_ai2_ihevc_trans_32[0][0] * pi2_tmp[0]; eeee[1] = g_ai2_ihevc_trans_32[0][1] * pi2_tmp[0]; /* Combining e and o terms at each hierarchy levels to calculate the final spatial domain vector */ eee[0] = eeee[0] + eeeo[0]; eee[3] = eeee[0] - eeeo[0]; eee[1] = eeee[1] + eeeo[1]; eee[2] = eeee[1] - eeeo[1]; for(k = 0; k < 4; k++) { ee[k] = eee[k] + eeo[k]; ee[k + 4] = eee[3 - k] - eeo[3 - k]; } for(k = 0; k < 8; k++) { e[k] = ee[k] + eo[k]; e[k + 8] = ee[7 - k] - eo[7 - k]; } for(k = 0; k < 16; k++) { WORD32 itrans_out; itrans_out = CLIP_S16(((e[k] + o[k] + add) >> shift)); pu1_dst[k] = CLIP_U8((itrans_out + pu1_pred[k])); itrans_out = CLIP_S16(((e[15 - k] - o[15 - k] + add) >> shift)); pu1_dst[k + 16] = CLIP_U8((itrans_out + pu1_pred[k + 16])); } pi2_tmp++; pu1_pred += pred_strd; pu1_dst += dst_strd; } } else /* All rows of output of 1st stage are non-zero */ { for(j = 0; j < trans_size; j++) { /* Utilizing symmetry properties to the maximum to minimize the number of multiplications */ for(k = 0; k < 16; k++) { o[k] = g_ai2_ihevc_trans_32[1][k] * pi2_tmp[trans_size] + g_ai2_ihevc_trans_32[3][k] * pi2_tmp[3 * trans_size] + g_ai2_ihevc_trans_32[5][k] * pi2_tmp[5 * trans_size] + g_ai2_ihevc_trans_32[7][k] * pi2_tmp[7 * trans_size] + g_ai2_ihevc_trans_32[9][k] * pi2_tmp[9 * trans_size] + g_ai2_ihevc_trans_32[11][k] * pi2_tmp[11 * trans_size] + g_ai2_ihevc_trans_32[13][k] * pi2_tmp[13 * trans_size] + g_ai2_ihevc_trans_32[15][k] * pi2_tmp[15 * trans_size] + g_ai2_ihevc_trans_32[17][k] * pi2_tmp[17 * trans_size] + g_ai2_ihevc_trans_32[19][k] * pi2_tmp[19 * trans_size] + g_ai2_ihevc_trans_32[21][k] * pi2_tmp[21 * trans_size] + g_ai2_ihevc_trans_32[23][k] * pi2_tmp[23 * trans_size] + g_ai2_ihevc_trans_32[25][k] * pi2_tmp[25 * trans_size] + g_ai2_ihevc_trans_32[27][k] * pi2_tmp[27 * trans_size] + g_ai2_ihevc_trans_32[29][k] * pi2_tmp[29 * trans_size] + g_ai2_ihevc_trans_32[31][k] * pi2_tmp[31 * trans_size]; } for(k = 0; k < 8; k++) { eo[k] = g_ai2_ihevc_trans_32[2][k] * pi2_tmp[2 * trans_size] + g_ai2_ihevc_trans_32[6][k] * pi2_tmp[6 * trans_size] + g_ai2_ihevc_trans_32[10][k] * pi2_tmp[10 * trans_size] + g_ai2_ihevc_trans_32[14][k] * pi2_tmp[14 * trans_size] + g_ai2_ihevc_trans_32[18][k] * pi2_tmp[18 * trans_size] + g_ai2_ihevc_trans_32[22][k] * pi2_tmp[22 * trans_size] + g_ai2_ihevc_trans_32[26][k] * pi2_tmp[26 * trans_size] + g_ai2_ihevc_trans_32[30][k] * pi2_tmp[30 * trans_size]; } for(k = 0; k < 4; k++) { eeo[k] = g_ai2_ihevc_trans_32[4][k] * pi2_tmp[4 * trans_size] + g_ai2_ihevc_trans_32[12][k] * pi2_tmp[12 * trans_size] + g_ai2_ihevc_trans_32[20][k] * pi2_tmp[20 * trans_size] + g_ai2_ihevc_trans_32[28][k] * pi2_tmp[28 * trans_size]; } eeeo[0] = g_ai2_ihevc_trans_32[8][0] * pi2_tmp[8 * trans_size] + g_ai2_ihevc_trans_32[24][0] * pi2_tmp[24 * trans_size]; eeeo[1] = g_ai2_ihevc_trans_32[8][1] * pi2_tmp[8 * trans_size] + g_ai2_ihevc_trans_32[24][1] * pi2_tmp[24 * trans_size]; eeee[0] = g_ai2_ihevc_trans_32[0][0] * pi2_tmp[0] + g_ai2_ihevc_trans_32[16][0] * pi2_tmp[16 * trans_size]; eeee[1] = g_ai2_ihevc_trans_32[0][1] * pi2_tmp[0] + g_ai2_ihevc_trans_32[16][1] * pi2_tmp[16 * trans_size]; /* Combining e and o terms at each hierarchy levels to calculate the final spatial domain vector */ eee[0] = eeee[0] + eeeo[0]; eee[3] = eeee[0] - eeeo[0]; eee[1] = eeee[1] + eeeo[1]; eee[2] = eeee[1] - eeeo[1]; for(k = 0; k < 4; k++) { ee[k] = eee[k] + eeo[k]; ee[k + 4] = eee[3 - k] - eeo[3 - k]; } for(k = 0; k < 8; k++) { e[k] = ee[k] + eo[k]; e[k + 8] = ee[7 - k] - eo[7 - k]; } for(k = 0; k < 16; k++) { WORD32 itrans_out; itrans_out = CLIP_S16(((e[k] + o[k] + add) >> shift)); pu1_dst[k] = CLIP_U8((itrans_out + pu1_pred[k])); itrans_out = CLIP_S16(((e[15 - k] - o[15 - k] + add) >> shift)); pu1_dst[k + 16] = CLIP_U8((itrans_out + pu1_pred[k + 16])); } pi2_tmp++; pu1_pred += pred_strd; pu1_dst += dst_strd; } } /************************************************************************************************/ /************************************END - IT_RECON_32x32****************************************/ /************************************************************************************************/ } else if((zero_rows & 0xFFFFFF00) == 0xFFFFFF00) /* First 8 rows of input are non-zero */ { /************************************************************************************************/ /**********************************START - IT_RECON_32x32****************************************/ /************************************************************************************************/ /* Inverse Transform 1st stage */ shift = IT_SHIFT_STAGE_1; add = 1 << (shift - 1); for(j = 0; j < row_limit_2nd_stage; j++) { /* Checking for Zero Cols */ if((zero_cols & 1) == 1) { memset(pi2_tmp, 0, trans_size * sizeof(WORD16)); } else { /* Utilizing symmetry properties to the maximum to minimize the number of multiplications */ for(k = 0; k < 16; k++) { o[k] = g_ai2_ihevc_trans_32[1][k] * pi2_src[src_strd] + g_ai2_ihevc_trans_32[3][k] * pi2_src[3 * src_strd] + g_ai2_ihevc_trans_32[5][k] * pi2_src[5 * src_strd] + g_ai2_ihevc_trans_32[7][k] * pi2_src[7 * src_strd]; } for(k = 0; k < 8; k++) { eo[k] = g_ai2_ihevc_trans_32[2][k] * pi2_src[2 * src_strd] + g_ai2_ihevc_trans_32[6][k] * pi2_src[6 * src_strd]; } for(k = 0; k < 4; k++) { eeo[k] = g_ai2_ihevc_trans_32[4][k] * pi2_src[4 * src_strd]; } eeeo[0] = 0; eeeo[1] = 0; eeee[0] = g_ai2_ihevc_trans_32[0][0] * pi2_src[0]; eeee[1] = g_ai2_ihevc_trans_32[0][1] * pi2_src[0]; /* Combining e and o terms at each hierarchy levels to calculate the final spatial domain