Commit 8e46b860 authored by Yi Luo's avatar Yi Luo
Browse files

Fix filter type mismatch warning on Visual Studio

- Move filter look-up functions to corresponding optimization modules.

BUG=webm:1296

Change-Id: I87f399609052db2dbc7e5a590afb08b82e3fa89f
parent bf043127
......@@ -72,9 +72,11 @@ AV1_COMMON_SRCS-yes += common/av1_fwd_txfm2d.c
AV1_COMMON_SRCS-yes += common/av1_fwd_txfm2d_cfg.h
AV1_COMMON_SRCS-yes += common/av1_inv_txfm2d.c
AV1_COMMON_SRCS-yes += common/av1_inv_txfm2d_cfg.h
AV1_COMMON_SRCS-$(HAVE_SSSE3) += common/x86/convolve_filter_ssse3.h
AV1_COMMON_SRCS-$(HAVE_SSSE3) += common/x86/av1_convolve_ssse3.c
AV1_COMMON_SRCS-$(HAVE_SSSE3) += common/x86/av1_convolve_filters_ssse3.c
ifeq ($(CONFIG_AOM_HIGHBITDEPTH),yes)
AV1_COMMON_SRCS-$(HAVE_SSE4_1) += common/x86/highbd_convolve_filter_sse4.h
AV1_COMMON_SRCS-$(HAVE_SSE4_1) += common/x86/av1_highbd_convolve_sse4.c
AV1_COMMON_SRCS-$(HAVE_SSE4_1) += common/x86/av1_highbd_convolve_filters_sse4.c
endif
......
......@@ -26,8 +26,8 @@ DECLARE_ALIGNED(256, static const InterpKernel,
};
#if USE_TEMPORALFILTER_12TAP
DECLARE_ALIGNED(16, static const int16_t,
sub_pel_filters_temporalfilter_12[SUBPEL_SHIFTS][12]) = {
DECLARE_ALIGNED(16, const int16_t,
av1_sub_pel_filters_temporalfilter_12[SUBPEL_SHIFTS][12]) = {
// intfilt 0.8
{ 0, 0, 0, 0, 0, 128, 0, 0, 0, 0, 0, 0 },
{ 0, 1, -1, 3, -7, 127, 8, -4, 2, -1, 0, 0 },
......@@ -77,8 +77,8 @@ DECLARE_ALIGNED(256, static const InterpKernel,
};
#endif // CONFIG_EXT_INTRA
DECLARE_ALIGNED(256, static const int16_t,
sub_pel_filters_10sharp[SUBPEL_SHIFTS][10]) = {
DECLARE_ALIGNED(256, const int16_t,
av1_sub_pel_filters_10sharp[SUBPEL_SHIFTS][10]) = {
// intfilt 0.77
{ 0, 0, 0, 0, 128, 0, 0, 0, 0, 0 },
{ 0, -1, 3, -6, 127, 8, -4, 2, -1, 0 },
......@@ -124,8 +124,8 @@ DECLARE_ALIGNED(256, static const InterpKernel,
{ 0, 2, -12, 37, 94, 14, -9, 2 }, { 0, 2, -11, 31, 95, 19, -10, 2 },
};
DECLARE_ALIGNED(16, static const int16_t,
sub_pel_filters_12sharp[SUBPEL_SHIFTS][12]) = {
DECLARE_ALIGNED(16, const int16_t,
av1_sub_pel_filters_12sharp[SUBPEL_SHIFTS][12]) = {
// intfilt 0.85
{ 0, 0, 0, 0, 0, 128, 0, 0, 0, 0, 0, 0 },
{ 0, 1, -2, 3, -7, 127, 8, -4, 2, -1, 1, 0 },
......@@ -200,9 +200,9 @@ static const InterpFilterParams
av1_interp_filter_params_list[SWITCHABLE_FILTERS + 1] = {
{ (const int16_t *)sub_pel_filters_8, SUBPEL_TAPS, SUBPEL_SHIFTS },
{ (const int16_t *)sub_pel_filters_8smooth, SUBPEL_TAPS, SUBPEL_SHIFTS },
{ (const int16_t *)sub_pel_filters_10sharp, 10, SUBPEL_SHIFTS },
{ (const int16_t *)av1_sub_pel_filters_10sharp, 10, SUBPEL_SHIFTS },
{ (const int16_t *)sub_pel_filters_8smooth2, SUBPEL_TAPS, SUBPEL_SHIFTS },
{ (const int16_t *)sub_pel_filters_12sharp, 12, SUBPEL_SHIFTS },
{ (const int16_t *)av1_sub_pel_filters_12sharp, 12, SUBPEL_SHIFTS },
{ (const int16_t *)bilinear_filters, SUBPEL_TAPS, SUBPEL_SHIFTS }
};
#else
