Skip to content

F
ffmpeg.wasm-core
Commit d4956b0b authored by Michael Niedermayer's avatar Michael Niedermayer
Browse files
Options

sws/input: replace hardcoded rgb2yuv coefficients by table 

Signed-off-by: 's avatarMichael Niedermayer <michaelni@gmx.at>
parent 41ebb645
Hide whitespace changes
Inline Side-by-side
Showing with 174 additions and 124 deletions
libswscale/input.c
View file @ d4956b0b
...@@ -37,15 +37,6 @@ ...@@ -37,15 +37,6 @@
#include "swscale_internal.h" #include "swscale_internal.h"
#define RGB2YUV_SHIFT 15 #define RGB2YUV_SHIFT 15
#define BY ((int)(0.114 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
#define BV (-(int)(0.081 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
#define BU ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
#define GY ((int)(0.587 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
#define GV (-(int)(0.419 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
#define GU (-(int)(0.331 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
#define RY ((int)(0.299 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
#define RV ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
#define RU (-(int)(0.169 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
#define input_pixel(pos) (isBE(origin) ? AV_RB16(pos) : AV_RL16(pos)) #define input_pixel(pos) (isBE(origin) ? AV_RB16(pos) : AV_RL16(pos))
...@@ -54,79 +45,84 @@ ...@@ -54,79 +45,84 @@
static av_always_inline void static av_always_inline void
rgb64ToY_c_template(uint16_t *dst, const uint16_t *src, int width, rgb64ToY_c_template(uint16_t *dst, const uint16_t *src, int width,
enum AVPixelFormat origin) enum AVPixelFormat origin, int32_t *rgb2yuv)
{ {
int32_t ry = rgb2yuv[RY_IDX], gy = rgb2yuv[GY_IDX], by = rgb2yuv[BY_IDX];
int i; int i;
for (i = 0; i < width; i++) { for (i = 0; i < width; i++) {
unsigned int r_b = input_pixel(&src[i*4+0]); unsigned int r_b = input_pixel(&src[i*4+0]);
unsigned int g = input_pixel(&src[i*4+1]); unsigned int g = input_pixel(&src[i*4+1]);
unsigned int b_r = input_pixel(&src[i*4+2]); unsigned int b_r = input_pixel(&src[i*4+2]);
dst[i] = (RY*r + GY*g + BY*b + (0x2001<<(RGB2YUV_SHIFT-1))) >> RGB2YUV_SHIFT; dst[i] = (ry*r + gy*g + by*b + (0x2001<<(RGB2YUV_SHIFT-1))) >> RGB2YUV_SHIFT;
} }
} }
static av_always_inline void static av_always_inline void
rgb64ToUV_c_template(uint16_t *dstU, uint16_t *dstV, rgb64ToUV_c_template(uint16_t *dstU, uint16_t *dstV,
const uint16_t *src1, const uint16_t *src2, const uint16_t *src1, const uint16_t *src2,
int width, enum AVPixelFormat origin) int width, enum AVPixelFormat origin, int32_t *rgb2yuv)
{ {
int i; int i;
int32_t ru = rgb2yuv[RU_IDX], gu = rgb2yuv[GU_IDX], bu = rgb2yuv[BU_IDX];
int32_t rv = rgb2yuv[RV_IDX], gv = rgb2yuv[GV_IDX], bv = rgb2yuv[BV_IDX];
av_assert1(src1==src2); av_assert1(src1==src2);
for (i = 0; i < width; i++) { for (i = 0; i < width; i++) {
int r_b = input_pixel(&src1[i*4+0]); int r_b = input_pixel(&src1[i*4+0]);
int g = input_pixel(&src1[i*4+1]); int g = input_pixel(&src1[i*4+1]);
int b_r = input_pixel(&src1[i*4+2]); int b_r = input_pixel(&src1[i*4+2]);
dstU[i] = (RU*r + GU*g + BU*b + (0x10001<<(RGB2YUV_SHIFT-1))) >> RGB2YUV_SHIFT; dstU[i] = (ru*r + gu*g + bu*b + (0x10001<<(RGB2YUV_SHIFT-1))) >> RGB2YUV_SHIFT;
dstV[i] = (RV*r + GV*g + BV*b + (0x10001<<(RGB2YUV_SHIFT-1))) >> RGB2YUV_SHIFT; dstV[i] = (rv*r + gv*g + bv*b + (0x10001<<(RGB2YUV_SHIFT-1))) >> RGB2YUV_SHIFT;
} }
} }
static av_always_inline void static av_always_inline void
rgb64ToUV_half_c_template(uint16_t *dstU, uint16_t *dstV, rgb64ToUV_half_c_template(uint16_t *dstU, uint16_t *dstV,
const uint16_t *src1, const uint16_t *src2, const uint16_t *src1, const uint16_t *src2,
int width, enum AVPixelFormat origin) int width, enum AVPixelFormat origin, int32_t *rgb2yuv)
{ {
int i; int i;
int32_t ru = rgb2yuv[RU_IDX], gu = rgb2yuv[GU_IDX], bu = rgb2yuv[BU_IDX];
int32_t rv = rgb2yuv[RV_IDX], gv = rgb2yuv[GV_IDX], bv = rgb2yuv[BV_IDX];
av_assert1(src1==src2); av_assert1(src1==src2);
for (i = 0; i < width; i++) { for (i = 0; i < width; i++) {
int r_b = (input_pixel(&src1[8 * i + 0]) + input_pixel(&src1[8 * i + 4]) + 1) >> 1; int r_b = (input_pixel(&src1[8 * i + 0]) + input_pixel(&src1[8 * i + 4]) + 1) >> 1;
int g = (input_pixel(&src1[8 * i + 1]) + input_pixel(&src1[8 * i + 5]) + 1) >> 1; int g = (input_pixel(&src1[8 * i + 1]) + input_pixel(&src1[8 * i + 5]) + 1) >> 1;
int b_r = (input_pixel(&src1[8 * i + 2]) + input_pixel(&src1[8 * i + 6]) + 1) >> 1; int b_r = (input_pixel(&src1[8 * i + 2]) + input_pixel(&src1[8 * i + 6]) + 1) >> 1;
dstU[i]= (RU*r + GU*g + BU*b + (0x10001<<(RGB2YUV_SHIFT-1))) >> RGB2YUV_SHIFT; dstU[i]= (ru*r + gu*g + bu*b + (0x10001<<(RGB2YUV_SHIFT-1))) >> RGB2YUV_SHIFT;
dstV[i]= (RV*r + GV*g + BV*b + (0x10001<<(RGB2YUV_SHIFT-1))) >> RGB2YUV_SHIFT; dstV[i]= (rv*r + gv*g + bv*b + (0x10001<<(RGB2YUV_SHIFT-1))) >> RGB2YUV_SHIFT;
} }
} }
#define rgb64funcs(pattern, BE_LE, origin) \ #define rgb64funcs(pattern, BE_LE, origin) \
