yading@11: /*
yading@11:  * software RGB to RGB converter
yading@11:  * pluralize by software PAL8 to RGB converter
yading@11:  *              software YUV to YUV converter
yading@11:  *              software YUV to RGB converter
yading@11:  * Written by Nick Kurshev.
yading@11:  * palette & YUV & runtime CPU stuff by Michael (michaelni@gmx.at)
yading@11:  * lot of big-endian byte order fixes by Alex Beregszaszi
yading@11:  *
yading@11:  * This file is part of FFmpeg.
yading@11:  *
yading@11:  * FFmpeg is free software; you can redistribute it and/or
yading@11:  * modify it under the terms of the GNU Lesser General Public
yading@11:  * License as published by the Free Software Foundation; either
yading@11:  * version 2.1 of the License, or (at your option) any later version.
yading@11:  *
yading@11:  * FFmpeg is distributed in the hope that it will be useful,
yading@11:  * but WITHOUT ANY WARRANTY; without even the implied warranty of
yading@11:  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
yading@11:  * Lesser General Public License for more details.
yading@11:  *
yading@11:  * You should have received a copy of the GNU Lesser General Public
yading@11:  * License along with FFmpeg; if not, write to the Free Software
yading@11:  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
yading@11:  */
yading@11: 
yading@11: #include <stddef.h>
yading@11: 
yading@11: static inline void rgb24tobgr32_c(const uint8_t *src, uint8_t *dst,
yading@11:                                   int src_size)
yading@11: {
yading@11:     uint8_t *dest      = dst;
yading@11:     const uint8_t *s   = src;
yading@11:     const uint8_t *end = s + src_size;
yading@11: 
yading@11:     while (s < end) {
yading@11: #if HAVE_BIGENDIAN
yading@11:         /* RGB24 (= R, G, B) -> RGB32 (= A, B, G, R) */
yading@11:         *dest++  = 255;
yading@11:         *dest++  = s[2];
yading@11:         *dest++  = s[1];
yading@11:         *dest++  = s[0];
yading@11:         s       += 3;
yading@11: #else
yading@11:         *dest++  = *s++;
yading@11:         *dest++  = *s++;
yading@11:         *dest++  = *s++;
yading@11:         *dest++  = 255;
yading@11: #endif
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void rgb32tobgr24_c(const uint8_t *src, uint8_t *dst,
yading@11:                                   int src_size)
yading@11: {
yading@11:     uint8_t *dest      = dst;
yading@11:     const uint8_t *s   = src;
yading@11:     const uint8_t *end = s + src_size;
yading@11: 
yading@11:     while (s < end) {
yading@11: #if HAVE_BIGENDIAN
yading@11:         /* RGB32 (= A, B, G, R) -> RGB24 (= R, G, B) */
yading@11:         s++;
yading@11:         dest[2]  = *s++;
yading@11:         dest[1]  = *s++;
yading@11:         dest[0]  = *s++;
yading@11:         dest    += 3;
yading@11: #else
yading@11:         *dest++  = *s++;
yading@11:         *dest++  = *s++;
yading@11:         *dest++  = *s++;
yading@11:         s++;
yading@11: #endif
yading@11:     }
yading@11: }
yading@11: 
yading@11: /*
yading@11:  * original by Strepto/Astral
yading@11:  * ported to gcc & bugfixed: A'rpi
yading@11:  * MMXEXT, 3DNOW optimization by Nick Kurshev
yading@11:  * 32-bit C version, and and&add trick by Michael Niedermayer
yading@11:  */
yading@11: static inline void rgb15to16_c(const uint8_t *src, uint8_t *dst, int src_size)
yading@11: {
yading@11:     register uint8_t *d         = dst;
yading@11:     register const uint8_t *s   = src;
yading@11:     register const uint8_t *end = s + src_size;
yading@11:     const uint8_t *mm_end       = end - 3;
yading@11: 
yading@11:     while (s < mm_end) {
yading@11:         register unsigned x = *((const uint32_t *)s);
yading@11:         *((uint32_t *)d)    = (x & 0x7FFF7FFF) + (x & 0x7FE07FE0);
yading@11:         d += 4;
yading@11:         s += 4;
yading@11:     }
yading@11:     if (s < end) {
yading@11:         register unsigned short x = *((const uint16_t *)s);
yading@11:         *((uint16_t *)d)          = (x & 0x7FFF) + (x & 0x7FE0);