......@@ -217,7 +217,7 @@ static const InterpFilterParams
#if USE_TEMPORALFILTER_12TAP
static const InterpFilterParams av1_interp_temporalfilter_12tap = {
(const int16_t *)sub_pel_filters_temporalfilter_12, 12, SUBPEL_SHIFTS
(const int16_t *)av1_sub_pel_filters_temporalfilter_12, 12, SUBPEL_SHIFTS
};
#endif // USE_TEMPORALFILTER_12TAP
......@@ -238,65 +238,3 @@ const int16_t *av1_get_interp_filter_kernel(const InterpFilter interp_filter) {
return (const int16_t *)av1_interp_filter_params_list[interp_filter]
.filter_ptr;
}
SubpelFilterCoeffs av1_get_subpel_filter_signal_dir(const InterpFilterParams p,
int index) {
#if CONFIG_EXT_INTERP && HAVE_SSSE3
if (p.filter_ptr == (const int16_t *)sub_pel_filters_12sharp) {
return &sub_pel_filters_12sharp_signal_dir[index][0];
}
if (p.filter_ptr == (const int16_t *)sub_pel_filters_10sharp) {
return &sub_pel_filters_10sharp_signal_dir[index][0];
}
#endif
#if USE_TEMPORALFILTER_12TAP && HAVE_SSSE3
if (p.filter_ptr == (const int16_t *)sub_pel_filters_temporalfilter_12) {
return &sub_pel_filters_temporalfilter_12_signal_dir[index][0];
}
#endif
(void)p;
(void)index;
return NULL;
}
SubpelFilterCoeffs av1_get_subpel_filter_ver_signal_dir(
const InterpFilterParams p, int index) {
#if CONFIG_EXT_INTERP && HAVE_SSSE3
if (p.filter_ptr == (const int16_t *)sub_pel_filters_12sharp) {
return &sub_pel_filters_12sharp_ver_signal_dir[index][0];
}
if (p.filter_ptr == (const int16_t *)sub_pel_filters_10sharp) {
return &sub_pel_filters_10sharp_ver_signal_dir[index][0];
}
#endif
#if USE_TEMPORALFILTER_12TAP && HAVE_SSSE3
if (p.filter_ptr == (const int16_t *)sub_pel_filters_temporalfilter_12) {
return &sub_pel_filters_temporalfilter_12_ver_signal_dir[index][0];
}
#endif
(void)p;
(void)index;
return NULL;
}
#if CONFIG_AOM_HIGHBITDEPTH
HbdSubpelFilterCoeffs av1_hbd_get_subpel_filter_ver_signal_dir(
const InterpFilterParams p, int index) {
#if CONFIG_EXT_INTERP && HAVE_SSE4_1
if (p.filter_ptr == (const int16_t *)sub_pel_filters_12sharp) {
return &sub_pel_filters_12sharp_highbd_ver_signal_dir[index][0];
}
if (p.filter_ptr == (const int16_t *)sub_pel_filters_10sharp) {
return &sub_pel_filters_10sharp_highbd_ver_signal_dir[index][0];
}
#endif
#if USE_TEMPORALFILTER_12TAP && HAVE_SSE4_1
if (p.filter_ptr == (const int16_t *)sub_pel_filters_temporalfilter_12) {
return &sub_pel_filters_temporalfilter_12_highbd_ver_signal_dir[index][0];
}
#endif
(void)p;
(void)index;
return NULL;
}
#endif
......@@ -92,24 +92,15 @@ static INLINE int av1_is_interpolating_filter(
}
#if USE_TEMPORALFILTER_12TAP
extern const int8_t sub_pel_filters_temporalfilter_12_signal_dir[15][2][16];
extern const int8_t sub_pel_filters_temporalfilter_12_ver_signal_dir[15][6][16];
#if CONFIG_AOM_HIGHBITDEPTH
extern const int16_t sub_pel_filters_temporalfilter_12_highbd_ver_signal_dir[15]
[6]
[8];
#endif
DECLARE_ALIGNED(16, extern const int16_t,
av1_sub_pel_filters_temporalfilter_12[SUBPEL_SHIFTS][12]);
