/******************************************************************************
|
*
|
* Copyright (C) 2015 The Android Open Source Project
|
*
|
* 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.
|
*
|
*****************************************************************************
|
* Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore
|
*/
|
/**
|
*******************************************************************************
|
* @file
|
* ih264_chroma_intra_pred_filters.c
|
*
|
* @brief
|
* Contains function definitions for chroma intra prediction filters
|
*
|
* @author
|
* Ittiam
|
*
|
* @par List of Functions:
|
* -ih264_intra_pred_chroma_8x8_mode_dc
|
* -ih264_intra_pred_chroma_8x8_mode_horz
|
* -ih264_intra_pred_chroma_8x8_mode_vert
|
* -ih264_intra_pred_chroma_8x8_mode_plane
|
*
|
* @remarks
|
* None
|
*
|
*******************************************************************************
|
*/
|
|
/*****************************************************************************/
|
/* File Includes */
|
/*****************************************************************************/
|
|
/* System include files */
|
#include <stdio.h>
|
#include <stddef.h>
|
#include <string.h>
|
|
/* User include files */
|
#include "ih264_defs.h"
|
#include "ih264_typedefs.h"
|
#include "ih264_macros.h"
|
#include "ih264_platform_macros.h"
|
#include "ih264_intra_pred_filters.h"
|
|
/* Global variables used only in assembly files*/
|
const WORD8 ih264_gai1_intrapred_chroma_plane_coeffs1[] =
|
{ 0x01,0x00,0x01,0x00,
|
0x02,0x00,0x02,0x00,
|
0x03,0x00,0x03,0x00,
|
0x04,0x00,0x04,0x00
|
};
|
const WORD8 ih264_gai1_intrapred_chroma_plane_coeffs2[] =
|
{ 0xfd,0xff,0xfe,0xff,
|
0xff,0xff,0x00,0x00,
|
0x01,0x00,0x02,0x00,
|
0x03,0x00,0x04,0x00,
|
};
|
|
/*****************************************************************************/
|
/* Chroma Intra prediction 8x8 filters */
|
/*****************************************************************************/
|
|
/**
|
*******************************************************************************
|
*
|
* ih264_intra_pred_chroma_8x8_mode_dc
|
*
|
* @brief
|
* Perform Intra prediction for chroma_8x8 mode:DC
|
*
|
* @par Description:
|
* Perform Intra prediction for chroma_8x8 mode:DC ,described in sec 8.3.4.1
|
*
|
* @param[in] pu1_src
|
* UWORD8 pointer to the source containing alternate U and V samples
|
*
|
* @param[out] pu1_dst
|
* UWORD8 pointer to the destination with alternate U and V samples
|
*
|
* @param[in] src_strd
|
* integer source stride
|
*
|
* @param[in] dst_strd
|
* integer destination stride
|
*
|
** @param[in] ngbr_avail
|
* availability of neighbouring pixels
|
*
|
* @returns
|
*
|
* @remarks
|
* None
|
*
|
******************************************************************************
|
*/
|
void ih264_intra_pred_chroma_8x8_mode_dc(UWORD8 *pu1_src,
|
UWORD8 *pu1_dst,
|
WORD32 src_strd,
|
WORD32 dst_strd,
|
WORD32 ngbr_avail)
|
{
|
WORD32 left_avail, left_avail1, left_avail2; /* availability of left predictors (only for DC) */
|
WORD32 top_avail; /* availability of top predictors (only for DC) */