static void pattern ## 64 ## BE_LE ## ToY_c(uint8_t *_dst, const uint8_t *_src, const uint8_t *unused0, const uint8_t *unused1,\ static void pattern ## 64 ## BE_LE ## ToY_c(uint8_t *_dst, const uint8_t *_src, const uint8_t *unused0, const uint8_t *unused1,\
int width, uint32_t *unused) \ int width, uint32_t *rgb2yuv) \
{ \ { \
const uint16_t *src = (const uint16_t *) _src; \ const uint16_t *src = (const uint16_t *) _src; \
uint16_t *dst = (uint16_t *) _dst; \ uint16_t *dst = (uint16_t *) _dst; \
rgb64ToY_c_template(dst, src, width, origin); \ rgb64ToY_c_template(dst, src, width, origin, rgb2yuv); \
} \ } \
\ \
static void pattern ## 64 ## BE_LE ## ToUV_c(uint8_t *_dstU, uint8_t *_dstV, \ static void pattern ## 64 ## BE_LE ## ToUV_c(uint8_t *_dstU, uint8_t *_dstV, \
const uint8_t *unused0, const uint8_t *_src1, const uint8_t *_src2, \ const uint8_t *unused0, const uint8_t *_src1, const uint8_t *_src2, \
int width, uint32_t *unused) \ int width, uint32_t *rgb2yuv) \
{ \ { \
const uint16_t *src1 = (const uint16_t *) _src1, \ const uint16_t *src1 = (const uint16_t *) _src1, \
*src2 = (const uint16_t *) _src2; \ *src2 = (const uint16_t *) _src2; \
uint16_t *dstU = (uint16_t *) _dstU, *dstV = (uint16_t *) _dstV; \ uint16_t *dstU = (uint16_t *) _dstU, *dstV = (uint16_t *) _dstV; \
rgb64ToUV_c_template(dstU, dstV, src1, src2, width, origin); \ rgb64ToUV_c_template(dstU, dstV, src1, src2, width, origin, rgb2yuv); \
} \ } \
\ \
static void pattern ## 64 ## BE_LE ## ToUV_half_c(uint8_t *_dstU, uint8_t *_dstV, \ static void pattern ## 64 ## BE_LE ## ToUV_half_c(uint8_t *_dstU, uint8_t *_dstV, \
const uint8_t *unused0, const uint8_t *_src1, const uint8_t *_src2, \ const uint8_t *unused0, const uint8_t *_src1, const uint8_t *_src2, \
int width, uint32_t *unused) \ int width, uint32_t *rgb2yuv) \
{ \ { \
const uint16_t *src1 = (const uint16_t *) _src1, \ const uint16_t *src1 = (const uint16_t *) _src1, \
*src2 = (const uint16_t *) _src2; \ *src2 = (const uint16_t *) _src2; \
uint16_t *dstU = (uint16_t *) _dstU, *dstV = (uint16_t *) _dstV; \ uint16_t *dstU = (uint16_t *) _dstU, *dstV = (uint16_t *) _dstV; \
rgb64ToUV_half_c_template(dstU, dstV, src1, src2, width, origin); \ rgb64ToUV_half_c_template(dstU, dstV, src1, src2, width, origin, rgb2yuv); \
} }
rgb64funcs(rgb, LE, AV_PIX_FMT_RGBA64LE) rgb64funcs(rgb, LE, AV_PIX_FMT_RGBA64LE)
...@@ -134,15 +130,17 @@ rgb64funcs(rgb, BE, AV_PIX_FMT_RGBA64BE) ...@@ -134,15 +130,17 @@ rgb64funcs(rgb, BE, AV_PIX_FMT_RGBA64BE)
static av_always_inline void rgb48ToY_c_template(uint16_t *dst, static av_always_inline void rgb48ToY_c_template(uint16_t *dst,
const uint16_t *src, int width, const uint16_t *src, int width,
enum AVPixelFormat origin) enum AVPixelFormat origin,
int32_t *rgb2yuv)
{ {
int32_t ry = rgb2yuv[RY_IDX], gy = rgb2yuv[GY_IDX], by = rgb2yuv[BY_IDX];
int i; int i;
for (i = 0; i < width; i++) { for (i = 0; i < width; i++) {
unsigned int r_b = input_pixel(&src[i * 3 + 0]); unsigned int r_b = input_pixel(&src[i * 3 + 0]);
unsigned int g = input_pixel(&src[i * 3 + 1]); unsigned int g = input_pixel(&src[i * 3 + 1]);
unsigned int b_r = input_pixel(&src[i * 3 + 2]); unsigned int b_r = input_pixel(&src[i * 3 + 2]);
dst[i] = (RY * r + GY * g + BY * b + (0x2001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT; dst[i] = (ry*r + gy*g + by*b + (0x2001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
} }
} }
...@@ -151,17 +149,20 @@ static av_always_inline void rgb48ToUV_c_template(uint16_t *dstU, ...@@ -151,17 +149,20 @@ static av_always_inline void rgb48ToUV_c_template(uint16_t *dstU,
const uint16_t *src1, const uint16_t *src1,
const uint16_t *src2, const uint16_t *src2,
int width, int width,
enum AVPixelFormat origin) enum AVPixelFormat origin,
int32_t *rgb2yuv)
{ {
int i; int i;
int32_t ru = rgb2yuv[RU_IDX], gu = rgb2yuv[GU_IDX], bu = rgb2yuv[BU_IDX];
int32_t rv = rgb2yuv[RV_IDX], gv = rgb2yuv[GV_IDX], bv = rgb2yuv[BV_IDX];
av_assert1(src1 == src2); av_assert1(src1 == src2);
for (i = 0; i < width; i++) { for (i = 0; i < width; i++) {
int r_b = input_pixel(&src1[i * 3 + 0]); int r_b = input_pixel(&src1[i * 3 + 0]);
int g = input_pixel(&src1[i * 3 + 1]); int g = input_pixel(&src1[i * 3 + 1]);
int b_r = input_pixel(&src1[i * 3 + 2]); int b_r = input_pixel(&src1[i * 3 + 2]);
dstU[i] = (RU * r + GU * g + BU * b + (0x10001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT; dstU[i] = (ru*r + gu*g + bu*b + (0x10001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
dstV[i] = (RV * r + GV * g + BV * b + (0x10001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT; dstV[i] = (rv*r + gv*g + bv*b + (0x10001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
} }
} }
...@@ -170,9 +171,12 @@ static av_always_inline void rgb48ToUV_half_c_template(uint16_t *dstU, ...@@ -170,9 +171,12 @@ static av_always_inline void rgb48ToUV_half_c_template(uint16_t *dstU,
const uint16_t *src1, const uint16_t *src1,
const uint16_t *src2, const uint16_t *src2,
int width, int width,
enum AVPixelFormat origin) enum AVPixelFormat origin,
int32_t *rgb2yuv)
{ {
int i; int i;
int32_t ru = rgb2yuv[RU_IDX], gu = rgb2yuv[GU_IDX], bu = rgb2yuv[BU_IDX];
int32_t rv = rgb2yuv[RV_IDX], gv = rgb2yuv[GV_IDX], bv = rgb2yuv[BV_IDX];
