generic-library/vpx
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
vpx/vp9/common/vp9_idct.h
Debargha Mukherjee e6790e30c5 Replace DST1 with DST2 for ext-tx experiment 
A small gain (0.1 - 0.2%) with this experiment on derflr/hevcmr.

The DST2 can be implemened very efficiently using sign flipping
of odd indexed inputs, followed by DCT, followed by reversal of
the output. This is how it is implemented in this patch.
SIMD optimization is pending.

Change-Id: Ic2fc211ce0e6b7c6702974d76d6573f55cc4da0e
2015-12-14 13:54:41 -08:00

464 lines
18 KiB
C
Raw Blame History

/*
* Copyright (c) 2010 The WebM project authors. All Rights Reserved.
*
* Use of this source code is governed by a BSD-style license
* that can be found in the LICENSE file in the root of the source
* tree. An additional intellectual property rights grant can be found
* in the file PATENTS. All contributing project authors may
* be found in the AUTHORS file in the root of the source tree.
*/
#ifndef VP9_COMMON_VP9_IDCT_H_
#define VP9_COMMON_VP9_IDCT_H_
#include <assert.h>
#include "./vpx_config.h"
#include "vp9/common/vp9_common.h"
#include "vp9/common/vp9_enums.h"
#ifdef __cplusplus
extern "C" {
#endif
// Constants and Macros used by all idct/dct functions
#define DCT_CONST_BITS 14
#define DCT_CONST_ROUNDING (1 << (DCT_CONST_BITS - 1))
#define UNIT_QUANT_SHIFT 2
#define UNIT_QUANT_FACTOR (1 << UNIT_QUANT_SHIFT)
#define pair_set_epi16(a, b) \
_mm_set_epi16(b, a, b, a, b, a, b, a)
#define dual_set_epi16(a, b) \
_mm_set_epi16(b, b, b, b, a, a, a, a)
#if CONFIG_TX_SKIP
#define TX_SKIP_SHIFT_LQ 2
#define TX_SKIP_SHIFT_HQ 3
#endif
// Constants:
// for (int i = 1; i< 32; ++i)
// printf("static const int cospi_%d_64 = %.0f;\n", i,
// round(16384 * cos(i*M_PI/64)));
// Note: sin(k*Pi/64) = cos((32-k)*Pi/64)
static const tran_high_t cospi_1_64 = 16364;
static const tran_high_t cospi_2_64 = 16305;
static const tran_high_t cospi_3_64 = 16207;
static const tran_high_t cospi_4_64 = 16069;
static const tran_high_t cospi_5_64 = 15893;
static const tran_high_t cospi_6_64 = 15679;
static const tran_high_t cospi_7_64 = 15426;
static const tran_high_t cospi_8_64 = 15137;
static const tran_high_t cospi_9_64 = 14811;
static const tran_high_t cospi_10_64 = 14449;
static const tran_high_t cospi_11_64 = 14053;
static const tran_high_t cospi_12_64 = 13623;
static const tran_high_t cospi_13_64 = 13160;
static const tran_high_t cospi_14_64 = 12665;
static const tran_high_t cospi_15_64 = 12140;
static const tran_high_t cospi_16_64 = 11585;
static const tran_high_t cospi_17_64 = 11003;
static const tran_high_t cospi_18_64 = 10394;
static const tran_high_t cospi_19_64 = 9760;
static const tran_high_t cospi_20_64 = 9102;
static const tran_high_t cospi_21_64 = 8423;
static const tran_high_t cospi_22_64 = 7723;
static const tran_high_t cospi_23_64 = 7005;