|
UWORD8 *pu1_left = NULL; /* Pointer to start of left predictors */
|
UWORD8 *pu1_top = NULL; /* Pointer to start of top predictors */
|
|
/* temporary variables to store accumulated first left half,second left half,
|
* first top half,second top half of U and V values*/
|
WORD32 val_u_l1 = 0, val_u_l2 = 0, val_u_t1 = 0, val_u_t2 = 0;
|
WORD32 val_v_l1 = 0, val_v_l2 = 0, val_v_t1 = 0, val_v_t2 = 0;
|
|
WORD32 val_u1 = 0, val_u2 = 0, val_v1 = 0, val_v2 = 0;
|
|
WORD32 col, row; /*loop variables*/
|
UNUSED(src_strd);
|
|
left_avail = ngbr_avail & 0x11;
|
left_avail1 = ngbr_avail & 1;
|
left_avail2 = (ngbr_avail >> 4) & 1;
|
top_avail = (ngbr_avail >> 2) & 1;
|
|
pu1_top = pu1_src + 2 * BLK8x8SIZE + 2;
|
pu1_left = pu1_src + 2 * BLK8x8SIZE - 2;
|
|
if(left_avail1)
|
{ /* First 4x4 block*/
|
val_u_l1 += *pu1_left;
|
val_v_l1 += *(pu1_left + 1);
|
pu1_left -= 2;
|
val_u_l1 += *pu1_left;
|
val_v_l1 += *(pu1_left + 1);
|
pu1_left -= 2;
|
val_u_l1 += *pu1_left;
|
val_v_l1 += *(pu1_left + 1);
|
pu1_left -= 2;
|
val_u_l1 += *pu1_left + 2;
|
val_v_l1 += *(pu1_left + 1) + 2;
|
pu1_left -= 2;
|
}
|
else
|
pu1_left -= 2 * 4;
|
|
if(left_avail2)
|
{
|
/* Second 4x4 block*/
|
val_u_l2 += *pu1_left;
|
val_v_l2 += *(pu1_left + 1);
|
pu1_left -= 2;
|
val_u_l2 += *pu1_left;
|
val_v_l2 += *(pu1_left + 1);
|
pu1_left -= 2;
|
val_u_l2 += *pu1_left;
|
val_v_l2 += *(pu1_left + 1);
|
pu1_left -= 2;
|
val_u_l2 += *pu1_left + 2;
|
val_v_l2 += *(pu1_left + 1) + 2;
|
pu1_left -= 2;
|
}
|
else
|
pu1_left -= 2 * 4;
|
|
if(top_avail)
|
{
|
val_u_t1 += *pu1_top + *(pu1_top + 2) + *(pu1_top + 4)
|
+ *(pu1_top + 6) + 2;
|
val_u_t2 += *(pu1_top + 8) + *(pu1_top + 10) + *(pu1_top + 12)
|
+ *(pu1_top + 14) + 2;
|
val_v_t1 += *(pu1_top + 1) + *(pu1_top + 3) + *(pu1_top + 5)
|
+ *(pu1_top + 7) + 2;
|
val_v_t2 += *(pu1_top + 9) + *(pu1_top + 11) + *(pu1_top + 13)
|
+ *(pu1_top + 15) + 2;
|
}
|
|
if(left_avail + top_avail)
|
{
|
val_u1 = (left_avail1 + top_avail) ?
|
((val_u_l1 + val_u_t1)
|
>> (1 + left_avail1 + top_avail)) :128;
|
val_v1 = (left_avail1 + top_avail) ?
|
((val_v_l1 + val_v_t1)
|
>> (1 + left_avail1 + top_avail)) :128;
|
if(top_avail)
|
{
|
val_u2 = val_u_t2 >> 2;
|
val_v2 = val_v_t2 >> 2;
|
}
|
else if(left_avail1)
|
{
|
val_u2 = val_u_l1 >> 2;
|
val_v2 = val_v_l1 >> 2;
|
}
|
else
|
{
|
val_u2 = val_v2 = 128;
|
}
|
|
for(row = 0; row < 4; row++)
|
{
|
/*top left 4x4 block*/
|
for(col = 0; col < 8; col += 2)
|
{
|
*(pu1_dst + row * dst_strd + col) = val_u1;
|
*(pu1_dst + row * dst_strd + col + 1) = val_v1;
|
}
|
/*top right 4x4 block*/
|
for(col = 8; col < 16; col += 2)
|
{
|
*(pu1_dst + row * dst_strd + col) = val_u2;
|
*(pu1_dst + row * dst_strd + col + 1) = val_v2;
|
}
|
}
|
|
if(left_avail2)
|
{
|
val_u1 = val_u_l2 >> 2;
|
val_v1 = val_v_l2 >> 2;
|
}
|
else if(top_avail)
|
{
|
val_u1 = val_u_t1 >> 2;
|
val_v1 = val_v_t1 >> 2;
|
}
|
else
|
{
|
val_u1 = val_v1 = 128;
|
}
|
val_u2 = (left_avail2 + top_avail) ?
|
((val_u_l2 + val_u_t2)
|
>> (1 + left_avail2 + top_avail)) : 128;
|
val_v2 = (left_avail2 + top_avail) ?