av_assert1(src1 == src2); av_assert1(src1 == src2);
for (i = 0; i < width; i++) { for (i = 0; i < width; i++) {
int r_b = (input_pixel(&src1[6 * i + 0]) + int r_b = (input_pixel(&src1[6 * i + 0]) +
...@@ -182,8 +186,8 @@ static av_always_inline void rgb48ToUV_half_c_template(uint16_t *dstU, ...@@ -182,8 +186,8 @@ static av_always_inline void rgb48ToUV_half_c_template(uint16_t *dstU,
int b_r = (input_pixel(&src1[6 * i + 2]) + int b_r = (input_pixel(&src1[6 * i + 2]) +
input_pixel(&src1[6 * i + 5]) + 1) >> 1; input_pixel(&src1[6 * i + 5]) + 1) >> 1;
dstU[i] = (RU * r + GU * g + BU * b + (0x10001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT; dstU[i] = (ru*r + gu*g + bu*b + (0x10001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
dstV[i] = (RV * r + GV * g + BV * b + (0x10001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT; dstV[i] = (rv*r + gv*g + bv*b + (0x10001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
} }
} }
...@@ -196,11 +200,11 @@ static void pattern ## 48 ## BE_LE ## ToY_c(uint8_t *_dst, \ ...@@ -196,11 +200,11 @@ static void pattern ## 48 ## BE_LE ## ToY_c(uint8_t *_dst, \
const uint8_t *_src, \ const uint8_t *_src, \
const uint8_t *unused0, const uint8_t *unused1,\ const uint8_t *unused0, const uint8_t *unused1,\
int width, \ int width, \
uint32_t *unused) \ uint32_t *rgb2yuv) \
{ \ { \
const uint16_t *src = (const uint16_t *)_src; \ const uint16_t *src = (const uint16_t *)_src; \
uint16_t *dst = (uint16_t *)_dst; \ uint16_t *dst = (uint16_t *)_dst; \
rgb48ToY_c_template(dst, src, width, origin); \ rgb48ToY_c_template(dst, src, width, origin, rgb2yuv); \
} \ } \
\ \
static void pattern ## 48 ## BE_LE ## ToUV_c(uint8_t *_dstU, \ static void pattern ## 48 ## BE_LE ## ToUV_c(uint8_t *_dstU, \
...@@ -209,13 +213,13 @@ static void pattern ## 48 ## BE_LE ## ToUV_c(uint8_t *_dstU, \ ...@@ -209,13 +213,13 @@ static void pattern ## 48 ## BE_LE ## ToUV_c(uint8_t *_dstU, \
const uint8_t *_src1, \ const uint8_t *_src1, \
const uint8_t *_src2, \ const uint8_t *_src2, \
int width, \ int width, \
uint32_t *unused) \ uint32_t *rgb2yuv) \
{ \ { \
const uint16_t *src1 = (const uint16_t *)_src1, \ const uint16_t *src1 = (const uint16_t *)_src1, \
*src2 = (const uint16_t *)_src2; \ *src2 = (const uint16_t *)_src2; \
uint16_t *dstU = (uint16_t *)_dstU, \ uint16_t *dstU = (uint16_t *)_dstU, \
*dstV = (uint16_t *)_dstV; \ *dstV = (uint16_t *)_dstV; \
rgb48ToUV_c_template(dstU, dstV, src1, src2, width, origin); \ rgb48ToUV_c_template(dstU, dstV, src1, src2, width, origin, rgb2yuv); \
} \ } \
\ \
static void pattern ## 48 ## BE_LE ## ToUV_half_c(uint8_t *_dstU, \ static void pattern ## 48 ## BE_LE ## ToUV_half_c(uint8_t *_dstU, \
...@@ -224,13 +228,13 @@ static void pattern ## 48 ## BE_LE ## ToUV_half_c(uint8_t *_dstU, \ ...@@ -224,13 +228,13 @@ static void pattern ## 48 ## BE_LE ## ToUV_half_c(uint8_t *_dstU, \
const uint8_t *_src1, \ const uint8_t *_src1, \
const uint8_t *_src2, \ const uint8_t *_src2, \
int width, \ int width, \
uint32_t *unused) \ uint32_t *rgb2yuv) \
{ \ { \
const uint16_t *src1 = (const uint16_t *)_src1, \ const uint16_t *src1 = (const uint16_t *)_src1, \
*src2 = (const uint16_t *)_src2; \ *src2 = (const uint16_t *)_src2; \
uint16_t *dstU = (uint16_t *)_dstU, \ uint16_t *dstU = (uint16_t *)_dstU, \
*dstV = (uint16_t *)_dstV; \ *dstV = (uint16_t *)_dstV; \
rgb48ToUV_half_c_template(dstU, dstV, src1, src2, width, origin); \ rgb48ToUV_half_c_template(dstU, dstV, src1, src2, width, origin, rgb2yuv); \
} }
rgb48funcs(rgb, LE, AV_PIX_FMT_RGB48LE) rgb48funcs(rgb, LE, AV_PIX_FMT_RGB48LE)
...@@ -254,9 +258,10 @@ static av_always_inline void rgb16_32ToY_c_template(int16_t *dst, ...@@ -254,9 +258,10 @@ static av_always_inline void rgb16_32ToY_c_template(int16_t *dst,
int shb, int shp, int shb, int shp,
int maskr, int maskg, int maskr, int maskg,
int maskb, int rsh, int maskb, int rsh,
int gsh, int bsh, int S) int gsh, int bsh, int S,
int32_t *rgb2yuv)
{ {
const int ry = RY << rsh, gy = GY << gsh, by = BY << bsh; const int ry = rgb2yuv[RY_IDX]<<rsh, gy = rgb2yuv[GY_IDX]<<gsh, by = rgb2yuv[BY_IDX]<<bsh;
const unsigned rnd = (32<<((S)-1)) + (1<<(S-7)); const unsigned rnd = (32<<((S)-1)) + (1<<(S-7));
int i; int i;
...@@ -279,10 +284,11 @@ static av_always_inline void rgb16_32ToUV_c_template(int16_t *dstU, ...@@ -279,10 +284,11 @@ static av_always_inline void rgb16_32ToUV_c_template(int16_t *dstU,
int shb, int shp, int shb, int shp,
int maskr, int maskg, int maskr, int maskg,
int maskb, int rsh, int maskb, int rsh,
int gsh, int bsh, int S) int gsh, int bsh, int S,
int32_t *rgb2yuv)
{ {
const int ru = RU << rsh, gu = GU << gsh, bu = BU << bsh, const int ru = rgb2yuv[RU_IDX] << rsh, gu = rgb2yuv[GU_IDX] << gsh, bu = rgb2yuv[BU_IDX] << bsh,
rv = RV << rsh, gv = GV << gsh, bv = BV << bsh; rv = rgb2yuv[RV_IDX] << rsh, gv = rgb2yuv[GV_IDX] << gsh, bv = rgb2yuv[BV_IDX] << bsh;
const unsigned rnd = (256u<<((S)-1)) + (1<<(S-7)); const unsigned rnd = (256u<<((S)-1)) + (1<<(S-7));
int i; int i;