static const tran_high_t cospi_24_64 = 6270;
static const tran_high_t cospi_25_64 = 5520;
static const tran_high_t cospi_26_64 = 4756;
static const tran_high_t cospi_27_64 = 3981;
static const tran_high_t cospi_28_64 = 3196;
static const tran_high_t cospi_29_64 = 2404;
static const tran_high_t cospi_30_64 = 1606;
static const tran_high_t cospi_31_64 = 804;
// 16384 * sqrt(2) * sin(kPi/9) * 2 / 3
static const tran_high_t sinpi_1_9 = 5283;
static const tran_high_t sinpi_2_9 = 9929;
static const tran_high_t sinpi_3_9 = 13377;
static const tran_high_t sinpi_4_9 = 15212;
static INLINE tran_low_t check_range(tran_high_t input) {
#if CONFIG_VP9_HIGHBITDEPTH
// For valid highbitdepth VP9 streams, intermediate stage coefficients will
// stay within the ranges:
// - 8 bit: signed 16 bit integer
// - 10 bit: signed 18 bit integer
// - 12 bit: signed 20 bit integer
#elif CONFIG_COEFFICIENT_RANGE_CHECKING
// For valid VP9 input streams, intermediate stage coefficients should always
// stay within the range of a signed 16 bit integer. Coefficients can go out
// of this range for invalid/corrupt VP9 streams. However, strictly checking
// this range for every intermediate coefficient can burdensome for a decoder,
// therefore the following assertion is only enabled when configured with
// --enable-coefficient-range-checking.
assert(INT16_MIN <= input);
assert(input <= INT16_MAX);
#endif
return (tran_low_t)input;
}
static INLINE tran_low_t dct_const_round_shift(tran_high_t input) {
tran_high_t rv = ROUND_POWER_OF_TWO(input, DCT_CONST_BITS);
return check_range(rv);
}
typedef void (*transform_1d)(const tran_low_t*, tran_low_t*);
typedef struct {
transform_1d cols, rows; // vertical and horizontal
} transform_2d;
#if CONFIG_VP9_HIGHBITDEPTH
typedef void (*highbd_transform_1d)(const tran_low_t*, tran_low_t*, int bd);
typedef struct {
highbd_transform_1d cols, rows; // vertical and horizontal
} highbd_transform_2d;
#endif // CONFIG_VP9_HIGHBITDEPTH
#if CONFIG_EMULATE_HARDWARE
// When CONFIG_EMULATE_HARDWARE is 1 the transform performs a
// non-normative method to handle overflows. A stream that causes
// overflows in the inverse transform is considered invalid in VP9,
// and a hardware implementer is free to choose any reasonable
// method to handle overflows. However to aid in hardware
// verification they can use a specific implementation of the
// WRAPLOW() macro below that is identical to their intended
// hardware implementation (and also use configure options to trigger
// the C-implementation of the transform).
//
// The particular WRAPLOW implementation below performs strict
// overflow wrapping to match common hardware implementations.
// bd of 8 uses trans_low with 16bits, need to remove 16bits
// bd of 10 uses trans_low with 18bits, need to remove 14bits
// bd of 12 uses trans_low with 20bits, need to remove 12bits