|
((val_v_l2 + val_v_t2)
|
>> (1 + left_avail2 + top_avail)) : 128;
|
|
for(row = 4; row < 8; row++)
|
{ /*bottom left 4x4 block*/
|
for(col = 0; col < 8; col += 2)
|
{
|
*(pu1_dst + row * dst_strd + col) = val_u1;
|
*(pu1_dst + row * dst_strd + col + 1) = val_v1;
|
}
|
/*bottom right 4x4 block*/
|
for(col = 8; col < 16; col += 2)
|
{
|
*(pu1_dst + row * dst_strd + col) = val_u2;
|
*(pu1_dst + row * dst_strd + col + 1) = val_v2;
|
}
|
}
|
}
|
else
|
{
|
/* Both left and top are unavailable, set the block to 128 */
|
for(row = 0; row < 8; row++)
|
{
|
memset(pu1_dst + row * dst_strd, 128, 8 * sizeof(UWORD16));
|
}
|
}
|
}
|
|
/**
|
*******************************************************************************
|
*
|
*ih264_intra_pred_chroma_8x8_mode_horz
|
*
|
* @brief
|
* Perform Intra prediction for chroma_8x8 mode:Horizontal
|
*
|
* @par Description:
|
* Perform Intra prediction for chroma_8x8 mode:Horizontal ,described in sec 8.3.4.2
|
*
|
* @param[in] pu1_src
|
* UWORD8 pointer to the source containing alternate U and V samples
|
*
|
* @param[out] pu1_dst
|
* UWORD8 pointer to the destination with alternate U and V samples
|
*
|
* @param[in] src_strd
|
* integer source stride
|
*
|
* @param[in] dst_strd
|
* integer destination stride
|
*
|
* @param[in] ngbr_avail
|
* availability of neighbouring pixels(Not used in this function)
|
*
|
* @returns
|
*
|
* @remarks
|
* None
|
*
|
******************************************************************************
|
*/
|
void ih264_intra_pred_chroma_8x8_mode_horz(UWORD8 *pu1_src,
|
UWORD8 *pu1_dst,
|
WORD32 src_strd,
|
WORD32 dst_strd,
|
WORD32 ngbr_avail)
|
{
|
|
UWORD8 *pu1_left = NULL; /* Pointer to start of top predictors */
|
WORD32 rows, cols; /* loop variables*/
|
UNUSED(src_strd);
|
UNUSED(ngbr_avail);
|
pu1_left = pu1_src + 2 * BLK8x8SIZE - 2;
|
for(rows = 0; rows < 8; rows++)
|
{
|
for(cols = 0; cols < 16; cols += 2)
|
{
|
*(pu1_dst + rows * dst_strd + cols) = *pu1_left;
|
|
*(pu1_dst + rows * dst_strd + cols + 1) = *(pu1_left + 1);
|
}
|
pu1_left -= 2;
|
}
|
|
}
|
|
/**
|
*******************************************************************************
|
*
|
*ih264_intra_pred_chroma_8x8_mode_vert
|
*
|
* @brief
|
* Perform Intra prediction for chroma_8x8 mode:vertical
|
*
|
* @par Description:
|
* Perform Intra prediction for chroma_8x8 mode:vertical ,described in sec 8.3.4.3
|
*
|
* @param[in] pu1_src
|
* UWORD8 pointer to the source containing alternate U and V samples
|
*
|
* @param[out] pu1_dst
|
* UWORD8 pointer to the destination with alternate U and V samples
|
*
|
* @param[in] src_strd
|
* integer source stride
|
*
|
* @param[in] dst_strd
|
* integer destination stride
|
*
|
* @param[in] ngbr_avail
|
* availability of neighbouring pixels(Not used in this function)
|
*
|
* @returns
|
*
|
* @remarks
|
* None
|
*
|
*******************************************************************************
|
*/
|
void ih264_intra_pred_chroma_8x8_mode_vert(UWORD8 *pu1_src,
|
UWORD8 *pu1_dst,
|
WORD32 src_strd,
|
WORD32 dst_strd,
|
WORD32 ngbr_avail)
|
{
|
|
UWORD8 *pu1_top = NULL; /* Pointer to start of top predictors */
|
WORD32 row;/*loop variable*/
|
UNUSED(src_strd);
|
UNUSED(ngbr_avail);
|
pu1_top = pu1_src + 2 * BLK8x8SIZE + 2;
|
|
/* 8 bytes are copied from src to dst */