#endif
#if CONFIG_EXT_INTERP
extern const int8_t sub_pel_filters_12sharp_signal_dir[15][2][16];
extern const int8_t sub_pel_filters_10sharp_signal_dir[15][2][16];
extern const int8_t sub_pel_filters_12sharp_ver_signal_dir[15][6][16];
extern const int8_t sub_pel_filters_10sharp_ver_signal_dir[15][6][16];
#if CONFIG_AOM_HIGHBITDEPTH
extern const int16_t sub_pel_filters_12sharp_highbd_ver_signal_dir[15][6][8];
extern const int16_t sub_pel_filters_10sharp_highbd_ver_signal_dir[15][6][8];
#endif
DECLARE_ALIGNED(256, extern const int16_t,
av1_sub_pel_filters_10sharp[SUBPEL_SHIFTS][10]);
DECLARE_ALIGNED(16, extern const int16_t,
av1_sub_pel_filters_12sharp[SUBPEL_SHIFTS][12]);
#endif
typedef const int8_t (*SubpelFilterCoeffs)[16];
......
......@@ -12,7 +12,7 @@
#if CONFIG_EXT_INTERP
DECLARE_ALIGNED(16, const int8_t,
sub_pel_filters_10sharp_signal_dir[15][2][16]) = {
av1_sub_pel_filters_10sharp_signal_dir[15][2][16]) = {
{
{ 0, 0, -1, 3, -6, 127, 8, -4, 2, -1, 0, 0, 0, 0, 0, 0 },
{ 0, 0, 0, 0, -1, 3, -6, 127, 8, -4, 2, -1, 0, 0, 0, 0 },
......@@ -77,7 +77,7 @@ DECLARE_ALIGNED(16, const int8_t,
#endif
#if CONFIG_EXT_INTERP
DECLARE_ALIGNED(16, const int8_t,
sub_pel_filters_10sharp_ver_signal_dir[15][6][16]) = {
av1_sub_pel_filters_10sharp_ver_signal_dir[15][6][16]) = {
{
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
{ -1, 3, -1, 3, -1, 3, -1, 3, -1, 3, -1, 3, -1, 3, -1, 3 },
......@@ -228,7 +228,7 @@ DECLARE_ALIGNED(16, const int8_t,
#endif
#if CONFIG_EXT_INTERP
DECLARE_ALIGNED(16, const int8_t,
sub_pel_filters_12sharp_signal_dir[15][2][16]) = {
av1_sub_pel_filters_12sharp_signal_dir[15][2][16]) = {
{
{ 0, 1, -2, 3, -7, 127, 8, -4, 2, -1, 1, 0, 0, 0, 0, 0 },
{ 0, 0, 0, 1, -2, 3, -7, 127, 8, -4, 2, -1, 1, 0, 0, 0 },
......@@ -293,7 +293,7 @@ DECLARE_ALIGNED(16, const int8_t,
#endif
#if CONFIG_EXT_INTERP
DECLARE_ALIGNED(16, const int8_t,
sub_pel_filters_12sharp_ver_signal_dir[15][6][16]) = {
av1_sub_pel_filters_12sharp_ver_signal_dir[15][6][16]) = {
{
{ 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1 },
{ -2, 3, -2, 3, -2, 3, -2, 3, -2, 3, -2, 3, -2, 3, -2, 3 },
......@@ -444,7 +444,7 @@ DECLARE_ALIGNED(16, const int8_t,
#endif
#if USE_TEMPORALFILTER_12TAP
DECLARE_ALIGNED(16, const int8_t,
sub_pel_filters_temporalfilter_12_signal_dir[15][2][16]) = {
av1_sub_pel_filters_temporalfilter_12_signal_dir[15][2][16]) = {
{
{ 0, 1, -1, 3, -7, 127, 8, -4, 2, -1, 0, 0, 0, 0, 0, 0 },
{ 0, 0, 0, 1, -1, 3, -7, 127, 8, -4, 2, -1, 0, 0, 0, 0 },
......@@ -509,7 +509,8 @@ DECLARE_ALIGNED(16, const int8_t,
#endif
#if USE_TEMPORALFILTER_12TAP
DECLARE_ALIGNED(16, const int8_t,
sub_pel_filters_temporalfilter_12_ver_signal_dir[15][6][16]) = {
av1_sub_pel_filters_temporalfilter_12_ver_signal_dir[15][6]
[16]) = {
{
{ 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1 },
{ -1, 3, -1, 3, -1, 3, -1, 3, -1, 3, -1, 3, -1, 3, -1, 3 },
......