vector */ eee[0] = eeee[0] + eeeo[0]; eee[3] = eeee[0] - eeeo[0]; eee[1] = eeee[1] + eeeo[1]; eee[2] = eeee[1] - eeeo[1]; for(k = 0; k < 4; k++) { ee[k] = eee[k] + eeo[k]; ee[k + 4] = eee[3 - k] - eeo[3 - k]; } for(k = 0; k < 8; k++) { e[k] = ee[k] + eo[k]; e[k + 8] = ee[7 - k] - eo[7 - k]; } for(k = 0; k < 16; k++) { pi2_tmp[k] = CLIP_S16(((e[k] + o[k] + add) >> shift)); pi2_tmp[k + 16] = CLIP_S16(((e[15 - k] - o[15 - k] + add) >> shift)); } } pi2_src++; pi2_tmp += trans_size; zero_cols = zero_cols >> 1; } pi2_tmp = pi2_tmp_orig; /* Inverse Transform 2nd stage */ shift = IT_SHIFT_STAGE_2; add = 1 << (shift - 1); if((zero_rows_2nd_stage & 0xFFFFFFF0) == 0xFFFFFFF0) /* First 4 rows of output of 1st stage are non-zero */ { for(j = 0; j < trans_size; j++) { /* Utilizing symmetry properties to the maximum to minimize the number of multiplications */ for(k = 0; k < 16; k++) { o[k] = g_ai2_ihevc_trans_32[1][k] * pi2_tmp[trans_size] + g_ai2_ihevc_trans_32[3][k] * pi2_tmp[3 * trans_size]; } for(k = 0; k < 8; k++) { eo[k] = g_ai2_ihevc_trans_32[2][k] * pi2_tmp[2 * trans_size]; } // for(k = 0; k < 4; k++) { eeo[0] = 0; eeo[1] = 0; eeo[2] = 0; eeo[3] = 0; } eeeo[0] = 0; eeeo[1] = 0; eeee[0] = g_ai2_ihevc_trans_32[0][0] * pi2_tmp[0]; eeee[1] = g_ai2_ihevc_trans_32[0][1] * pi2_tmp[0]; /* Combining e and o terms at each hierarchy levels to calculate the final spatial domain vector */ eee[0] = eeee[0] + eeeo[0]; eee[3] = eeee[0] - eeeo[0]; eee[1] = eeee[1] + eeeo[1]; eee[2] = eeee[1] - eeeo[1]; for(k = 0; k < 4; k++) { ee[k] = eee[k] + eeo[k]; ee[k + 4] = eee[3 - k] - eeo[3 - k]; } for(k = 0; k < 8; k++) { e[k] = ee[k] + eo[k]; e[k + 8] = ee[7 - k] - eo[7 - k]; } for(k = 0; k < 16; k++) { WORD32 itrans_out; itrans_out = CLIP_S16(((e[k] + o[k] + add) >> shift)); pu1_dst[k] = CLIP_U8((itrans_out + pu1_pred[k])); itrans_out = CLIP_S16(((e[15 - k] - o[15 - k] + add) >> shift)); pu1_dst[k + 16] = CLIP_U8((itrans_out + pu1_pred[k + 16])); } pi2_tmp++; pu1_pred += pred_strd; pu1_dst += dst_strd; } } else if((zero_rows_2nd_stage & 0xFFFFFF00) == 0xFFFFFF00) /* First 8 rows of output of 1st stage are non-zero */ { for(j = 0; j < trans_size; j++) { /* Utilizing symmetry properties to the maximum to minimize the number of multiplications */ for(k = 0; k < 16; k++) { o[k] = g_ai2_ihevc_trans_32[1][k] * pi2_tmp[trans_size] + g_ai2_ihevc_trans_32[3][k] * pi2_tmp[3 * trans_size] + g_ai2_ihevc_trans_32[5][k] * pi2_tmp[5 * trans_size] + g_ai2_ihevc_trans_32[7][k] * pi2_tmp[7 * trans_size]; } for(k = 0; k < 8; k++) { eo[k] = g_ai2_ihevc_trans_32[2][k] * pi2_tmp[2 * trans_size] + g_ai2_ihevc_trans_32[6][k] * pi2_tmp[6 * trans_size]; } for(k = 0; k < 4; k++) { eeo[k] = g_ai2_ihevc_trans_32[4][k] * pi2_tmp[4 * trans_size]; } eeeo[0] = 0; eeeo[1] = 