|
for(row = 0; row < 2; row++)
|
{
|
memcpy(pu1_dst, pu1_top, 16);
|
|
pu1_dst += dst_strd;
|
memcpy(pu1_dst, pu1_top, 16);
|
|
pu1_dst += dst_strd;
|
memcpy(pu1_dst, pu1_top, 16);
|
|
pu1_dst += dst_strd;
|
memcpy(pu1_dst, pu1_top, 16);
|
|
pu1_dst += dst_strd;
|
}
|
}
|
|
/**
|
*******************************************************************************
|
*
|
* ih264_intra_pred_chroma_8x8_mode_plane
|
*
|
* @brief
|
* Perform Intra prediction for chroma_8x8 mode:PLANE
|
*
|
* @par Description:
|
* Perform Intra prediction for chroma_8x8 mode:PLANE ,described in sec 8.3.4.4
|
*
|
* @param[in] pu1_src
|
* UWORD8 pointer to the source containing alternate U and V samples
|
*
|
* @param[out] pu1_dst
|
* UWORD8 pointer to the destination with alternate U and V samples
|
*
|
* @param[in] src_strd
|
* integer source stride
|
*
|
* @param[in] dst_strd
|
* integer destination stride
|
*
|
* @param[in] ngbr_avail
|
* availability of neighbouring pixels(Not used in this function)
|
*
|
* @returns
|
*
|
* @remarks
|
* None
|
*
|
******************************************************************************
|
*/
|
void ih264_intra_pred_chroma_8x8_mode_plane(UWORD8 *pu1_src,
|
UWORD8 *pu1_dst,
|
WORD32 src_strd,
|
WORD32 dst_strd,
|
WORD32 ngbr_avail)
|
{
|
|
UWORD8 *pu1_left = NULL; /* Pointer to start of left predictors */
|
UWORD8 *pu1_top = NULL; /* Pointer to start of top predictors */
|
WORD32 val = 0;
|
WORD32 rows, cols; /* loop variables*/
|
WORD32 a_u, b_u, c_u, h_u, v_u; /* Implementing section 8.3.4.4 . The variables represent the corresponding variables in the section*/
|
WORD32 a_v, b_v, c_v, h_v, v_v;
|
UNUSED(src_strd);
|
UNUSED(ngbr_avail);
|
a_u = b_u = c_u = h_u = v_u = 0;
|
a_v = b_v = c_v = h_v = v_v = 0;
|
/* As chroma format 4:2:0 is used,xCF = 4 * ( chroma_format_idc = = 3 ) = 0 and
|
yCF = 4 * ( chroma_format_idc != 1 ) = 0 */
|
pu1_top = pu1_src + 2 * BLK8x8SIZE + 2;
|
pu1_left = pu1_src + 2 * BLK8x8SIZE - 2;
|
/* Implementing section 8.3.4.4 */
|
for(cols = 0; cols < 4; cols++)
|
{
|
h_u += (cols + 1) * (pu1_top[8 + 2 * cols] - pu1_top[4 - 2 * cols]);/*section 8.3.4.4 equation (8-144)*/
|
h_v += (cols + 1) * (pu1_top[8 + 2 * cols + 1] - pu1_top[4 - 2 * cols+ 1]);
|
|
v_u += (cols + 1) * (pu1_left[(4 + cols) * (-2)] - pu1_left[(2 - cols) * (-2)]);
|
v_v += (cols + 1) * (pu1_left[(4 + cols) * (-2) + 1] - pu1_left[(2 - cols) * (-2) + 1]);/*section 8.3.4.4 equation (8-145)*/
|
}
|
a_u = 16 * (pu1_left[7 * (-2)] + pu1_top[14]);
|
a_v = 16 * (pu1_left[7 * (-2) + 1] + pu1_top[15]);/*section 8.3.3.4 equation (8-141)*/
|
b_u = (34 * h_u + 32) >> 6;/*section 8.3.3.4 equation (8-142)*/
|
b_v = (34 * h_v + 32) >> 6;/*section 8.3.3.4 equation (8-142)*/
|
c_u = (34 * v_u + 32) >> 6;/*section 8.3.3.4 equation (8-143)*/
|
c_v = (34 * v_v + 32) >> 6;/*section 8.3.3.4 equation (8-143)*/
|
|
for(rows = 0; rows < 8; rows++)
|
{
|
for(cols = 0; cols < 8; cols++)
|
{
|
val = (a_u + b_u * (cols - 3) + c_u * (rows - 3) );/*section 8.3.4.4 equation (8-140)*/
|
val = (val + 16) >> 5;
|
*(pu1_dst + rows * dst_strd + 2 * cols) = CLIP_U8(val);
|
val = (a_v + b_v * (cols - 3) + c_v * (rows - 3) );/*section 8.3.4.4 equation (8-140)*/
|
val = (val + 16) >> 5;
|
*(pu1_dst + rows * dst_strd + 2 * cols + 1) = CLIP_U8(val);
|
}
|
}
|
}
|