......@@ -13,11 +13,51 @@
#include <tmmintrin.h>
#include "./av1_rtcd.h"
#include "av1/common/filter.h"
#include "av1/common/x86/convolve_filter_ssse3.h"
#define WIDTH_BOUND (16)
#define HEIGHT_BOUND (16)
static SubpelFilterCoeffs get_subpel_filter_signal_dir(
const InterpFilterParams p, int index) {
#if CONFIG_EXT_INTERP
if (p.filter_ptr == (const int16_t *)av1_sub_pel_filters_12sharp) {
return &av1_sub_pel_filters_12sharp_signal_dir[index][0];
}
if (p.filter_ptr == (const int16_t *)av1_sub_pel_filters_10sharp) {
return &av1_sub_pel_filters_10sharp_signal_dir[index][0];
}
#endif
#if USE_TEMPORALFILTER_12TAP
if (p.filter_ptr == (const int16_t *)av1_sub_pel_filters_temporalfilter_12) {
return &av1_sub_pel_filters_temporalfilter_12_signal_dir[index][0];
}
#endif
(void)p;
(void)index;
return NULL;
}
static SubpelFilterCoeffs get_subpel_filter_ver_signal_dir(
const InterpFilterParams p, int index) {
#if CONFIG_EXT_INTERP
if (p.filter_ptr == (const int16_t *)av1_sub_pel_filters_12sharp) {
return &av1_sub_pel_filters_12sharp_ver_signal_dir[index][0];
}
if (p.filter_ptr == (const int16_t *)av1_sub_pel_filters_10sharp) {
return &av1_sub_pel_filters_10sharp_ver_signal_dir[index][0];
}
#endif
#if USE_TEMPORALFILTER_12TAP
if (p.filter_ptr == (const int16_t *)av1_sub_pel_filters_temporalfilter_12) {
return &av1_sub_pel_filters_temporalfilter_12_ver_signal_dir[index][0];
}
#endif
(void)p;
(void)index;
return NULL;
}
static INLINE void transpose_4x8(const __m128i *in, __m128i *out) {
__m128i t0, t1;
......@@ -636,9 +676,8 @@ void av1_convolve_horiz_ssse3(const uint8_t *src, int src_stride, uint8_t *dst,
return;
}
hCoeffs = av1_get_subpel_filter_signal_dir(filter_params, subpel_x_q4 - 1);
vCoeffs =
av1_get_subpel_filter_ver_signal_dir(filter_params, subpel_x_q4 - 1);
hCoeffs = get_subpel_filter_signal_dir(filter_params, subpel_x_q4 - 1);
vCoeffs = get_subpel_filter_ver_signal_dir(filter_params, subpel_x_q4 - 1);
if (!hCoeffs || !vCoeffs) {
av1_convolve_horiz_c(src, src_stride, dst, dst_stride, w, h, filter_params,
......@@ -845,8 +884,7 @@ void av1_convolve_vert_ssse3(const uint8_t *src, int src_stride, uint8_t *dst,
return;
}
vCoeffs =
av1_get_subpel_filter_ver_signal_dir(filter_params, subpel_y_q4 - 1);
vCoeffs = get_subpel_filter_ver_signal_dir(filter_params, subpel_y_q4 - 1);
if (!vCoeffs) {
av1_convolve_vert_c(src, src_stride, dst, dst_stride, w, h, filter_params,
......