// bd of x uses trans_low with 8+x bits, need to remove 24-x bits
#define WRAPLOW(x, bd) ((((int32_t)(x)) << (24 - bd)) >> (24 - bd))
#else
#define WRAPLOW(x, bd) (x)
#endif // CONFIG_EMULATE_HARDWARE
static INLINE uint8_t clip_pixel_add(uint8_t dest, tran_high_t trans) {
trans = WRAPLOW(trans, 8);
return clip_pixel(WRAPLOW(dest + trans, 8));
}
#if CONFIG_EXT_TX
#define USE_DST2 1
#if USE_DST2
static const tran_high_t Tx4[4 * 4] = {
// dst2
6270, 15137, 15137, 6270,
11585, 11585, -11585, -11585,
15137, -6270, -6270, 15137,
11585, -11585, 11585, -11585,
};
static const tran_high_t Tx8[8 * 8] = {
// dst2
3196, 9102, 13623, 16069, 16069, 13623, 9102, 3196,
6270, 15137, 15137, 6270, -6270, -15137, -15137, -6270,
9102, 16069, 3196, -13623, -13623, 3196, 16069, 9102,
11585, 11585, -11585, -11585, 11585, 11585, -11585, -11585,
13623, 3196, -16069, 9102, 9102, -16069, 3196, 13623,
15137, -6270, -6270, 15137, -15137, 6270, 6270, -15137,
16069, -13623, 9102, -3196, -3196, 9102, -13623, 16069,
11585, -11585, 11585, -11585, 11585, -11585, 11585, -11585,
};
static const tran_high_t Tx16[16 * 16] = {
// dst2
1606, 4756, 7723, 10394, 12665, 14449, 15679, 16305,
16305, 15679, 14449, 12665, 10394, 7723, 4756, 1606,
3196, 9102, 13623, 16069, 16069, 13623, 9102, 3196,
-3196, -9102, -13623, -16069, -16069, -13623, -9102, -3196,
4756, 12665, 16305, 14449, 7723, -1606, -10394, -15679,
-15679, -10394, -1606, 7723, 14449, 16305, 12665, 4756,
6270, 15137, 15137, 6270, -6270, -15137, -15137, -6270,
6270, 15137, 15137, 6270, -6270, -15137, -15137, -6270,
7723, 16305, 10394, -4756, -15679, -12665, 1606, 14449,
14449, 1606, -12665, -15679, -4756, 10394, 16305, 7723,
9102, 16069, 3196, -13623, -13623, 3196, 16069, 9102,
-9102, -16069, -3196, 13623, 13623, -3196, -16069, -9102,
10394, 14449, -4756, -16305, -1606, 15679, 7723, -12665,
-12665, 7723, 15679, -1606, -16305, -4756, 14449, 10394,
11585, 11585, -11585, -11585, 11585, 11585, -11585, -11585,
11585, 11585, -11585, -11585, 11585, 11585, -11585, -11585,
12665, 7723, -15679, -1606, 16305, -4756, -14449, 10394,
10394, -14449, -4756, 16305, -1606, -15679, 7723, 12665,
13623, 3196, -16069, 9102, 9102, -16069, 3196, 13623,
-13623, -3196, 16069, -9102, -9102, 16069, -3196, -13623,
14449, -1606, -12665, 15679, -4756, -10394, 16305, -7723,
-7723, 16305, -10394, -4756, 15679, -12665, -1606, 14449,
15137, -6270, -6270, 15137, -15137, 6270, 6270, -15137,
15137, -6270, -6270, 15137, -15137, 6270, 6270, -15137,
15679, -10394, 1606, 7723, -14449, 16305, -12665, 4756,
4756, -12665, 16305, -14449, 7723, 1606, -10394, 15679,
16069, -13623, 9102, -3196, -3196, 9102, -13623, 16069,
-16069, 13623, -9102, 3196, 3196, -9102, 13623, -16069,
16305, -15679, 14449, -12665, 10394, -7723, 4756, -1606,
-1606, 4756, -7723, 10394, -12665, 14449, -15679, 16305,
11585, -11585, 11585, -11585, 11585, -11585, 11585, -11585,