...@@ -306,10 +312,11 @@ static av_always_inline void rgb16_32ToUV_half_c_template(int16_t *dstU, ...@@ -306,10 +312,11 @@ static av_always_inline void rgb16_32ToUV_half_c_template(int16_t *dstU,
int shb, int shp, int shb, int shp,
int maskr, int maskg, int maskr, int maskg,
int maskb, int rsh, int maskb, int rsh,
int gsh, int bsh, int S) int gsh, int bsh, int S,
int32_t *rgb2yuv)
{ {
const int ru = RU << rsh, gu = GU << gsh, bu = BU << bsh, const int ru = rgb2yuv[RU_IDX] << rsh, gu = rgb2yuv[GU_IDX] << gsh, bu = rgb2yuv[BU_IDX] << bsh,
rv = RV << rsh, gv = GV << gsh, bv = BV << bsh, rv = rgb2yuv[RV_IDX] << rsh, gv = rgb2yuv[GV_IDX] << gsh, bv = rgb2yuv[BV_IDX] << bsh,
maskgx = ~(maskr | maskb); maskgx = ~(maskr | maskb);
const unsigned rnd = (256U<<(S)) + (1<<(S-6)); const unsigned rnd = (256U<<(S)) + (1<<(S-6));
int i; int i;
...@@ -343,30 +350,30 @@ static av_always_inline void rgb16_32ToUV_half_c_template(int16_t *dstU, ...@@ -343,30 +350,30 @@ static av_always_inline void rgb16_32ToUV_half_c_template(int16_t *dstU,
#define rgb16_32_wrapper(fmt, name, shr, shg, shb, shp, maskr, \ #define rgb16_32_wrapper(fmt, name, shr, shg, shb, shp, maskr, \
maskg, maskb, rsh, gsh, bsh, S) \ maskg, maskb, rsh, gsh, bsh, S) \
static void name ## ToY_c(uint8_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, \ static void name ## ToY_c(uint8_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, \
int width, uint32_t *unused) \ int width, uint32_t *tab) \
{ \ { \
rgb16_32ToY_c_template((int16_t*)dst, src, width, fmt, shr, shg, shb, shp, \ rgb16_32ToY_c_template((int16_t*)dst, src, width, fmt, shr, shg, shb, shp, \
maskr, maskg, maskb, rsh, gsh, bsh, S); \ maskr, maskg, maskb, rsh, gsh, bsh, S, tab); \
} \ } \
\ \
static void name ## ToUV_c(uint8_t *dstU, uint8_t *dstV, \ static void name ## ToUV_c(uint8_t *dstU, uint8_t *dstV, \
const uint8_t *unused0, const uint8_t *src, const uint8_t *dummy, \ const uint8_t *unused0, const uint8_t *src, const uint8_t *dummy, \
int width, uint32_t *unused) \ int width, uint32_t *tab) \
{ \ { \
rgb16_32ToUV_c_template((int16_t*)dstU, (int16_t*)dstV, src, width, fmt, \ rgb16_32ToUV_c_template((int16_t*)dstU, (int16_t*)dstV, src, width, fmt, \
shr, shg, shb, shp, \ shr, shg, shb, shp, \
maskr, maskg, maskb, rsh, gsh, bsh, S); \ maskr, maskg, maskb, rsh, gsh, bsh, S, tab);\
} \ } \
\ \
static void name ## ToUV_half_c(uint8_t *dstU, uint8_t *dstV, \ static void name ## ToUV_half_c(uint8_t *dstU, uint8_t *dstV, \
const uint8_t *unused0, const uint8_t *src, \ const uint8_t *unused0, const uint8_t *src, \
const uint8_t *dummy, \ const uint8_t *dummy, \
int width, uint32_t *unused) \ int width, uint32_t *tab) \
{ \ { \
rgb16_32ToUV_half_c_template((int16_t*)dstU, (int16_t*)dstV, src, width, fmt, \ rgb16_32ToUV_half_c_template((int16_t*)dstU, (int16_t*)dstV, src, width, fmt, \
shr, shg, shb, shp, \ shr, shg, shb, shp, \
maskr, maskg, maskb, \ maskr, maskg, maskb, \
rsh, gsh, bsh, S); \ rsh, gsh, bsh, S, tab); \
} }
rgb16_32_wrapper(AV_PIX_FMT_BGR32, bgr32, 16, 0, 0, 0, 0xFF0000, 0xFF00, 0x00FF, 8, 0, 8, RGB2YUV_SHIFT + 8) rgb16_32_wrapper(AV_PIX_FMT_BGR32, bgr32, 16, 0, 0, 0, 0xFF0000, 0xFF00, 0x00FF, 8, 0, 8, RGB2YUV_SHIFT + 8)
...@@ -388,18 +395,21 @@ rgb16_32_wrapper(AV_PIX_FMT_RGB444BE, rgb12be, 0, 0, 0, 0, 0x0F00, 0x00F0, ...@@ -388,18 +395,21 @@ rgb16_32_wrapper(AV_PIX_FMT_RGB444BE, rgb12be, 0, 0, 0, 0, 0x0F00, 0x00F0,
static void gbr24pToUV_half_c(uint8_t *_dstU, uint8_t *_dstV, static void gbr24pToUV_half_c(uint8_t *_dstU, uint8_t *_dstV,
const uint8_t *gsrc, const uint8_t *bsrc, const uint8_t *rsrc, const uint8_t *gsrc, const uint8_t *bsrc, const uint8_t *rsrc,
int width, uint32_t *unused) int width, uint32_t *rgb2yuv)
{ {
uint16_t *dstU = (uint16_t *)_dstU; uint16_t *dstU = (uint16_t *)_dstU;
uint16_t *dstV = (uint16_t *)_dstV; uint16_t *dstV = (uint16_t *)_dstV;
int32_t ru = rgb2yuv[RU_IDX], gu = rgb2yuv[GU_IDX], bu = rgb2yuv[BU_IDX];
int32_t rv = rgb2yuv[RV_IDX], gv = rgb2yuv[GV_IDX], bv = rgb2yuv[BV_IDX];
int i; int i;
for (i = 0; i < width; i++) { for (i = 0; i < width; i++) {
unsigned int g = gsrc[2*i] + gsrc[2*i+1]; unsigned int g = gsrc[2*i] + gsrc[2*i+1];
unsigned int b = bsrc[2*i] + bsrc[2*i+1]; unsigned int b = bsrc[2*i] + bsrc[2*i+1];
unsigned int r = rsrc[2*i] + rsrc[2*i+1]; unsigned int r = rsrc[2*i] + rsrc[2*i+1];
dstU[i] = (RU*r + GU*g + BU*b + (0x4001<<(RGB2YUV_SHIFT-6))) >> (RGB2YUV_SHIFT-6+1); dstU[i] = (ru*r + gu*g + bu*b + (0x4001<<(RGB2YUV_SHIFT-6))) >> (RGB2YUV_SHIFT-6+1);
dstV[i] = (RV*r + GV*g + BV*b + (0x4001<<(RGB2YUV_SHIFT-6))) >> (RGB2YUV_SHIFT-6+1); dstV[i] = (rv*r + gv*g + bv*b + (0x4001<<(RGB2YUV_SHIFT-6))) >> (RGB2YUV_SHIFT-6+1);
} }
} }
...@@ -593,273 +603,289 @@ static void nv21ToUV_c(uint8_t *dstU, uint8_t *dstV, ...@@ -593,273 +603,289 @@ static void nv21ToUV_c(uint8_t *dstU, uint8_t *dstV,
#define input_pixel(pos) (isBE(origin) ? AV_RB16(pos) : AV_RL16(pos)) #define input_pixel(pos) (isBE(origin) ? AV_RB16(pos) : AV_RL16(pos))
static void bgr24ToY_c(uint8_t *_dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, static void bgr24ToY_c(uint8_t *_dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2,