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void rgb16to15_c(const uint8_t *src, uint8_t *dst, int src_size)
yading@11: {
yading@11:     register uint8_t *d         = dst;
yading@11:     register const uint8_t *s   = src;
yading@11:     register const uint8_t *end = s + src_size;
yading@11:     const uint8_t *mm_end       = end - 3;
yading@11: 
yading@11:     while (s < mm_end) {
yading@11:         register uint32_t x  = *((const uint32_t *)s);
yading@11:         *((uint32_t *)d)     = ((x >> 1) & 0x7FE07FE0) | (x & 0x001F001F);
yading@11:         s                   += 4;
yading@11:         d                   += 4;
yading@11:     }
yading@11:     if (s < end) {
yading@11:         register uint16_t x = *((const uint16_t *)s);
yading@11:         *((uint16_t *)d)    = ((x >> 1) & 0x7FE0) | (x & 0x001F);
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void rgb32to16_c(const uint8_t *src, uint8_t *dst, int src_size)
yading@11: {
yading@11:     uint16_t *d        = (uint16_t *)dst;
yading@11:     const uint8_t *s   = src;
yading@11:     const uint8_t *end = s + src_size;
yading@11: 
yading@11:     while (s < end) {
yading@11:         register int rgb  = *(const uint32_t *)s;
yading@11:         s                += 4;
yading@11:         *d++              = ((rgb & 0xFF)     >> 3) +
yading@11:                             ((rgb & 0xFC00)   >> 5) +
yading@11:                             ((rgb & 0xF80000) >> 8);
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void rgb32tobgr16_c(const uint8_t *src, uint8_t *dst,
yading@11:                                   int src_size)
yading@11: {
yading@11:     uint16_t *d        = (uint16_t *)dst;
yading@11:     const uint8_t *s   = src;
yading@11:     const uint8_t *end = s + src_size;
yading@11: 
yading@11:     while (s < end) {
yading@11:         register int rgb  = *(const uint32_t *)s;
yading@11:         s                += 4;
yading@11:         *d++              = ((rgb & 0xF8)     << 8) +
yading@11:                             ((rgb & 0xFC00)   >> 5) +
yading@11:                             ((rgb & 0xF80000) >> 19);
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void rgb32to15_c(const uint8_t *src, uint8_t *dst, int src_size)
yading@11: {
yading@11:     uint16_t *d        = (uint16_t *)dst;
yading@11:     const uint8_t *s   = src;
yading@11:     const uint8_t *end = s + src_size;
yading@11: 
yading@11:     while (s < end) {
yading@11:         register int rgb  = *(const uint32_t *)s;
yading@11:         s                += 4;
yading@11:         *d++              = ((rgb & 0xFF)     >> 3) +
yading@11:                             ((rgb & 0xF800)   >> 6) +
yading@11:                             ((rgb & 0xF80000) >> 9);
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void rgb32tobgr15_c(const uint8_t *src, uint8_t *dst,
yading@11:                                   int src_size)
yading@11: {
yading@11:     uint16_t *d        = (uint16_t *)dst;
yading@11:     const uint8_t *s   = src;
yading@11:     const uint8_t *end = s + src_size;
yading@11: 
yading@11:     while (s < end) {
yading@11:         register int rgb  = *(const uint32_t *)s;
yading@11:         s                += 4;
yading@11:         *d++              = ((rgb & 0xF8)     <<  7) +
yading@11:                             ((rgb & 0xF800)   >>  6) +
yading@11:                             ((rgb & 0xF80000) >> 19);
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void rgb24tobgr16_c(const uint8_t *src, uint8_t *dst,
yading@11:                                   int src_size)
yading@11: {
yading@11:     uint16_t *d        = (uint16_t *)dst;
yading@11:     const uint8_t *s   = src;
yading@11:     const uint8_t *end = s + src_size;
yading@11: 
yading@11:     while (s < end) {
yading@11:         const int b = *s++;
yading@11:         const int g = *s++;
yading@11:         const int r = *s++;
yading@11:         *d++        = (b >> 3) | ((g & 0xFC) << 3) | ((r & 0xF8) << 8);
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void rgb24to16_c(const uint8_t *src, uint8_t *dst, int src_size)
