mozilla
/
aom
зеркало из
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность

Enable hybrid 1-D/2-D transform coding for highbd setting 

This commit enables the hybrid 1-D/2-D transform coding scheme for
high bit-depth setting. It improves the compression performance of
ext-tx experiment by 0.98% for lowres_all set.

Change-Id: Ic27f5037f2c36b095a93b9f15dbae34bdcdf00aa
This commit is contained in:
Jingning Han 2016-03-09 16:40:08 -08:00
Родитель e0413094fb
Коммит c453ae53d0
2 изменённых файлов: 88 добавлений и 5 удалений
Показать статистику Скачать Patch файл Скачать Diff файл Показать все файлы Свернуть все файлы

85
vp10/common/idct.c
Просмотреть файл

@ -700,11 +700,78 @@ void highbd_idst16_c(const tran_low_t *input, tran_low_t *output, int bd) {
}
static void highbd_inv_idtx_add_c(const tran_low_t *input, uint8_t *dest8,
int stride, int bs, int bd) {
int stride, int bs, int tx_type, int bd) {
int r, c;
const int shift = bs < 32 ? 3 : 2;
uint16_t *dest = CONVERT_TO_SHORTPTR(dest8);
tran_low_t temp_in[32], temp_out[32];
highbd_transform_2d ht = {vpx_highbd_idct4_c, vpx_highbd_idct4_c};
int out_scale = 1;
int coeff_stride = 0;
switch (bs) {
case 4:
ht.cols = vpx_highbd_idct4_c;
ht.rows = vpx_highbd_idct4_c;
out_scale = cospi_16_64 >> 3;
coeff_stride = 4;
break;
case 8:
ht.cols = vpx_highbd_idct8_c;
ht.rows = vpx_highbd_idct8_c;
out_scale = (1 << (DCT_CONST_BITS - 4));
coeff_stride = 8;
break;
case 16:
ht.cols = vpx_highbd_idct16_c;
ht.rows = vpx_highbd_idct16_c;
out_scale = cospi_16_64 >> 4;
coeff_stride = 16;
break;
case 32:
ht.cols = vpx_highbd_idct32_c;
ht.rows = vpx_highbd_idct32_c;
out_scale = (1 << (DCT_CONST_BITS - 4));
coeff_stride = 32;
break;
default:
assert(0);
}
// Columns
if (tx_type == V_DCT) {
for (c = 0; c < bs; ++c) {
for (r = 0; r < bs; ++r)
temp_in[r] = input[r * coeff_stride + c];
ht.cols(temp_in, temp_out, bd);
for (r = 0; r < bs; ++r) {
tran_high_t temp = (tran_high_t)temp_out[r] * out_scale;
temp >>= DCT_CONST_BITS;
dest[r * stride + c] = highbd_clip_pixel_add(dest[r * stride + c],
(tran_low_t)temp, bd);
}
}
return;
}
if (tx_type == H_DCT) {
for (r = 0; r < bs; ++r) {
for (c = 0; c < bs; ++c)
temp_in[c] = input[r * coeff_stride + c];
ht.rows(temp_in, temp_out, bd);
for (c = 0; c < bs; ++c) {
tran_high_t temp = (tran_high_t)temp_out[c] * out_scale;
temp >>= DCT_CONST_BITS;
dest[r * stride + c] = highbd_clip_pixel_add(dest[r * stride + c],
(tran_low_t)temp, bd);
}
}
return;
}
for (r = 0; r < bs; ++r) {
for (c = 0; c < bs; ++c)
dest[c] = highbd_clip_pixel_add(dest[c], input[c] >> shift, bd);
@ -1593,8 +1660,10 @@ void vp10_highbd_inv_txfm_add_4x4(const tran_low_t *input, uint8_t *dest,
// Use C version since DST only exists in C code
vp10_highbd_iht4x4_16_add_c(input, dest, stride, tx_type, bd);
break;
case H_DCT:
case V_DCT:
case IDTX:
highbd_inv_idtx_add_c(input, dest, stride, 4, bd);
highbd_inv_idtx_add_c(input, dest, stride, 4, tx_type, bd);
break;
#endif // CONFIG_EXT_TX
default:
@ -1633,8 +1702,10 @@ void vp10_highbd_inv_txfm_add_8x8(const tran_low_t *input, uint8_t *dest,
// Use C version since DST only exists in C code
vp10_highbd_iht8x8_64_add_c(input, dest, stride, tx_type, bd);
break;
case H_DCT:
case V_DCT:
case IDTX:
highbd_inv_idtx_add_c(input, dest, stride, 8, bd);
highbd_inv_idtx_add_c(input, dest, stride, 8, tx_type, bd);
break;
#endif // CONFIG_EXT_TX
default:
@ -1673,8 +1744,10 @@ void vp10_highbd_inv_txfm_add_16x16(const tran_low_t *input, uint8_t *dest,
// Use C version since DST only exists in C code
vp10_highbd_iht16x16_256_add_c(input, dest, stride, tx_type, bd);
break;
case H_DCT:
case V_DCT:
case IDTX:
highbd_inv_idtx_add_c(input, dest, stride, 16, bd);
highbd_inv_idtx_add_c(input, dest, stride, 16, tx_type, bd);
break;
#endif // CONFIG_EXT_TX
default:
@ -1708,8 +1781,10 @@ void vp10_highbd_inv_txfm_add_32x32(const tran_low_t *input, uint8_t *dest,
case DST_FLIPADST:
vp10_highbd_iht32x32_1024_add_c(input, dest, stride, tx_type, bd);
break;
case H_DCT:
case V_DCT:
case IDTX:
highbd_inv_idtx_add_c(input, dest, stride, 32, bd);
highbd_inv_idtx_add_c(input, dest, stride, 32, tx_type, bd);
break;
#endif // CONFIG_EXT_TX
default:

8
vp10/encoder/hybrid_fwd_txfm.c
Просмотреть файл

@ -232,6 +232,8 @@ void vp10_highbd_fwd_txfm_4x4(const int16_t *src_diff, tran_low_t *coeff,
// Use C version since DST exists only in C
vp10_highbd_fht4x4_c(src_diff, coeff, diff_stride, tx_type);
break;
case H_DCT:
case V_DCT:
case IDTX:
vp10_fwd_idtx_c(src_diff, coeff, diff_stride, 4, tx_type);
break;
@ -274,6 +276,8 @@ static void highbd_fwd_txfm_8x8(const int16_t *src_diff, tran_low_t *coeff,
// Use C version since DST exists only in C
vp10_highbd_fht8x8_c(src_diff, coeff, diff_stride, tx_type);
break;
case H_DCT:
case V_DCT:
case IDTX:
vp10_fwd_idtx_c(src_diff, coeff, diff_stride, 8, tx_type);
break;
@ -316,6 +320,8 @@ static void highbd_fwd_txfm_16x16(const int16_t *src_diff, tran_low_t *coeff,
// Use C version since DST exists only in C
vp10_highbd_fht16x16_c(src_diff, coeff, diff_stride, tx_type);
break;
case H_DCT:
case V_DCT:
case IDTX:
vp10_fwd_idtx_c(src_diff, coeff, diff_stride, 16, tx_type);
break;
@ -354,6 +360,8 @@ static void highbd_fwd_txfm_32x32(int rd_transform, const int16_t *src_diff,
case FLIPADST_DST:
vp10_highbd_fht32x32_c(src_diff, coeff, diff_stride, tx_type);
break;
case H_DCT:
case V_DCT:
case IDTX:
vp10_fwd_idtx_c(src_diff, coeff, diff_stride, 32, tx_type);
break;