0; eeee[0] = g_ai2_ihevc_trans_32[0][0] * pi2_tmp[0]; eeee[1] = g_ai2_ihevc_trans_32[0][1] * pi2_tmp[0]; /* Combining e and o terms at each hierarchy levels to calculate the final spatial domain vector */ eee[0] = eeee[0] + eeeo[0]; eee[3] = eeee[0] - eeeo[0]; eee[1] = eeee[1] + eeeo[1]; eee[2] = eeee[1] - eeeo[1]; for(k = 0; k < 4; k++) { ee[k] = eee[k] + eeo[k]; ee[k + 4] = eee[3 - k] - eeo[3 - k]; } for(k = 0; k < 8; k++) { e[k] = ee[k] + eo[k]; e[k + 8] = ee[7 - k] - eo[7 - k]; } for(k = 0; k < 16; k++) { WORD32 itrans_out; itrans_out = CLIP_S16(((e[k] + o[k] + add) >> shift)); pu1_dst[k] = CLIP_U8((itrans_out + pu1_pred[k])); itrans_out = CLIP_S16(((e[15 - k] - o[15 - k] + add) >> shift)); pu1_dst[k + 16] = CLIP_U8((itrans_out + pu1_pred[k + 16])); } pi2_tmp++; pu1_pred += pred_strd; pu1_dst += dst_strd; } } else /* All rows of output of 1st stage are non-zero */ { for(j = 0; j < trans_size; j++) { /* Utilizing symmetry properties to the maximum to minimize the number of multiplications */ for(k = 0; k < 16; k++) { o[k] = g_ai2_ihevc_trans_32[1][k] * pi2_tmp[trans_size] + g_ai2_ihevc_trans_32[3][k] * pi2_tmp[3 * trans_size] + g_ai2_ihevc_trans_32[5][k] * pi2_tmp[5 * trans_size] + g_ai2_ihevc_trans_32[7][k] * pi2_tmp[7 * trans_size] + g_ai2_ihevc_trans_32[9][k] * pi2_tmp[9 * trans_size] + g_ai2_ihevc_trans_32[11][k] * pi2_tmp[11 * trans_size] + g_ai2_ihevc_trans_32[13][k] * pi2_tmp[13 * trans_size] + g_ai2_ihevc_trans_32[15][k] * pi2_tmp[15 * trans_size] + g_ai2_ihevc_trans_32[17][k] * pi2_tmp[17 * trans_size] + g_ai2_ihevc_trans_32[19][k] * pi2_tmp[19 * trans_size] + g_ai2_ihevc_trans_32[21][k] * pi2_tmp[21 * trans_size] + g_ai2_ihevc_trans_32[23][k] * pi2_tmp[23 * trans_size] + g_ai2_ihevc_trans_32[25][k] * pi2_tmp[25 * trans_size] + g_ai2_ihevc_trans_32[27][k] * pi2_tmp[27 * trans_size] + g_ai2_ihevc_trans_32[29][k] * pi2_tmp[29 * trans_size] + g_ai2_ihevc_trans_32[31][k] * pi2_tmp[31 * trans_size]; } for(k = 0; k < 8; k++) { eo[k] = g_ai2_ihevc_trans_32[2][k] * pi2_tmp[2 * trans_size] + g_ai2_ihevc_trans_32[6][k] * pi2_tmp[6 * trans_size] + g_ai2_ihevc_trans_32[10][k] * pi2_tmp[10 * trans_size] + g_ai2_ihevc_trans_32[14][k] * pi2_tmp[14 * trans_size] + g_ai2_ihevc_trans_32[18][k] * pi2_tmp[18 * trans_size] + g_ai2_ihevc_trans_32[22][k] * pi2_tmp[22 * trans_size] + g_ai2_ihevc_trans_32[26][k] * pi2_tmp[26 * trans_size] + g_ai2_ihevc_trans_32[30][k] * pi2_tmp[30 * trans_size]; } for(k = 0; k < 4; k++) { eeo[k] = g_ai2_ihevc_trans_32[4][k] * pi2_tmp[4 * trans_size] + g_ai2_ihevc_trans_32[12][k] * pi2_tmp[12 * trans_size] + g_ai2_ihevc_trans_32[20][k] * pi2_tmp[20 * trans_size] + g_ai2_ihevc_trans_32[28][k] * pi2_tmp[28 * trans_size]; } eeeo[0] = g_ai2_ihevc_trans_32[8][0] * pi2_tmp[8 * trans_size] + g_ai2_ihevc_trans_32[24][0] * pi2_tmp[24 * trans_size]; eeeo[1] = g_ai2_ihevc_trans_32[8][1] * pi2_tmp[8 * trans_size] + g_ai2_ihevc_trans_32[24][1] * pi2_tmp[24 * trans_size]; eeee[0] = g_ai2_ihevc_trans_32[0][0] * pi2_tmp[0] + g_ai2_ihevc_trans_32[16][0] * pi2_tmp[16 * trans_size]; eeee[1] = g_ai2_ihevc_trans_32[0][1] * pi2_tmp[0] + g_ai2_ihevc_trans_32[16][1] * pi2_tmp[16 * trans_size]; /* Combining e and o terms at each hierarchy levels to calculate the final spatial domain vector */ eee[0] = eeee[0] + eeeo[0]; eee[3] = eeee[0] - eeeo[0]; eee[1] = eeee[1] + eeeo[1]; eee[2] = eeee[1] - eeeo[1]; for(k = 0; k < 4; k++) { ee[k] = eee[k] + eeo[k]; ee[k + 4] = eee[3 - k] - eeo[3 - k]; } for(k = 0; k < 8; k++) { e[k] = ee[k] + eo[k]; e[k + 8] = ee[7 - k] - eo[7 - k]; } for(k = 0; k < 16; k++) { WORD32 itrans_out; itrans_out = CLIP_S16(((e[k] + o[k] + add) >> shift)); pu1_dst[k] = CLIP_U8((itrans_out + pu1_pred[k])); itrans_out = CLIP_S16(((e[15 - k] - o[15 - k] + add) >> shift)); pu1_dst[k + 16] = CLIP_U8((itrans_out + pu1_pred[k + 16])); } pi2_tmp++; pu1_pred += pred_strd; pu1_dst += dst_strd; } } /************************************************************************************************/ /************************************END - IT_RECON_32x32****************************************/ /************************************************************************************************/ } else /* All rows of input are non-zero */ { /************************************************************************************************/ /**********************************START - IT_RECON_32x32****************************************/ /************************************************************************************************/ /* Inverse Transform 1st stage */ shift = IT_SHIFT_STAGE_1; add = 1 << (shift - 1); for(j = 0; j < row_limit_2nd_stage; j++) { /* Checking for Zero Cols */ if((zero_cols & 1) == 1) { memset(pi2_tmp, 0, trans_size * sizeof(WORD16)); } else { /* Utilizing symmetry properties to the maximum to minimize the number of multiplications */ for(k = 0; k < 16; k++) { o[k] = g_ai2_ihevc_trans_32[1][k] * pi2_src[src_strd] + g_ai2_ihevc_trans_32[3][k] * pi2_src[3 * src_strd] + g_ai2_ihevc_trans_32[5][k] * pi2_src[5 * src_strd] + g_ai2_ihevc_trans_32[7][k] * pi2_src[7 * src_strd] + g_ai2_ihevc_trans_32[9][k] * pi2_src[9 * src_strd] + g_ai2_ihevc_trans_32[11][k] * pi2_src[11 * src_strd] + g_ai2_ihevc_trans_32[13][k] * pi2_src[13 * src_strd] + g_ai2_ihevc_trans_32[15][k] * pi2_src[15 * src_strd] + g_ai2_ihevc_trans_32[17][k] * pi2_src[17 * src_strd] + g_ai2_ihevc_trans_32[19][k] * pi2_src[19 * src_strd] + g_ai2_ihevc_trans_32[21][k] * pi2_src[21 * src_strd] + g_ai2_ihevc_trans_32[23][k] * pi2_src[23 * src_strd] + g_ai2_ihevc_trans_32[25][k] * pi2_src[25 * src_strd] + g_ai2_ihevc_trans_32[27][k] * pi2_src[27 * src_strd] + g_ai2_ihevc_trans_32[29][k] * pi2_src[29 * src_strd] + g_ai2_ihevc_trans_32[31][k] * pi2_src[31 * src_strd]; } for(k = 0; k < 8; k++) { eo[k] = g_ai2_ihevc_trans_32[2][k] * pi2_src[2 * src_strd] + g_ai2_ihevc_trans_32[6][k] * pi2_src[6 * src_strd] + g_ai2_ihevc_trans_32[10][k] * pi2_src[10 * src_strd] + g_ai2_ihevc_trans_32[14][k] * pi2_src[14 * src_strd] + g_ai2_ihevc_trans_32[18][k] * pi2_src[18 * src_strd] + g_ai2_ihevc_trans_32[22][k] * pi2_src[22 * src_strd] + g_ai2_ihevc_trans_32[26][k] * pi2_src[26 * src_strd] + g_ai2_ihevc_trans_32[30][k] * pi2_src[30 * src_strd]; } for(k = 0; k < 4; k++) { eeo[k] = g_ai2_ihevc_trans_32[4][k] * pi2_src[4 * src_strd] + g_ai2_ihevc_trans_32[12][k] * pi2_src[12 * src_strd] + g_ai2_ihevc_trans_32[20][k] * pi2_src[20 * src_strd] + g_ai2_ihevc_trans_32[28][k] * pi2_src[28 * src_strd]; } eeeo[0] = g_ai2_ihevc_trans_32[8][0] * pi2_src[8 * src_strd] + g_ai2_ihevc_trans_32[24][0] * pi2_src[24 * src_strd]; eeeo[1] = g_ai2_ihevc_trans_32[8][1] * pi2_src[8 * src_strd] + g_ai2_ihevc_trans_32[24][1] * pi2_src[24 * src_strd]; eeee[0] = g_ai2_ihevc_trans_32[0][0] * pi2_src[0] + g_ai2_ihevc_trans_32[16][0] * pi2_src[16 * src_strd]; eeee[1] = g_ai2_ihevc_trans_32[0][1] * pi2_src[0] + g_ai2_ihevc_trans_32[16][1] * pi2_src[16 * src_strd]; /* Combining e and o terms at each hierarchy levels to calculate the final spatial domain vector */ eee[0] = eeee[0] + eeeo[0]; eee[3] = eeee[0] - eeeo[0]; eee[1] = eeee[1] + eeeo[1]; eee[2] = eeee[1] - eeeo[1]; for(k = 0; k < 4; k++) { ee[k] = eee[k] + eeo[k]; ee[k + 4] = eee[3 - k] - eeo[3 - k]; } for(k = 0; k < 8; k++) { e[k] = ee[k] + eo[k]; e[k + 8] = ee[7 - k] - eo[7 - k]; } for(k = 0; k < 16; k++) { pi2_tmp[k] = CLIP_S16(((e[k] + o[k] + add) >> shift)); pi2_tmp[k + 16] = CLIP_S16(((e[15 - k] - o[15 - k] + add) >> shift)); } } pi2_src++; pi2_tmp += trans_size; zero_cols = zero_cols >> 1; } pi2_tmp = pi2_tmp_orig; /* Inverse Transform 2nd stage */ shift = IT_SHIFT_STAGE_2; add = 1 << (shift - 1); if((zero_rows_2nd_stage & 0xFFFFFFF0) == 0xFFFFFFF0) /* First 4 rows of output of 1st stage are non-zero */ { for(j = 0; j < trans_size; j++) { /* Utilizing symmetry properties to the maximum to minimize the number of multiplications */ for(k = 0; k < 16; k++) { o[k] = g_ai2_ihevc_trans_32[1][k] * pi2_tmp[trans_size] + g_ai2_ihevc_trans_32[3][k] * pi2_tmp[3 * trans_size]; } for(k = 0; k < 8; k++) { eo[k] = g_ai2_ihevc_trans_32[2][k] * pi2_tmp[2 * trans_size]; } // for(k = 0; k < 4; k++) { eeo[0] = 0; eeo[1] = 0; eeo[2] = 0; eeo[3] = 0; } eeeo[0] = 0; eeeo[1] = 0; eeee[0] = g_ai2_ihevc_trans_32[0][0] * pi2_tmp[0]; eeee[1] = g_ai2_ihevc_trans_32[0][1] * pi2_tmp[0]; /* Combining e and o terms at each hierarchy levels to calculate the final