11585, -11585, 11585, -11585, 11585, -11585, 11585, -11585,
};
#endif // USE_DST2
static INLINE void vp9_fgentx4(const tran_low_t *input, tran_low_t *output,
const tran_high_t *T) {
tran_high_t sum;
int i;
const tran_high_t *Tx = T;
for (i = 0; i < 4; ++i, Tx += 4) {
sum = Tx[0] * input[0] + Tx[1] * input[1] +
Tx[2] * input[2] + Tx[3] * input[3];
output[i] = ROUND_POWER_OF_TWO(sum, DCT_CONST_BITS);
}
}
static INLINE void vp9_fgentx8(const tran_low_t *input, tran_low_t *output,
const tran_high_t *T) {
tran_high_t sum;
int i;
const tran_high_t *Tx = T;
for (i = 0; i < 8; ++i, Tx += 8) {
sum = Tx[0] * input[0] + Tx[1] * input[1] +
Tx[2] * input[2] + Tx[3] * input[3] +
Tx[4] * input[4] + Tx[5] * input[5] +
Tx[6] * input[6] + Tx[7] * input[7];
output[i] = ROUND_POWER_OF_TWO(sum, DCT_CONST_BITS);
}
}
static INLINE void vp9_fgentx16(const tran_low_t *input, tran_low_t *output,
const tran_high_t *T) {
tran_high_t sum;
int i;
const tran_high_t *Tx = T;
for (i = 0; i < 16; ++i, Tx += 16) {
sum = Tx[0] * input[0] + Tx[1] * input[1] +
Tx[2] * input[2] + Tx[3] * input[3] +
Tx[4] * input[4] + Tx[5] * input[5] +
Tx[6] * input[6] + Tx[7] * input[7] +
Tx[8] * input[8] + Tx[9] * input[9] +
Tx[10] * input[10] + Tx[11] * input[11] +
Tx[12] * input[12] + Tx[13] * input[13] +
Tx[14] * input[14] + Tx[15] * input[15];
output[i] = ROUND_POWER_OF_TWO(sum, DCT_CONST_BITS);
}
}
static INLINE void vp9_igentx4(const tran_low_t *input, tran_low_t *output,
const tran_high_t *T) {
tran_high_t sum[4];
int i;
const tran_high_t *Tx = T;
for (i = 0; i < 4; ++i, ++Tx) {
sum[i] = Tx[0] * input[0] + Tx[4] * input[1] +
Tx[8] * input[2] + Tx[12] * input[3];
}
for (i = 0; i < 4; ++i) {
output[i] = WRAPLOW(ROUND_POWER_OF_TWO(sum[i], DCT_CONST_BITS), 8);
}
}
static INLINE void vp9_igentx8(const tran_low_t *input, tran_low_t *output,
const tran_high_t *T) {
tran_high_t sum[8];
int i;
const tran_high_t *Tx = T;
for (i = 0; i < 8; ++i, ++Tx) {
sum[i] = Tx[0] * input[0] + Tx[8] * input[1] +
Tx[16] * input[2] + Tx[24] * input[3] +
Tx[32] * input[4] + Tx[40] * input[5] +
Tx[48] * input[6] + Tx[56] * input[7];
}
for (i = 0; i < 8; ++i) {
output[i] = WRAPLOW(ROUND_POWER_OF_TWO(sum[i], DCT_CONST_BITS), 8);
}
}
static INLINE void vp9_igentx16(const tran_low_t *input, tran_low_t *output,
const tran_high_t *T) {
tran_high_t sum[16];
int i;
const tran_high_t *Tx = T;
for (i = 0; i < 16; ++i, ++Tx) {
sum[i] = Tx[0] * input[0] + Tx[16] * input[1] +
Tx[32] * input[2] + Tx[48] * input[3] +
Tx[64] * input[4] + Tx[80] * input[5] +
Tx[96] * input[6] + Tx[112] * input[7] +
Tx[128] * input[8] + Tx[144] * input[9] +
Tx[160] * input[10] + Tx[176] * input[11] +
Tx[192] * input[12] + Tx[208] * input[13] +
Tx[224] * input[14] + Tx[240] * input[15];
}
for (i = 0; i < 16; ++i) {
output[i] = WRAPLOW(ROUND_POWER_OF_TWO(sum[i], DCT_CONST_BITS), 8);
}
}
#if CONFIG_VP9_HIGHBITDEPTH
static INLINE void vp9_highbd_igentx4(const tran_low_t *input,
tran_low_t *output,
int bd, const tran_high_t *T) {
tran_high_t sum[4];
int i;
const tran_high_t *Tx = T;
(void) bd;
for (i = 0; i < 4; ++i, Tx += 1) {
sum[i] = Tx[0] * input[0] + Tx[4] * input[1] +
Tx[8] * input[2] + Tx[12] * input[3];
}
for (i = 0; i < 4; ++i) {
output[i] = WRAPLOW(ROUND_POWER_OF_TWO(sum[i], DCT_CONST_BITS), bd);
}
}
static INLINE void vp9_highbd_igentx8(const tran_low_t *input,
tran_low_t *output,
int bd, const tran_high_t *T) {
tran_high_t sum[8];
int i;
const tran_high_t *Tx = T;
(void) bd;
for (i = 0; i < 8; ++i, Tx += 1) {
sum[i] = Tx[0] * input[0] + Tx[8] * input[1] +
Tx[16] * input[2] + Tx[24] * input[3] +
Tx[32] * input[4] + Tx[40] * input[5] +
Tx[48] * input[6] + Tx[56] * input[7];
}
for (i = 0; i < 8; ++i) {
output[i] = WRAPLOW(ROUND_POWER_OF_TWO(sum[i], DCT_CONST_BITS), bd);
}
}
static INLINE void vp9_highbd_igentx16(const tran_low_t *input,
tran_low_t *output,
int bd, const tran_high_t *T) {
tran_high_t sum[16];
int i;
const tran_high_t *Tx = T;
(void) bd;
for (i = 0; i < 16; ++i, Tx += 1) {
sum[i] = Tx[0] * input[0] + Tx[16] * input[1] +
Tx[32] * input[2] + Tx[48] * input[3] +
Tx[64] * input[4] + Tx[80] * input[5] +
Tx[96] * input[6] + Tx[112] * input[7] +
Tx[128] * input[8] + Tx[144] * input[9] +
Tx[160] * input[10] + Tx[176] * input[11] +
Tx[192] * input[12] + Tx[208] * input[13] +
Tx[224] * input[14] + Tx[240] * input[15];
}
for (i = 0; i < 16; ++i) {
output[i] = WRAPLOW(ROUND_POWER_OF_TWO(sum[i], DCT_CONST_BITS), bd);
}
}
#endif // CONFIG_VP9_HIGHBITDEPTH
#endif // CONFIG_EXT_TX
void vp9_iwht4x4_add(const tran_low_t *input, uint8_t *dest, int stride,
int eob);
void vp9_idct4x4_add(const tran_low_t *input, uint8_t *dest, int stride,
int eob);
void vp9_idct8x8_add(const tran_low_t *input, uint8_t *dest, int stride,
int eob);
void vp9_idct16x16_add(const tran_low_t *input, uint8_t *dest, int stride,
int eob);
void vp9_idct32x32_add(const tran_low_t *input, uint8_t *dest, int stride,
int eob);
#if CONFIG_TX64X64
void vp9_idct64x64_add(const tran_low_t *input, uint8_t *dest, int stride,
int eob);
#endif // CONFIG_TX64X64
void vp9_iht4x4_add(TX_TYPE tx_type, const tran_low_t *input, uint8_t *dest,
int stride, int eob);
void vp9_iht8x8_add(TX_TYPE tx_type, const tran_low_t *input, uint8_t *dest,
int stride, int eob);
void vp9_iht16x16_add(TX_TYPE tx_type, const tran_low_t *input, uint8_t *dest,
int stride, int eob);
#if CONFIG_VP9_HIGHBITDEPTH
void vp9_highbd_idct4(const tran_low_t *input, tran_low_t *output, int bd);
void vp9_highbd_idct8(const tran_low_t *input, tran_low_t *output, int bd);
void vp9_highbd_idct16(const tran_low_t *input, tran_low_t *output, int bd);
void vp9_highbd_iwht4x4_add(const tran_low_t *input, uint8_t *dest, int stride,
int eob, int bd);
void vp9_highbd_idct4x4_add(const tran_low_t *input, uint8_t *dest, int stride,