int width, uint32_t *unused) int width, uint32_t *rgb2yuv)
{ {
int16_t *dst = (int16_t *)_dst; int16_t *dst = (int16_t *)_dst;
int32_t ry = rgb2yuv[RY_IDX], gy = rgb2yuv[GY_IDX], by = rgb2yuv[BY_IDX];
int i; int i;
for (i = 0; i < width; i++) { for (i = 0; i < width; i++) {
int b = src[i * 3 + 0]; int b = src[i * 3 + 0];
int g = src[i * 3 + 1]; int g = src[i * 3 + 1];
int r = src[i * 3 + 2]; int r = src[i * 3 + 2];
dst[i] = ((RY*r + GY*g + BY*b + (32<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6)); dst[i] = ((ry*r + gy*g + by*b + (32<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6));
} }
} }
static void bgr24ToUV_c(uint8_t *_dstU, uint8_t *_dstV, const uint8_t *unused0, const uint8_t *src1, static void bgr24ToUV_c(uint8_t *_dstU, uint8_t *_dstV, const uint8_t *unused0, const uint8_t *src1,
const uint8_t *src2, int width, uint32_t *unused) const uint8_t *src2, int width, uint32_t *rgb2yuv)
{ {
int16_t *dstU = (int16_t *)_dstU; int16_t *dstU = (int16_t *)_dstU;
int16_t *dstV = (int16_t *)_dstV; int16_t *dstV = (int16_t *)_dstV;
int32_t ru = rgb2yuv[RU_IDX], gu = rgb2yuv[GU_IDX], bu = rgb2yuv[BU_IDX];
int32_t rv = rgb2yuv[RV_IDX], gv = rgb2yuv[GV_IDX], bv = rgb2yuv[BV_IDX];
int i; int i;
for (i = 0; i < width; i++) { for (i = 0; i < width; i++) {
int b = src1[3 * i + 0]; int b = src1[3 * i + 0];
int g = src1[3 * i + 1]; int g = src1[3 * i + 1];
int r = src1[3 * i + 2]; int r = src1[3 * i + 2];
dstU[i] = (RU*r + GU*g + BU*b + (256<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6); dstU[i] = (ru*r + gu*g + bu*b + (256<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6);
dstV[i] = (RV*r + GV*g + BV*b + (256<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6); dstV[i] = (rv*r + gv*g + bv*b + (256<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6);
} }
av_assert1(src1 == src2); av_assert1(src1 == src2);
} }
static void bgr24ToUV_half_c(uint8_t *_dstU, uint8_t *_dstV, const uint8_t *unused0, const uint8_t *src1, static void bgr24ToUV_half_c(uint8_t *_dstU, uint8_t *_dstV, const uint8_t *unused0, const uint8_t *src1,
const uint8_t *src2, int width, uint32_t *unused) const uint8_t *src2, int width, uint32_t *rgb2yuv)
{ {
int16_t *dstU = (int16_t *)_dstU; int16_t *dstU = (int16_t *)_dstU;
int16_t *dstV = (int16_t *)_dstV; int16_t *dstV = (int16_t *)_dstV;
int i; int i;
int32_t ru = rgb2yuv[RU_IDX], gu = rgb2yuv[GU_IDX], bu = rgb2yuv[BU_IDX];
int32_t rv = rgb2yuv[RV_IDX], gv = rgb2yuv[GV_IDX], bv = rgb2yuv[BV_IDX];
for (i = 0; i < width; i++) { for (i = 0; i < width; i++) {
int b = src1[6 * i + 0] + src1[6 * i + 3]; int b = src1[6 * i + 0] + src1[6 * i + 3];
int g = src1[6 * i + 1] + src1[6 * i + 4]; int g = src1[6 * i + 1] + src1[6 * i + 4];
int r = src1[6 * i + 2] + src1[6 * i + 5]; int r = src1[6 * i + 2] + src1[6 * i + 5];
dstU[i] = (RU*r + GU*g + BU*b + (256<<RGB2YUV_SHIFT) + (1<<(RGB2YUV_SHIFT-6)))>>(RGB2YUV_SHIFT-5); dstU[i] = (ru*r + gu*g + bu*b + (256<<RGB2YUV_SHIFT) + (1<<(RGB2YUV_SHIFT-6)))>>(RGB2YUV_SHIFT-5);
dstV[i] = (RV*r + GV*g + BV*b + (256<<RGB2YUV_SHIFT) + (1<<(RGB2YUV_SHIFT-6)))>>(RGB2YUV_SHIFT-5); dstV[i] = (rv*r + gv*g + bv*b + (256<<RGB2YUV_SHIFT) + (1<<(RGB2YUV_SHIFT-6)))>>(RGB2YUV_SHIFT-5);
} }
av_assert1(src1 == src2); av_assert1(src1 == src2);
} }
static void rgb24ToY_c(uint8_t *_dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, int width, static void rgb24ToY_c(uint8_t *_dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, int width,
uint32_t *unused) uint32_t *rgb2yuv)
{ {
int16_t *dst = (int16_t *)_dst; int16_t *dst = (int16_t *)_dst;
int32_t ry = rgb2yuv[RY_IDX], gy = rgb2yuv[GY_IDX], by = rgb2yuv[BY_IDX];
int i; int i;
for (i = 0; i < width; i++) { for (i = 0; i < width; i++) {
int r = src[i * 3 + 0]; int r = src[i * 3 + 0];
int g = src[i * 3 + 1]; int g = src[i * 3 + 1];
int b = src[i * 3 + 2]; int b = src[i * 3 + 2];
dst[i] = ((RY*r + GY*g + BY*b + (32<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6)); dst[i] = ((ry*r + gy*g + by*b + (32<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6));
} }
} }
static void rgb24ToUV_c(uint8_t *_dstU, uint8_t *_dstV, const uint8_t *unused0, const uint8_t *src1, static void rgb24ToUV_c(uint8_t *_dstU, uint8_t *_dstV, const uint8_t *unused0, const uint8_t *src1,
const uint8_t *src2, int width, uint32_t *unused) const uint8_t *src2, int width, uint32_t *rgb2yuv)
{ {
int16_t *dstU = (int16_t *)_dstU; int16_t *dstU = (int16_t *)_dstU;
int16_t *dstV = (int16_t *)_dstV; int16_t *dstV = (int16_t *)_dstV;
int i; int i;
int32_t ru = rgb2yuv[RU_IDX], gu = rgb2yuv[GU_IDX], bu = rgb2yuv[BU_IDX];
int32_t rv = rgb2yuv[RV_IDX], gv = rgb2yuv[GV_IDX], bv = rgb2yuv[BV_IDX];
av_assert1(src1 == src2); av_assert1(src1 == src2);
for (i = 0; i < width; i++) { for (i = 0; i < width; i++) {
int r = src1[3 * i + 0]; int r = src1[3 * i + 0];
int g = src1[3 * i + 1]; int g = src1[3 * i + 1];
int b = src1[3 * i + 2]; int b = src1[3 * i + 2];
dstU[i] = (RU*r + GU*g + BU*b + (256<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6); dstU[i] = (ru*r + gu*g + bu*b + (256<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6);