spatial domain vector */ eee[0] = eeee[0] + eeeo[0]; eee[3] = eeee[0] - eeeo[0]; eee[1] = eeee[1] + eeeo[1]; eee[2] = eeee[1] - eeeo[1]; for(k = 0; k < 4; k++) { ee[k] = eee[k] + eeo[k]; ee[k + 4] = eee[3 - k] - eeo[3 - k]; } for(k = 0; k < 8; k++) { e[k] = ee[k] + eo[k]; e[k + 8] = ee[7 - k] - eo[7 - k]; } for(k = 0; k < 16; k++) { WORD32 itrans_out; itrans_out = CLIP_S16(((e[k] + o[k] + add) >> shift)); pu1_dst[k] = CLIP_U8((itrans_out + pu1_pred[k])); itrans_out = CLIP_S16(((e[15 - k] - o[15 - k] + add) >> shift)); pu1_dst[k + 16] = CLIP_U8((itrans_out + pu1_pred[k + 16])); } pi2_tmp++; pu1_pred += pred_strd; pu1_dst += dst_strd; } } else if((zero_rows_2nd_stage & 0xFFFFFF00) == 0xFFFFFF00) /* First 8 rows of output of 1st stage are non-zero */ { for(j = 0; j < trans_size; j++) { /* Utilizing symmetry properties to the maximum to minimize the number of multiplications */ for(k = 0; k < 16; k++) { o[k] = g_ai2_ihevc_trans_32[1][k] * pi2_tmp[trans_size] + g_ai2_ihevc_trans_32[3][k] * pi2_tmp[3 * trans_size] + g_ai2_ihevc_trans_32[5][k] * pi2_tmp[5 * trans_size] + g_ai2_ihevc_trans_32[7][k] * pi2_tmp[7 * trans_size]; } for(k = 0; k < 8; k++) { eo[k] = g_ai2_ihevc_trans_32[2][k] * pi2_tmp[2 * trans_size] + g_ai2_ihevc_trans_32[6][k] * pi2_tmp[6 * trans_size]; } for(k = 0; k < 4; k++) { eeo[k] = g_ai2_ihevc_trans_32[4][k] * pi2_tmp[4 * trans_size]; } eeeo[0] = 0; eeeo[1] = 0; eeee[0] = g_ai2_ihevc_trans_32[0][0] * pi2_tmp[0]; eeee[1] = g_ai2_ihevc_trans_32[0][1] * pi2_tmp[0]; /* Combining e and o terms at each hierarchy levels to calculate the final spatial domain vector */ eee[0] = eeee[0] + eeeo[0]; eee[3] = eeee[0] - eeeo[0]; eee[1] = eeee[1] + eeeo[1]; eee[2] = eeee[1] - eeeo[1]; for(k = 0; k < 4; k++) { ee[k] = eee[k] + eeo[k]; ee[k + 4] = eee[3 - k] - eeo[3 - k]; } for(k = 0; k < 8; k++) { e[k] = ee[k] + eo[k]; e[k + 8] = ee[7 - k] - eo[7 - k]; } for(k = 0; k < 16; k++) { WORD32 itrans_out; itrans_out = CLIP_S16(((e[k] + o[k] + add) >> shift)); pu1_dst[k] = CLIP_U8((itrans_out + pu1_pred[k])); itrans_out = CLIP_S16(((e[15 - k] - o[15 - k] + add) >> shift)); pu1_dst[k + 16] = CLIP_U8((itrans_out + pu1_pred[k + 16])); } pi2_tmp++; pu1_pred += pred_strd; pu1_dst += dst_strd; } } else /* All rows of output of 1st stage are non-zero */ { for(j = 0; j < trans_size; j++) { /* Utilizing symmetry properties to the maximum to minimize the number of multiplications */ for(k = 0; k < 16; k++) { o[k] = g_ai2_ihevc_trans_32[1][k] * pi2_tmp[trans_size] + g_ai2_ihevc_trans_32[3][k] * pi2_tmp[3 * trans_size] + g_ai2_ihevc_trans_32[5][k] * pi2_tmp[5 * trans_size] + g_ai2_ihevc_trans_32[7][k] * pi2_tmp[7 * trans_size] + g_ai2_ihevc_trans_32[9][k] * pi2_tmp[9 * trans_size] + g_ai2_ihevc_trans_32[11][k] * pi2_tmp[11 * trans_size] + g_ai2_ihevc_trans_32[13][k] * pi2_tmp[13 * trans_size] + g_ai2_ihevc_trans_32[15][k] * pi2_tmp[15 * trans_size] + g_ai2_ihevc_trans_32[17][k] * pi2_tmp[17 * trans_size] + g_ai2_ihevc_trans_32[19][k] * pi2_tmp[19 * trans_size] + g_ai2_ihevc_trans_32[21][k] * pi2_tmp[21 * trans_size] + g_ai2_ihevc_trans_32[23][k] * pi2_tmp[23 * trans_size] + g_ai2_ihevc_trans_32[25][k] * pi2_tmp[25 * trans_size] + g_ai2_ihevc_trans_32[27][k] * pi2_tmp[27 * trans_size] + g_ai2_ihevc_trans_32[29][k] * pi2_tmp[29 * trans_size] + g_ai2_ihevc_trans_32[31][k] * pi2_tmp[31 * trans_size]; } for(k = 0; k < 8; k++) { eo[k] = g_ai2_ihevc_trans_32[2][k] * pi2_tmp[2 * trans_size] + g_ai2_ihevc_trans_32[6][k] * pi2_tmp[6 * trans_size] + g_ai2_ihevc_trans_32[10][k] * pi2_tmp[10 * trans_size] + g_ai2_ihevc_trans_32[14][k] * pi2_tmp[14 * trans_size] + g_ai2_ihevc_trans_32[18][k] * pi2_tmp[18 * trans_size] + g_ai2_ihevc_trans_32[22][k] * pi2_tmp[22 * trans_size] + g_ai2_ihevc_trans_32[26][k] * pi2_tmp[26 * trans_size] + g_ai2_ihevc_trans_32[30][k] * pi2_tmp[30 * trans_size]; } for(k = 0; k < 4; k++) { eeo[k] = g_ai2_ihevc_trans_32[4][k] * pi2_tmp[4 * trans_size] + g_ai2_ihevc_trans_32[12][k] * pi2_tmp[12 * trans_size] + g_ai2_ihevc_trans_32[20][k] * pi2_tmp[20 * trans_size] + g_ai2_ihevc_trans_32[28][k] * pi2_tmp[28 * trans_size]; } eeeo[0] = g_ai2_ihevc_trans_32[8][0] * pi2_tmp[8 * trans_size] + g_ai2_ihevc_trans_32[24][0] * pi2_tmp[24 * trans_size]; eeeo[1] = g_ai2_ihevc_trans_32[8][1] * pi2_tmp[8 * trans_size] + g_ai2_ihevc_trans_32[24][1] * pi2_tmp[24 * trans_size]; eeee[0] = g_ai2_ihevc_trans_32[0][0] * pi2_tmp[0] + g_ai2_ihevc_trans_32[16][0] * pi2_tmp[16 * trans_size]; eeee[1] = g_ai2_ihevc_trans_32[0][1] * pi2_tmp[0] + g_ai2_ihevc_trans_32[16][1] * pi2_tmp[16 * trans_size]; /* Combining e and o terms at each hierarchy levels to calculate the final spatial domain vector */ eee[0] = eeee[0] + eeeo[0]; eee[3] = eeee[0] - eeeo[0]; eee[1] = eeee[1] + eeeo[1]; eee[2] = eeee[1] - eeeo[1]; for(k = 0; k < 4; k++) { ee[k] = eee[k] + eeo[k]; ee[k + 4] = eee[3 - k] - eeo[3 - k]; } for(k = 0; k < 8; k++) { e[k] = ee[k] + eo[k]; e[k + 8] = ee[7 - k] - eo[7 - k]; } for(k = 0; k < 16; k++) { WORD32 itrans_out; itrans_out = CLIP_S16(((e[k] + o[k] + add) >> shift)); pu1_dst[k] = CLIP_U8((itrans_out + pu1_pred[k])); itrans_out = CLIP_S16(((e[15 - k] - o[15 - k] + add) >> shift)); pu1_dst[k + 16] = CLIP_U8((itrans_out + pu1_pred[k + 16])); } pi2_tmp++; pu1_pred += pred_strd; pu1_dst += dst_strd; } } /************************************************************************************************/ /************************************END - IT_RECON_32x32****************************************/ /************************************************************************************************/ } }