int eob, int bd);
void vp9_highbd_idct8x8_add(const tran_low_t *input, uint8_t *dest, int stride,
int eob, int bd);
void vp9_highbd_idct16x16_add(const tran_low_t *input, uint8_t *dest,
int stride, int eob, int bd);
void vp9_highbd_idct32x32_add(const tran_low_t *input, uint8_t *dest,
int stride, int eob, int bd);
#if CONFIG_TX64X64
void vp9_highbd_idct64x64_add(const tran_low_t *input, uint8_t *dest,
int stride, int eob, int bd);
#endif
void vp9_highbd_iht4x4_add(TX_TYPE tx_type, const tran_low_t *input,
uint8_t *dest, int stride, int eob, int bd);
void vp9_highbd_iht8x8_add(TX_TYPE tx_type, const tran_low_t *input,
uint8_t *dest, int stride, int eob, int bd);
void vp9_highbd_iht16x16_add(TX_TYPE tx_type, const tran_low_t *input,
uint8_t *dest, int stride, int eob, int bd);
static INLINE uint16_t highbd_clip_pixel_add(uint16_t dest, tran_high_t trans,
int bd) {
trans = WRAPLOW(trans, bd);
return clip_pixel_highbd(WRAPLOW(dest + trans, bd), bd);
}
#endif // CONFIG_VP9_HIGHBITDEPTH
#if CONFIG_TX_SKIP
void vp9_tx_identity_add_rect(const tran_low_t *input, uint8_t *dest,
int row, int col, int stride_in,
int stride_out, int shift);
void vp9_tx_identity_add(const tran_low_t *input, uint8_t *dest,
int stride, int bs, int shift);
#if CONFIG_VP9_HIGHBITDEPTH
void vp9_highbd_tx_identity_add_rect(const tran_low_t *input, uint8_t *dest,
int row, int col, int stride_in,
int stride_out, int shift, int bd);
void vp9_highbd_tx_identity_add(const tran_low_t *input, uint8_t *dest,
int stride, int bs, int shift, int bd);
#endif // CONFIG_VP9_HIGHBITDEPTH
#endif // CONFIG_TX_SKIP
void vp9_dst1d_type1(int64_t *in, int64_t *out, int N);
void vp9_idst4x4_add(const tran_low_t *input, uint8_t *dest, int stride);
void vp9_idst8x8_add(const tran_low_t *input, uint8_t *dest, int stride);
void vp9_idst16x16_add(const tran_low_t *input, uint8_t *dest, int stride);
#ifdef __cplusplus
} // extern "C"
#endif
#endif // VP9_COMMON_VP9_IDCT_H_
#if CONFIG_SR_MODE
void vp9_iwht4x4(const tran_low_t *input, int16_t *dest, int stride,
int eob);
void vp9_idct4x4(const tran_low_t *input, int16_t *dest, int stride,
int eob);
void vp9_idct8x8(const tran_low_t *input, int16_t *dest, int stride,
int eob);
void vp9_idct16x16(const tran_low_t *input, int16_t *dest, int stride,
int eob);
void vp9_idct32x32(const tran_low_t *input, int16_t *dest, int stride,
int eob);
void vp9_iht4x4(TX_TYPE tx_type, const tran_low_t *input, int16_t *dest,
int stride, int eob);
void vp9_iht8x8(TX_TYPE tx_type, const tran_low_t *input, int16_t *dest,
int stride, int eob);
void vp9_iht16x16(TX_TYPE tx_type, const tran_low_t *input, int16_t *dest,
int stride, int eob);
#if CONFIG_TX64X64
void vp9_idct64x64(const tran_low_t *input, int16_t *dest, int stride,
int eob);
#endif // CONFIG_TX64X64
#endif // CONFIG_SR_MODE
Reference in New Issue View Git Blame Copy Permalink