dstV[i] = (RV*r + GV*g + BV*b + (256<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6); dstV[i] = (rv*r + gv*g + bv*b + (256<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6);
} }
} }
static void rgb24ToUV_half_c(uint8_t *_dstU, uint8_t *_dstV, const uint8_t *unused0, const uint8_t *src1, static void rgb24ToUV_half_c(uint8_t *_dstU, uint8_t *_dstV, const uint8_t *unused0, const uint8_t *src1,
const uint8_t *src2, int width, uint32_t *unused) const uint8_t *src2, int width, uint32_t *rgb2yuv)
{ {
int16_t *dstU = (int16_t *)_dstU; int16_t *dstU = (int16_t *)_dstU;
int16_t *dstV = (int16_t *)_dstV; int16_t *dstV = (int16_t *)_dstV;
int i; int i;
int32_t ru = rgb2yuv[RU_IDX], gu = rgb2yuv[GU_IDX], bu = rgb2yuv[BU_IDX];
int32_t rv = rgb2yuv[RV_IDX], gv = rgb2yuv[GV_IDX], bv = rgb2yuv[BV_IDX];
av_assert1(src1 == src2); av_assert1(src1 == src2);
for (i = 0; i < width; i++) { for (i = 0; i < width; i++) {
int r = src1[6 * i + 0] + src1[6 * i + 3]; int r = src1[6 * i + 0] + src1[6 * i + 3];
int g = src1[6 * i + 1] + src1[6 * i + 4]; int g = src1[6 * i + 1] + src1[6 * i + 4];
int b = src1[6 * i + 2] + src1[6 * i + 5]; int b = src1[6 * i + 2] + src1[6 * i + 5];
dstU[i] = (RU*r + GU*g + BU*b + (256<<RGB2YUV_SHIFT) + (1<<(RGB2YUV_SHIFT-6)))>>(RGB2YUV_SHIFT-5); dstU[i] = (ru*r + gu*g + bu*b + (256<<RGB2YUV_SHIFT) + (1<<(RGB2YUV_SHIFT-6)))>>(RGB2YUV_SHIFT-5);
dstV[i] = (RV*r + GV*g + BV*b + (256<<RGB2YUV_SHIFT) + (1<<(RGB2YUV_SHIFT-6)))>>(RGB2YUV_SHIFT-5); dstV[i] = (rv*r + gv*g + bv*b + (256<<RGB2YUV_SHIFT) + (1<<(RGB2YUV_SHIFT-6)))>>(RGB2YUV_SHIFT-5);
} }
} }
static void planar_rgb_to_y(uint8_t *_dst, const uint8_t *src[4], int width) static void planar_rgb_to_y(uint8_t *_dst, const uint8_t *src[4], int width, int32_t *rgb2yuv)
{ {
uint16_t *dst = (uint16_t *)_dst; uint16_t *dst = (uint16_t *)_dst;
int32_t ry = rgb2yuv[RY_IDX], gy = rgb2yuv[GY_IDX], by = rgb2yuv[BY_IDX];
int i; int i;
for (i = 0; i < width; i++) { for (i = 0; i < width; i++) {
int g = src[0][i]; int g = src[0][i];
int b = src[1][i]; int b = src[1][i];
int r = src[2][i]; int r = src[2][i];
dst[i] = (RY*r + GY*g + BY*b + (0x801<<(RGB2YUV_SHIFT-7))) >> (RGB2YUV_SHIFT-6); dst[i] = (ry*r + gy*g + by*b + (0x801<<(RGB2YUV_SHIFT-7))) >> (RGB2YUV_SHIFT-6);
} }
} }
static void planar_rgb_to_uv(uint8_t *_dstU, uint8_t *_dstV, const uint8_t *src[4], int width) static void planar_rgb_to_uv(uint8_t *_dstU, uint8_t *_dstV, const uint8_t *src[4], int width, int32_t *rgb2yuv)
{ {
uint16_t *dstU = (uint16_t *)_dstU; uint16_t *dstU = (uint16_t *)_dstU;
uint16_t *dstV = (uint16_t *)_dstV; uint16_t *dstV = (uint16_t *)_dstV;
int32_t ru = rgb2yuv[RU_IDX], gu = rgb2yuv[GU_IDX], bu = rgb2yuv[BU_IDX];
int32_t rv = rgb2yuv[RV_IDX], gv = rgb2yuv[GV_IDX], bv = rgb2yuv[BV_IDX];
int i; int i;
for (i = 0; i < width; i++) { for (i = 0; i < width; i++) {
int g = src[0][i]; int g = src[0][i];
int b = src[1][i]; int b = src[1][i];
int r = src[2][i]; int r = src[2][i];
dstU[i] = (RU*r + GU*g + BU*b + (0x4001<<(RGB2YUV_SHIFT-7))) >> (RGB2YUV_SHIFT-6); dstU[i] = (ru*r + gu*g + bu*b + (0x4001<<(RGB2YUV_SHIFT-7))) >> (RGB2YUV_SHIFT-6);
dstV[i] = (RV*r + GV*g + BV*b + (0x4001<<(RGB2YUV_SHIFT-7))) >> (RGB2YUV_SHIFT-6); dstV[i] = (rv*r + gv*g + bv*b + (0x4001<<(RGB2YUV_SHIFT-7))) >> (RGB2YUV_SHIFT-6);
} }
} }
#define rdpx(src) \ #define rdpx(src) \
is_be ? AV_RB16(src) : AV_RL16(src) is_be ? AV_RB16(src) : AV_RL16(src)
static av_always_inline void planar_rgb16_to_y(uint8_t *_dst, const uint8_t *_src[4], static av_always_inline void planar_rgb16_to_y(uint8_t *_dst, const uint8_t *_src[4],
int width, int bpc, int is_be) int width, int bpc, int is_be, int32_t *rgb2yuv)
{ {
int i; int i;
const uint16_t **src = (const uint16_t **)_src; const uint16_t **src = (const uint16_t **)_src;
uint16_t *dst = (uint16_t *)_dst; uint16_t *dst = (uint16_t *)_dst;
int32_t ry = rgb2yuv[RY_IDX], gy = rgb2yuv[GY_IDX], by = rgb2yuv[BY_IDX];
for (i = 0; i < width; i++) { for (i = 0; i < width; i++) {
int g = rdpx(src[0] + i); int g = rdpx(src[0] + i);
int b = rdpx(src[1] + i); int b = rdpx(src[1] + i);
int r = rdpx(src[2] + i); int r = rdpx(src[2] + i);
dst[i] = ((RY * r + GY * g + BY * b + (33 << (RGB2YUV_SHIFT + bpc - 9))) >> (RGB2YUV_SHIFT + bpc - 14)); dst[i] = ((ry*r + gy*g + by*b + (33 << (RGB2YUV_SHIFT + bpc - 9))) >> (RGB2YUV_SHIFT + bpc - 14));
} }
} }
static void planar_rgb9le_to_y(uint8_t *dst, const uint8_t *src[4], int w) static void planar_rgb9le_to_y(uint8_t *dst, const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_y(dst, src, w, 9, 0); planar_rgb16_to_y(dst, src, w, 9, 0, rgb2yuv);
} }
static void planar_rgb9be_to_y(uint8_t *dst, const uint8_t *src[4], int w) static void planar_rgb9be_to_y(uint8_t *dst, const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_y(dst, src, w, 9, 1); planar_rgb16_to_y(dst, src, w, 9, 1, rgb2yuv);
} }
static void planar_rgb10le_to_y(uint8_t *dst, const uint8_t *src[4], int w) static void planar_rgb10le_to_y(uint8_t *dst, const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_y(dst, src, w, 10, 0); planar_rgb16_to_y(dst, src, w, 10, 0, rgb2yuv);