......@@ -13,7 +13,7 @@
#if CONFIG_AOM_HIGHBITDEPTH
#if CONFIG_EXT_INTERP
DECLARE_ALIGNED(16, const int16_t,
sub_pel_filters_10sharp_highbd_ver_signal_dir[15][6][8]) = {
av1_sub_pel_filters_10sharp_highbd_ver_signal_dir[15][6][8]) = {
{
{ 0, 0, 0, 0, 0, 0, 0, 0 },
{ -1, 3, -1, 3, -1, 3, -1, 3 },
......@@ -140,7 +140,7 @@ DECLARE_ALIGNED(16, const int16_t,
#if CONFIG_AOM_HIGHBITDEPTH
#if CONFIG_EXT_INTERP
DECLARE_ALIGNED(16, const int16_t,
sub_pel_filters_12sharp_highbd_ver_signal_dir[15][6][8]) = {
av1_sub_pel_filters_12sharp_highbd_ver_signal_dir[15][6][8]) = {
{
{ 0, 1, 0, 1, 0, 1, 0, 1 },
{ -2, 3, -2, 3, -2, 3, -2, 3 },
......@@ -266,9 +266,9 @@ DECLARE_ALIGNED(16, const int16_t,
#endif
#if CONFIG_AOM_HIGHBITDEPTH
#if USE_TEMPORALFILTER_12TAP
DECLARE_ALIGNED(16, const int16_t,
sub_pel_filters_temporalfilter_12_highbd_ver_signal_dir[15][6]
[8]) = {
DECLARE_ALIGNED(
16, const int16_t,
av1_sub_pel_filters_temporalfilter_12_highbd_ver_signal_dir[15][6][8]) = {
{
{ 0, 1, 0, 1, 0, 1, 0, 1 },
{ -1, 3, -1, 3, -1, 3, -1, 3 },
......
......@@ -13,12 +13,33 @@
#include <smmintrin.h>
#include "./av1_rtcd.h"
#include "av1/common/filter.h"
#include "av1/common/x86/highbd_convolve_filter_sse4.h"
typedef void (*TransposeSave)(const int width, int pixelsNum, uint32_t *src,
int src_stride, uint16_t *dst, int dst_stride,
int bd);
static HbdSubpelFilterCoeffs hbd_get_subpel_filter_ver_signal_dir(
const InterpFilterParams p, int index) {
#if CONFIG_EXT_INTERP
if (p.filter_ptr == (const int16_t *)av1_sub_pel_filters_12sharp) {
return &av1_sub_pel_filters_12sharp_highbd_ver_signal_dir[index][0];
}
if (p.filter_ptr == (const int16_t *)av1_sub_pel_filters_10sharp) {
return &av1_sub_pel_filters_10sharp_highbd_ver_signal_dir[index][0];
}
#endif
#if USE_TEMPORALFILTER_12TAP
if (p.filter_ptr == (const int16_t *)av1_sub_pel_filters_temporalfilter_12) {
return &av1_sub_pel_filters_temporalfilter_12_highbd_ver_signal_dir[index]
[0];
}
#endif
(void)p;
(void)index;
return NULL;
}
// pixelsNum 0: write all 4 pixels
// 1/2/3: residual pixels 1/2/3
static void writePixel(__m128i *u, int width, int pixelsNum, uint16_t *dst,
......@@ -235,7 +256,7 @@ void av1_highbd_convolve_horiz_sse4_1(const uint16_t *src, int src_stride,
}
vCoeffs =
av1_hbd_get_subpel_filter_ver_signal_dir(filter_params, subpel_x_q4 - 1);
hbd_get_subpel_filter_ver_signal_dir(filter_params, subpel_x_q4 - 1);
if (!vCoeffs) {
av1_highbd_convolve_horiz_c(src, src_stride, dst, dst_stride, w, h,
filter_params, subpel_x_q4, x_step_q4, avg, bd);
......@@ -439,7 +460,7 @@ void av1_highbd_convolve_vert_sse4_1(const uint16_t *src, int src_stride,
}
vCoeffs =
av1_hbd_get_subpel_filter_ver_signal_dir(filter_params, subpel_y_q4 - 1);
hbd_get_subpel_filter_ver_signal_dir(filter_params, subpel_y_q4 - 1);
if (!vCoeffs) {
av1_highbd_convolve_vert_c(src, src_stride, dst, dst_stride, w, h,
filter_params, subpel_y_q4, y_step_q4, avg, bd);
......
