diff options
Diffstat (limited to 'libswscale/rgb2rgb.c')
-rw-r--r-- | libswscale/rgb2rgb.c | 131 |
1 files changed, 124 insertions, 7 deletions
diff --git a/libswscale/rgb2rgb.c b/libswscale/rgb2rgb.c index 74cc42e3b0..adc5d59c8c 100644 --- a/libswscale/rgb2rgb.c +++ b/libswscale/rgb2rgb.c @@ -6,24 +6,25 @@ * Written by Nick Kurshev. * palette & YUV & runtime CPU stuff by Michael (michaelni@gmx.at) * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg 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 * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ #include <inttypes.h> #include "config.h" +#include "libavutil/x86_cpu.h" #include "libavutil/bswap.h" #include "rgb2rgb.h" #include "swscale.h" @@ -94,6 +95,45 @@ void (*yuyvtoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t long width, long height, long lumStride, long chromStride, long srcStride); + +#if ARCH_X86 +DECLARE_ASM_CONST(8, uint64_t, mmx_ff) = 0x00000000000000FFULL; +DECLARE_ASM_CONST(8, uint64_t, mmx_null) = 0x0000000000000000ULL; +DECLARE_ASM_CONST(8, uint64_t, mmx_one) = 0xFFFFFFFFFFFFFFFFULL; +DECLARE_ASM_CONST(8, uint64_t, mask32b) = 0x000000FF000000FFULL; +DECLARE_ASM_CONST(8, uint64_t, mask32g) = 0x0000FF000000FF00ULL; +DECLARE_ASM_CONST(8, uint64_t, mask32r) = 0x00FF000000FF0000ULL; +DECLARE_ASM_CONST(8, uint64_t, mask32a) = 0xFF000000FF000000ULL; +DECLARE_ASM_CONST(8, uint64_t, mask32) = 0x00FFFFFF00FFFFFFULL; +DECLARE_ASM_CONST(8, uint64_t, mask3216br) = 0x00F800F800F800F8ULL; +DECLARE_ASM_CONST(8, uint64_t, mask3216g) = 0x0000FC000000FC00ULL; +DECLARE_ASM_CONST(8, uint64_t, mask3215g) = 0x0000F8000000F800ULL; +DECLARE_ASM_CONST(8, uint64_t, mul3216) = 0x2000000420000004ULL; +DECLARE_ASM_CONST(8, uint64_t, mul3215) = 0x2000000820000008ULL; +DECLARE_ASM_CONST(8, uint64_t, mask24b) = 0x00FF0000FF0000FFULL; +DECLARE_ASM_CONST(8, uint64_t, mask24g) = 0xFF0000FF0000FF00ULL; +DECLARE_ASM_CONST(8, uint64_t, mask24r) = 0x0000FF0000FF0000ULL; +DECLARE_ASM_CONST(8, uint64_t, mask24l) = 0x0000000000FFFFFFULL; +DECLARE_ASM_CONST(8, uint64_t, mask24h) = 0x0000FFFFFF000000ULL; +DECLARE_ASM_CONST(8, uint64_t, mask24hh) = 0xffff000000000000ULL; +DECLARE_ASM_CONST(8, uint64_t, mask24hhh) = 0xffffffff00000000ULL; +DECLARE_ASM_CONST(8, uint64_t, mask24hhhh) = 0xffffffffffff0000ULL; +DECLARE_ASM_CONST(8, uint64_t, mask15b) = 0x001F001F001F001FULL; /* 00000000 00011111 xxB */ +DECLARE_ASM_CONST(8, uint64_t, mask15rg) = 0x7FE07FE07FE07FE0ULL; /* 01111111 11100000 RGx */ +DECLARE_ASM_CONST(8, uint64_t, mask15s) = 0xFFE0FFE0FFE0FFE0ULL; +DECLARE_ASM_CONST(8, uint64_t, mask15g) = 0x03E003E003E003E0ULL; +DECLARE_ASM_CONST(8, uint64_t, mask15r) = 0x7C007C007C007C00ULL; +#define mask16b mask15b +DECLARE_ASM_CONST(8, uint64_t, mask16g) = 0x07E007E007E007E0ULL; +DECLARE_ASM_CONST(8, uint64_t, mask16r) = 0xF800F800F800F800ULL; +DECLARE_ASM_CONST(8, uint64_t, red_16mask) = 0x0000f8000000f800ULL; +DECLARE_ASM_CONST(8, uint64_t, green_16mask) = 0x000007e0000007e0ULL; +DECLARE_ASM_CONST(8, uint64_t, blue_16mask) = 0x0000001f0000001fULL; +DECLARE_ASM_CONST(8, uint64_t, red_15mask) = 0x00007c0000007c00ULL; +DECLARE_ASM_CONST(8, uint64_t, green_15mask) = 0x000003e0000003e0ULL; +DECLARE_ASM_CONST(8, uint64_t, blue_15mask) = 0x0000001f0000001fULL; +#endif /* ARCH_X86 */ + #define RGB2YUV_SHIFT 8 #define BY ((int)( 0.098*(1<<RGB2YUV_SHIFT)+0.5)) #define BV ((int)(-0.071*(1<<RGB2YUV_SHIFT)+0.5)) @@ -105,9 +145,50 @@ void (*yuyvtoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t #define RV ((int)( 0.439*(1<<RGB2YUV_SHIFT)+0.5)) #define RU ((int)(-0.148*(1<<RGB2YUV_SHIFT)+0.5)) +//Note: We have C, MMX, MMX2, 3DNOW versions, there is no 3DNOW + MMX2 one. //plain C versions +#define COMPILE_TEMPLATE_MMX 0 +#define COMPILE_TEMPLATE_MMX2 0 +#define COMPILE_TEMPLATE_AMD3DNOW 0 +#define COMPILE_TEMPLATE_SSE2 0 +#define RENAME(a) a ## _C #include "rgb2rgb_template.c" +#if ARCH_X86 + +//MMX versions +#undef RENAME +#undef COMPILE_TEMPLATE_MMX +#define COMPILE_TEMPLATE_MMX 1 +#define RENAME(a) a ## _MMX +#include "rgb2rgb_template.c" + +//MMX2 versions +#undef RENAME +#undef COMPILE_TEMPLATE_MMX2 +#define COMPILE_TEMPLATE_MMX2 1 +#define RENAME(a) a ## _MMX2 +#include "rgb2rgb_template.c" + +//SSE2 versions +#undef RENAME +#undef COMPILE_TEMPLATE_SSE2 +#define COMPILE_TEMPLATE_SSE2 1 +#define RENAME(a) a ## _SSE2 +#include "rgb2rgb_template.c" + +//3DNOW versions +#undef RENAME +#undef COMPILE_TEMPLATE_MMX2 +#undef COMPILE_TEMPLATE_SSE2 +#undef COMPILE_TEMPLATE_AMD3DNOW +#define COMPILE_TEMPLATE_MMX2 0 +#define COMPILE_TEMPLATE_SSE2 1 +#define COMPILE_TEMPLATE_AMD3DNOW 1 +#define RENAME(a) a ## _3DNOW +#include "rgb2rgb_template.c" + +#endif //ARCH_X86 || ARCH_X86_64 /* RGB15->RGB16 original by Strepto/Astral @@ -118,12 +199,48 @@ void (*yuyvtoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t void sws_rgb2rgb_init(int flags) { - rgb2rgb_init_c(); -#if HAVE_MMX2 || HAVE_AMD3DNOW || HAVE_MMX - rgb2rgb_init_x86(flags); +#if HAVE_SSE2 || HAVE_MMX2 || HAVE_AMD3DNOW || HAVE_MMX + if (flags & SWS_CPU_CAPS_SSE2) + rgb2rgb_init_SSE2(); + else if (flags & SWS_CPU_CAPS_MMX2) + rgb2rgb_init_MMX2(); + else if (flags & SWS_CPU_CAPS_3DNOW) + rgb2rgb_init_3DNOW(); + else if (flags & SWS_CPU_CAPS_MMX) + rgb2rgb_init_MMX(); + else #endif /* HAVE_MMX2 || HAVE_AMD3DNOW || HAVE_MMX */ + rgb2rgb_init_C(); +} + +#if LIBSWSCALE_VERSION_MAJOR < 1 +void palette8topacked32(const uint8_t *src, uint8_t *dst, long num_pixels, const uint8_t *palette) +{ + sws_convertPalette8ToPacked32(src, dst, num_pixels, palette); +} + +void palette8topacked24(const uint8_t *src, uint8_t *dst, long num_pixels, const uint8_t *palette) +{ + sws_convertPalette8ToPacked24(src, dst, num_pixels, palette); } +/** + * Palette is assumed to contain BGR16, see rgb32to16 to convert the palette. + */ +void palette8torgb16(const uint8_t *src, uint8_t *dst, long num_pixels, const uint8_t *palette) +{ + long i; + for (i=0; i<num_pixels; i++) + ((uint16_t *)dst)[i] = ((const uint16_t *)palette)[src[i]]; +} +void palette8tobgr16(const uint8_t *src, uint8_t *dst, long num_pixels, const uint8_t *palette) +{ + long i; + for (i=0; i<num_pixels; i++) + ((uint16_t *)dst)[i] = av_bswap16(((const uint16_t *)palette)[src[i]]); +} +#endif + void rgb32to24(const uint8_t *src, uint8_t *dst, long src_size) { long i; |