} }
static void planar_rgb10be_to_y(uint8_t *dst, const uint8_t *src[4], int w) static void planar_rgb10be_to_y(uint8_t *dst, const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_y(dst, src, w, 10, 1); planar_rgb16_to_y(dst, src, w, 10, 1, rgb2yuv);
} }
static void planar_rgb12le_to_y(uint8_t *dst, const uint8_t *src[4], int w) static void planar_rgb12le_to_y(uint8_t *dst, const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_y(dst, src, w, 12, 0); planar_rgb16_to_y(dst, src, w, 12, 0, rgb2yuv);
} }
static void planar_rgb12be_to_y(uint8_t *dst, const uint8_t *src[4], int w) static void planar_rgb12be_to_y(uint8_t *dst, const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_y(dst, src, w, 12, 1); planar_rgb16_to_y(dst, src, w, 12, 1, rgb2yuv);
} }
static void planar_rgb14le_to_y(uint8_t *dst, const uint8_t *src[4], int w) static void planar_rgb14le_to_y(uint8_t *dst, const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_y(dst, src, w, 14, 0); planar_rgb16_to_y(dst, src, w, 14, 0, rgb2yuv);
} }
static void planar_rgb14be_to_y(uint8_t *dst, const uint8_t *src[4], int w) static void planar_rgb14be_to_y(uint8_t *dst, const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_y(dst, src, w, 14, 1); planar_rgb16_to_y(dst, src, w, 14, 1, rgb2yuv);
} }
static void planar_rgb16le_to_y(uint8_t *dst, const uint8_t *src[4], int w) static void planar_rgb16le_to_y(uint8_t *dst, const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_y(dst, src, w, 16, 0); planar_rgb16_to_y(dst, src, w, 16, 0, rgb2yuv);
} }
static void planar_rgb16be_to_y(uint8_t *dst, const uint8_t *src[4], int w) static void planar_rgb16be_to_y(uint8_t *dst, const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_y(dst, src, w, 16, 1); planar_rgb16_to_y(dst, src, w, 16, 1, rgb2yuv);
} }
static av_always_inline void planar_rgb16_to_uv(uint8_t *_dstU, uint8_t *_dstV, static av_always_inline void planar_rgb16_to_uv(uint8_t *_dstU, uint8_t *_dstV,
const uint8_t *_src[4], int width, const uint8_t *_src[4], int width,
int bpc, int is_be) int bpc, int is_be, int32_t *rgb2yuv)
{ {
int i; int i;
const uint16_t **src = (const uint16_t **)_src; const uint16_t **src = (const uint16_t **)_src;
uint16_t *dstU = (uint16_t *)_dstU; uint16_t *dstU = (uint16_t *)_dstU;
uint16_t *dstV = (uint16_t *)_dstV; uint16_t *dstV = (uint16_t *)_dstV;
int32_t ru = rgb2yuv[RU_IDX], gu = rgb2yuv[GU_IDX], bu = rgb2yuv[BU_IDX];
int32_t rv = rgb2yuv[RV_IDX], gv = rgb2yuv[GV_IDX], bv = rgb2yuv[BV_IDX];
for (i = 0; i < width; i++) { for (i = 0; i < width; i++) {
int g = rdpx(src[0] + i); int g = rdpx(src[0] + i);
int b = rdpx(src[1] + i); int b = rdpx(src[1] + i);
int r = rdpx(src[2] + i); int r = rdpx(src[2] + i);
dstU[i] = (RU * r + GU * g + BU * b + (257 << (RGB2YUV_SHIFT + bpc - 9))) >> (RGB2YUV_SHIFT + bpc - 14); dstU[i] = (ru*r + gu*g + bu*b + (257 << (RGB2YUV_SHIFT + bpc - 9))) >> (RGB2YUV_SHIFT + bpc - 14);
dstV[i] = (RV * r + GV * g + BV * b + (257 << (RGB2YUV_SHIFT + bpc - 9))) >> (RGB2YUV_SHIFT + bpc - 14); dstV[i] = (rv*r + gv*g + bv*b + (257 << (RGB2YUV_SHIFT + bpc - 9))) >> (RGB2YUV_SHIFT + bpc - 14);
} }
} }
#undef rdpx #undef rdpx
static void planar_rgb9le_to_uv(uint8_t *dstU, uint8_t *dstV, static void planar_rgb9le_to_uv(uint8_t *dstU, uint8_t *dstV,
const uint8_t *src[4], int w) const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_uv(dstU, dstV, src, w, 9, 0); planar_rgb16_to_uv(dstU, dstV, src, w, 9, 0, rgb2yuv);
} }
static void planar_rgb9be_to_uv(uint8_t *dstU, uint8_t *dstV, static void planar_rgb9be_to_uv(uint8_t *dstU, uint8_t *dstV,
const uint8_t *src[4], int w) const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_uv(dstU, dstV, src, w, 9, 1); planar_rgb16_to_uv(dstU, dstV, src, w, 9, 1, rgb2yuv);
} }
static void planar_rgb10le_to_uv(uint8_t *dstU, uint8_t *dstV, static void planar_rgb10le_to_uv(uint8_t *dstU, uint8_t *dstV,
const uint8_t *src[4], int w) const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_uv(dstU, dstV, src, w, 10, 0); planar_rgb16_to_uv(dstU, dstV, src, w, 10, 0, rgb2yuv);
} }
static void planar_rgb10be_to_uv(uint8_t *dstU, uint8_t *dstV, static void planar_rgb10be_to_uv(uint8_t *dstU, uint8_t *dstV,
const uint8_t *src[4], int w) const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_uv(dstU, dstV, src, w, 10, 1); planar_rgb16_to_uv(dstU, dstV, src, w, 10, 1, rgb2yuv);
} }
static void planar_rgb12le_to_uv(uint8_t *dstU, uint8_t *dstV, static void planar_rgb12le_to_uv(uint8_t *dstU, uint8_t *dstV,
const uint8_t *src[4], int w) const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_uv(dstU, dstV, src, w, 12, 0); planar_rgb16_to_uv(dstU, dstV, src, w, 12, 0, rgb2yuv);
} }
static void planar_rgb12be_to_uv(uint8_t *dstU, uint8_t *dstV, static void planar_rgb12be_to_uv(uint8_t *dstU, uint8_t *dstV,
const uint8_t *src[4], int w) const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_uv(dstU, dstV, src, w, 12, 1); planar_rgb16_to_uv(dstU, dstV, src, w, 12, 1, rgb2yuv);
} }
static void planar_rgb14le_to_uv(uint8_t *dstU, uint8_t *dstV, static void planar_rgb14le_to_uv(uint8_t *dstU, uint8_t *dstV,