/*
* Copyright (c) 2016, Alliance for Open Media. All rights reserved
*
* This source code is subject to the terms of the BSD 2 Clause License and
* the Alliance for Open Media Patent License 1.0. If the BSD 2 Clause License
* was not distributed with this source code in the LICENSE file, you can
* obtain it at www.aomedia.org/license/software. If the Alliance for Open
* Media Patent License 1.0 was not distributed with this source code in the
* PATENTS file, you can obtain it at www.aomedia.org/license/patent.
*/
#ifndef AV1_COMMON_X86_CONVOLVE_FILTER_SSSE3_H_
#define AV1_COMMON_X86_CONVOLVE_FILTER_SSSE3_H_
#include "./aom_config.h"
#include "av1/common/filter.h"
#if USE_TEMPORALFILTER_12TAP
DECLARE_ALIGNED(16, extern const int8_t,
av1_sub_pel_filters_temporalfilter_12_signal_dir[15][2][16]);
DECLARE_ALIGNED(16, extern const int8_t,
av1_sub_pel_filters_temporalfilter_12_ver_signal_dir[15][6]
[16]);
#endif
#if CONFIG_EXT_INTERP
DECLARE_ALIGNED(16, extern const int8_t,
av1_sub_pel_filters_12sharp_signal_dir[15][2][16]);
DECLARE_ALIGNED(16, extern const int8_t,
av1_sub_pel_filters_10sharp_signal_dir[15][2][16]);
DECLARE_ALIGNED(16, extern const int8_t,
av1_sub_pel_filters_12sharp_ver_signal_dir[15][6][16]);
DECLARE_ALIGNED(16, extern const int8_t,
av1_sub_pel_filters_10sharp_ver_signal_dir[15][6][16]);
#endif
#endif // AV1_COMMON_X86_CONVOLVE_FILTER_SSSE3_H_
/*
* Copyright (c) 2016, Alliance for Open Media. All rights reserved
*
* This source code is subject to the terms of the BSD 2 Clause License and
* the Alliance for Open Media Patent License 1.0. If the BSD 2 Clause License
* was not distributed with this source code in the LICENSE file, you can
* obtain it at www.aomedia.org/license/software. If the Alliance for Open
* Media Patent License 1.0 was not distributed with this source code in the
* PATENTS file, you can obtain it at www.aomedia.org/license/patent.
*/
#ifndef AV1_COMMON_X86_HIGHBD_CONVOLVE_FILTER_SSE4_H_
#define AV1_COMMON_X86_HIGHBD_CONVOLVE_FILTER_SSE4_H_
#include "./aom_config.h"
#include "av1/common/filter.h"
#if USE_TEMPORALFILTER_12TAP
DECLARE_ALIGNED(16, extern const int16_t,
av1_sub_pel_filters_temporalfilter_12_highbd_ver_signal_dir[15]
[6]
[8]);
#endif
#if CONFIG_EXT_INTERP
DECLARE_ALIGNED(16, extern const int16_t,
av1_sub_pel_filters_12sharp_highbd_ver_signal_dir[15][6][8]);
DECLARE_ALIGNED(16, extern const int16_t,
av1_sub_pel_filters_10sharp_highbd_ver_signal_dir[15][6][8]);
#endif
#endif // AV1_COMMON_X86_HIGHBD_CONVOLVE_FILTER_SSE4_H_
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment