www.pudn.com > T264-src-0.02.zip > dct.c
/*****************************************************************************
*
* T264 AVC CODEC
*
* Copyright(C) 2004-2005 llcc
* 2004-2005 visionany
*
* This program is free software ; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation ; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY ; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program ; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*
****************************************************************************/
#include "stdio.h"
#include "T264.h"
#define Q_BITS 15
#define DQ_BITS 6
#define DQ_ROUND (1<<(DQ_BITS-1))
//////////////////////////////////////////////////////
// static var
DECLARE_ALIGNED2_MATRIX_H(quant, 6, 4 * 4, int16_t, CACHE_SIZE) =
{
13107, 8066, 13107, 8066, 8066, 5243, 8066, 5243,
13107, 8066, 13107, 8066, 8066, 5243, 8066, 5243,
11916, 7490, 11916, 7490, 7490, 4660, 7490, 4660,
11916, 7490, 11916, 7490, 7490, 4660, 7490, 4660,
10082, 6554, 10082, 6554, 6554, 4194, 6554, 4194,
10082, 6554, 10082, 6554, 6554, 4194, 6554, 4194,
9362, 5825, 9362, 5825, 5825, 3647, 5825, 3647,
9362, 5825, 9362, 5825, 5825, 3647, 5825, 3647,
8192, 5243, 8192, 5243, 5243, 3355, 5243, 3355,
8192, 5243, 8192, 5243, 5243, 3355, 5243, 3355,
7282, 4559, 7282, 4559, 4559, 2893, 4559, 2893,
7282, 4559, 7282, 4559, 4559, 2893, 4559, 2893
};
DECLARE_ALIGNED2_MATRIX_H(dequant, 6, 4 * 4, int16_t, CACHE_SIZE) =
{
10, 13, 10, 13, 13, 16, 13, 16, 10, 13, 10, 13, 13, 16, 13, 16,
11, 14, 11, 14, 14, 18, 14, 18, 11, 14, 11, 14, 14, 18, 14, 18,
13, 16, 13, 16, 16, 20, 16, 20, 13, 16, 13, 16, 16, 20, 16, 20,
14, 18, 14, 18, 18, 23, 18, 23, 14, 18, 14, 18, 18, 23, 18, 23,
16, 20, 16, 20, 20, 25, 20, 25, 16, 20, 16, 20, 20, 25, 20, 25,
18, 23, 18, 23, 23, 29, 23, 29, 18, 23, 18, 23, 23, 29, 23, 29
};
//////////////////////////////////////////////////////
// DCT & IDCT
void
dct4x4_c(int16_t* data)
{
int32_t i;
int16_t s[4];
//
// horizontal
//
for(i = 0 ; i < 4 ; i ++)
{
s[0] = data[i * 4 + 0] + data[i * 4 + 3];
s[3] = data[i * 4 + 0] - data[i * 4 + 3];
s[1] = data[i * 4 + 1] + data[i * 4 + 2];
s[2] = data[i * 4 + 1] - data[i * 4 + 2];
data[i * 4 + 0] = s[0] + s[1];
data[i * 4 + 2] = s[0] - s[1];
data[i * 4 + 1] = (s[3] << 1) + s[2];
data[i * 4 + 3] = s[3] - (s[2] << 1);
}
//
// vertical
//
for(i = 0 ; i < 4 ; i ++)
{
s[0] = data[0 * 4 + i] + data[3 * 4 + i];
s[3] = data[0 * 4 + i] - data[3 * 4 + i];
s[1] = data[1 * 4 + i] + data[2 * 4 + i];
s[2] = data[1 * 4 + i] - data[2 * 4 + i];
data[0 * 4 + i] = s[0] + s[1];
data[2 * 4 + i] = s[0] - s[1];
data[1 * 4 + i] = (s[3] << 1) + s[2];
data[3 * 4 + i] = s[3] - (s[2] << 1);
}
}
void
dct4x4dc_c(int16_t* data)
{
int32_t i;
int16_t s[4];
for(i = 0 ; i < 4 ; i ++)
{
s[0] = data[i * 4 + 0] + data[i * 4 + 3];
s[3] = data[i * 4 + 0] - data[i * 4 + 3];
s[1] = data[i * 4 + 1] + data[i * 4 + 2];
s[2] = data[i * 4 + 1] - data[i * 4 + 2];
data[i * 4 + 0] = s[0] + s[1];
data[i * 4 + 2] = s[0] - s[1];
data[i * 4 + 1] = s[3] + s[2];
data[i * 4 + 3] = s[3] - s[2];
}
for(i = 0 ; i < 4 ; i ++)
{
s[0] = data[0 * 4 + i] + data[3 * 4 + i];
s[3] = data[0 * 4 + i] - data[3 * 4 + i];
s[1] = data[1 * 4 + i] + data[2 * 4 + i];
s[2] = data[1 * 4 + i] - data[2 * 4 + i];
data[0 * 4 + i] = (s[0] + s[1] + 1) >> 1;
data[2 * 4 + i] = (s[0] - s[1] + 1) >> 1;
data[1 * 4 + i] = (s[3] + s[2] + 1) >> 1;
data[3 * 4 + i] = (s[3] - s[2] + 1) >> 1;
}
}
void
dct2x2dc_c(int16_t* data)
{
int16_t s[4];
s[0] = data[0];
s[1] = data[1];
s[2] = data[2];
s[3] = data[3];
data[0] = s[0] + s[2] + s[1] + s[3];
data[1] = s[0] + s[2] - s[1] - s[3];
data[2] = s[0] - s[2] + s[1] - s[3];
data[3] = s[0] - s[2] - s[1] + s[3];
}
void
idct4x4_c(int16_t* data)
{
int32_t i;
int16_t s[4];
for (i = 0; i < 4; i ++)
{
s[0] = data[i * 4 + 0] + data[i * 4 + 2];
s[1] = data[i * 4 + 0] - data[i * 4 + 2];
s[2] = (data[i * 4 + 1] >> 1) - data[i * 4 + 3];
s[3] = data[i * 4 + 1] + (data[i * 4 + 3] >> 1);
data[i * 4 + 0] = s[0] + s[3];
data[i * 4 + 3] = s[0] - s[3];
data[i * 4 + 1] = s[1] + s[2];
data[i * 4 + 2] = s[1] - s[2];
}
for (i = 0; i < 4; i ++)
{
s[0] = data[0 * 4 + i] + data[2 * 4 + i];
s[1] = data[0 * 4 + i] - data[2 * 4 + i];
s[2] = (data[1 * 4 + i] >> 1) - data[3 * 4 + i];
s[3] = data[1 * 4 + i] + (data[3 * 4 + i] >> 1);
data[0 * 4 + i] = (s[0] + s[3] + 32) >> 6;
data[3 * 4 + i] = (s[0] - s[3] + 32) >> 6;
data[1 * 4 + i] = (s[1] + s[2] + 32) >> 6;
data[2 * 4 + i] = (s[1] - s[2] + 32) >> 6;
}
}
void
idct4x4dc_c(int16_t* data)
{
int32_t i;
int16_t s[4];
for (i = 0; i < 4; i ++)
{
s[0] = data[i * 4 + 0] + data[i * 4 + 2];
s[1] = data[i * 4 + 0] - data[i * 4 + 2];
s[2] = data[i * 4 + 1] - data[i * 4 + 3];
s[3] = data[i * 4 + 1] + data[i * 4 + 3];
data[i * 4 + 0] = s[0] + s[3];
data[i * 4 + 3] = s[0] - s[3];
data[i * 4 + 1] = s[1] + s[2];
data[i * 4 + 2] = s[1] - s[2];
}
for (i = 0; i < 4; i ++)
{
s[0] = data[0 * 4 + i] + data[2 * 4 + i];
s[1] = data[0 * 4 + i] - data[2 * 4 + i];
s[2] = data[1 * 4 + i] - data[3 * 4 + i];
s[3] = data[1 * 4 + i] + data[3 * 4 + i];
data[0 * 4 + i] = s[0] + s[3];
data[3 * 4 + i] = s[0] - s[3];
data[1 * 4 + i] = s[1] + s[2];
data[2 * 4 + i] = s[1] - s[2];
}
}
void
idct2x2dc_c(int16_t* data)
{
int16_t s[4];
s[0] = data[0];
s[1] = data[1];
s[2] = data[2];
s[3] = data[3];
data[0] = s[0] + s[2] + s[1] + s[3];
data[1] = s[0] + s[2] - s[1] - s[3];
data[2] = s[0] - s[2] + s[1] - s[3];
data[3] = s[0] - s[2] - s[1] + s[3];
}
///////////////////////////////////////////////////////////
// Quant & IQuant
void
quant4x4_c(int16_t* data, const int32_t Qp, int32_t is_intra)
{
const int32_t qbits = 15 + Qp / 6;
const int32_t mf_index = Qp % 6;
int32_t i;
const int32_t f = (1 << qbits) / (is_intra ? 3 : 6);
for(i = 0 ; i < 16 ; i ++)
{
if (data[i] > 0)
data[i] = (data[i] * quant[mf_index][i] + f) >> qbits;
else
data[i] = -((-(data[i] * quant[mf_index][i]) + f) >> qbits);
}
}
void
quant4x4dc_c(int16_t* data, const int32_t Qp)
{
const int32_t qbits = 15 + Qp / 6;
const int32_t mf_index = Qp % 6;
const int32_t mf00 = quant[mf_index][0];
const int32_t f2 = (2 << qbits) / 3; // Only 16x16 intra mode
int32_t i;
for(i = 0 ; i < 16 ; i ++)
{
if (data[i] > 0)
data[i] = (data[i] * mf00 + f2) >> (qbits + 1);
else
data[i] = -((-(data[i] * mf00) + f2) >> (qbits + 1));
}
}
void
quant2x2dc_c(int16_t* data, const int32_t Qp, int32_t is_intra)
{
const int32_t qbits = 15 + Qp / 6;
const int32_t mf_index = Qp % 6;
const int32_t mf00 = quant[mf_index][0];
const int32_t f2 = (2 << qbits) / (is_intra ? 3 : 6);
int32_t i;
for(i = 0 ; i < 4 ; i ++)
{
if (data[i] > 0)
data[i] = (data[i] * mf00 + f2) >> (qbits + 1);
else
data[i] = -((-(data[i] * mf00) + f2) >> (qbits + 1));
}
}
void
iquant4x4_c(int16_t* data, const int32_t Qp)
{
const int32_t qbits = Qp / 6;
const int32_t index_mf = Qp % 6;
int32_t i;
for(i = 0 ; i < 16 ; i ++)
{
data[i] = (data[i] * dequant[index_mf][i]) << qbits;
}
}
void
iquant4x4dc_c(int16_t* data, const int32_t Qp)
{
const int32_t qbits = Qp / 6 - 2;
int32_t i;
if (qbits >= 0)
{
const int32_t mf_index = Qp % 6;
const int32_t dmf = dequant[mf_index][0] << qbits;
for(i = 0 ; i < 16 ; i ++)
{
data[i] = data[i] * dmf;
}
}
else
{
const int32_t dmf = dequant[Qp % 6][0];
const int32_t t2 = 1 << (1 + qbits);
for(i = 0 ; i < 16 ; i ++)
{
data[i] = (data[i] * dmf + t2) >> (-qbits);
}
}
}
void
iquant2x2dc_c(int16_t* data, const int32_t Qp)
{
const int32_t qbits = Qp / 6 - 1;
if (qbits >= 0 )
{
const int32_t dmf = dequant[Qp % 6][0] << qbits;
data[0] = data[0] * dmf;
data[1] = data[1] * dmf;
data[2] = data[2] * dmf;
data[3] = data[3] * dmf;
}
else
{
const int32_t dmf = dequant[Qp % 6][0];
data[0] = (data[0] * dmf) >> 1;
data[1] = (data[1] * dmf) >> 1;
data[2] = (data[2] * dmf) >> 1;
data[3] = (data[3] * dmf) >> 1;
}
}