const uint8_t *src[4], int w) const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_uv(dstU, dstV, src, w, 14, 0); planar_rgb16_to_uv(dstU, dstV, src, w, 14, 0, rgb2yuv);
} }
static void planar_rgb14be_to_uv(uint8_t *dstU, uint8_t *dstV, static void planar_rgb14be_to_uv(uint8_t *dstU, uint8_t *dstV,
const uint8_t *src[4], int w) const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_uv(dstU, dstV, src, w, 14, 1); planar_rgb16_to_uv(dstU, dstV, src, w, 14, 1, rgb2yuv);
} }
static void planar_rgb16le_to_uv(uint8_t *dstU, uint8_t *dstV, static void planar_rgb16le_to_uv(uint8_t *dstU, uint8_t *dstV,
const uint8_t *src[4], int w) const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_uv(dstU, dstV, src, w, 16, 0); planar_rgb16_to_uv(dstU, dstV, src, w, 16, 0, rgb2yuv);
} }
static void planar_rgb16be_to_uv(uint8_t *dstU, uint8_t *dstV, static void planar_rgb16be_to_uv(uint8_t *dstU, uint8_t *dstV,
const uint8_t *src[4], int w) const uint8_t *src[4], int w, int32_t *rgb2yuv)
{ {
planar_rgb16_to_uv(dstU, dstV, src, w, 16, 1); planar_rgb16_to_uv(dstU, dstV, src, w, 16, 1, rgb2yuv);
} }
av_cold void ff_sws_init_input_funcs(SwsContext *c) av_cold void ff_sws_init_input_funcs(SwsContext *c)
......
libswscale/swscale.c
View file @ d4956b0b
...@@ -254,7 +254,7 @@ static av_always_inline void hyscale(SwsContext *c, int16_t *dst, int dstWidth, ...@@ -254,7 +254,7 @@ static av_always_inline void hyscale(SwsContext *c, int16_t *dst, int dstWidth,
toYV12(formatConvBuffer, src, src_in[1], src_in[2], srcW, pal); toYV12(formatConvBuffer, src, src_in[1], src_in[2], srcW, pal);
src = formatConvBuffer; src = formatConvBuffer;
} else if (c->readLumPlanar && !isAlpha) { } else if (c->readLumPlanar && !isAlpha) {
c->readLumPlanar(formatConvBuffer, src_in, srcW); c->readLumPlanar(formatConvBuffer, src_in, srcW, c->input_rgb2yuv_table);
src = formatConvBuffer; src = formatConvBuffer;
} }
...@@ -307,7 +307,7 @@ static av_always_inline void hcscale(SwsContext *c, int16_t *dst1, ...@@ -307,7 +307,7 @@ static av_always_inline void hcscale(SwsContext *c, int16_t *dst1,
} else if (c->readChrPlanar) { } else if (c->readChrPlanar) {
uint8_t *buf2 = formatConvBuffer + uint8_t *buf2 = formatConvBuffer +
FFALIGN(srcW*2+78, 16); FFALIGN(srcW*2+78, 16);
c->readChrPlanar(formatConvBuffer, buf2, src_in, srcW); c->readChrPlanar(formatConvBuffer, buf2, src_in, srcW, c->input_rgb2yuv_table);
src1 = formatConvBuffer; src1 = formatConvBuffer;
src2 = buf2; src2 = buf2;
} }
...@@ -383,6 +383,20 @@ static int swScale(SwsContext *c, const uint8_t *src[], ...@@ -383,6 +383,20 @@ static int swScale(SwsContext *c, const uint8_t *src[],
int lastInLumBuf = c->lastInLumBuf; int lastInLumBuf = c->lastInLumBuf;
int lastInChrBuf = c->lastInChrBuf; int lastInChrBuf = c->lastInChrBuf;
if (!usePal(c->srcFormat)) {
pal = c->input_rgb2yuv_table;
#define RGB2YUV_SHIFT 15
pal[BY_IDX] = ((int)(0.114 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
pal[BV_IDX] = (-(int)(0.081 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
pal[BU_IDX] = ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
pal[GY_IDX] = ((int)(0.587 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
pal[GV_IDX] = (-(int)(0.419 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
pal[GU_IDX] = (-(int)(0.331 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
pal[RY_IDX] = ((int)(0.299 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
pal[RV_IDX] = ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
pal[RU_IDX] = (-(int)(0.169 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
}
if (isPacked(c->srcFormat)) { if (isPacked(c->srcFormat)) {
src[0] = src[0] =
src[1] = src[1] =
......
libswscale/swscale_internal.h
View file @ d4956b0b
...@@ -360,6 +360,16 @@ typedef struct SwsContext { ...@@ -360,6 +360,16 @@ typedef struct SwsContext {
uint8_t *table_gU[256 + 2*YUVRGB_TABLE_HEADROOM]; uint8_t *table_gU[256 + 2*YUVRGB_TABLE_HEADROOM];
int table_gV[256 + 2*YUVRGB_TABLE_HEADROOM]; int table_gV[256 + 2*YUVRGB_TABLE_HEADROOM];
uint8_t *table_bU[256 + 2*YUVRGB_TABLE_HEADROOM]; uint8_t *table_bU[256 + 2*YUVRGB_TABLE_HEADROOM];
int32_t input_rgb2yuv_table[16]; // This table can contain both C and SIMD formatted values, teh C vales are always at the XY_IDX points
#define RY_IDX 0
#define GY_IDX 1
#define BY_IDX 2
#define RU_IDX 3
#define GU_IDX 4
#define BU_IDX 5
#define RV_IDX 6
#define GV_IDX 7
#define BV_IDX 8
int *dither_error[4]; int *dither_error[4];
...@@ -489,9 +499,9 @@ typedef struct SwsContext { ...@@ -489,9 +499,9 @@ typedef struct SwsContext {
* internally to Y/UV. * internally to Y/UV.
*/ */
/** @{ */ /** @{ */
void (*readLumPlanar)(uint8_t *dst, const uint8_t *src[4], int width); void (*readLumPlanar)(uint8_t *dst, const uint8_t *src[4], int width, int32_t *rgb2yuv);
void (*readChrPlanar)(uint8_t *dstU, uint8_t *dstV, const uint8_t *src[4], void (*readChrPlanar)(uint8_t *dstU, uint8_t *dstV, const uint8_t *src[4],
int width); int width, int32_t *rgb2yuv);
/** @} */ /** @} */
/** /**
......
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