yading@11: {
yading@11:     uint16_t *d        = (uint16_t *)dst;
yading@11:     const uint8_t *s   = src;
yading@11:     const uint8_t *end = s + src_size;
yading@11: 
yading@11:     while (s < end) {
yading@11:         const int r = *s++;
yading@11:         const int g = *s++;
yading@11:         const int b = *s++;
yading@11:         *d++        = (b >> 3) | ((g & 0xFC) << 3) | ((r & 0xF8) << 8);
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void rgb24tobgr15_c(const uint8_t *src, uint8_t *dst,
yading@11:                                   int src_size)
yading@11: {
yading@11:     uint16_t *d        = (uint16_t *)dst;
yading@11:     const uint8_t *s   = src;
yading@11:     const uint8_t *end = s + src_size;
yading@11: 
yading@11:     while (s < end) {
yading@11:         const int b = *s++;
yading@11:         const int g = *s++;
yading@11:         const int r = *s++;
yading@11:         *d++        = (b >> 3) | ((g & 0xF8) << 2) | ((r & 0xF8) << 7);
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void rgb24to15_c(const uint8_t *src, uint8_t *dst, int src_size)
yading@11: {
yading@11:     uint16_t *d        = (uint16_t *)dst;
yading@11:     const uint8_t *s   = src;
yading@11:     const uint8_t *end = s + src_size;
yading@11: 
yading@11:     while (s < end) {
yading@11:         const int r = *s++;
yading@11:         const int g = *s++;
yading@11:         const int b = *s++;
yading@11:         *d++        = (b >> 3) | ((g & 0xF8) << 2) | ((r & 0xF8) << 7);
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void rgb15tobgr24_c(const uint8_t *src, uint8_t *dst,
yading@11:                                   int src_size)
yading@11: {
yading@11:     uint8_t *d          = dst;
yading@11:     const uint16_t *s   = (const uint16_t *)src;
yading@11:     const uint16_t *end = s + src_size / 2;
yading@11: 
yading@11:     while (s < end) {
yading@11:         register uint16_t bgr = *s++;
yading@11:         *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
yading@11:         *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7);
yading@11:         *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void rgb16tobgr24_c(const uint8_t *src, uint8_t *dst,
yading@11:                                   int src_size)
yading@11: {
yading@11:     uint8_t *d          = (uint8_t *)dst;
yading@11:     const uint16_t *s   = (const uint16_t *)src;
yading@11:     const uint16_t *end = s + src_size / 2;
yading@11: 
yading@11:     while (s < end) {
yading@11:         register uint16_t bgr = *s++;
yading@11:         *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
yading@11:         *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9);
yading@11:         *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void rgb15to32_c(const uint8_t *src, uint8_t *dst, int src_size)
yading@11: {
yading@11:     uint8_t *d          = dst;
yading@11:     const uint16_t *s   = (const uint16_t *)src;
yading@11:     const uint16_t *end = s + src_size / 2;
yading@11: 
yading@11:     while (s < end) {
yading@11:         register uint16_t bgr = *s++;
yading@11: #if HAVE_BIGENDIAN
yading@11:         *d++ = 255;
yading@11:         *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
yading@11:         *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7);
yading@11:         *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
yading@11: #else
yading@11:         *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
yading@11:         *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7);
yading@11:         *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
yading@11:         *d++ = 255;
yading@11: #endif
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void rgb16to32_c(const uint8_t *src, uint8_t *dst, int src_size)
yading@11: {
yading@11:     uint8_t *d          = dst;
yading@11:     const uint16_t *s   = (const uint16_t *)src;
yading@11:     const uint16_t *end = s + src_size / 2;
yading@11: 
yading@11:     while (s < end) {
yading@11:         register uint16_t bgr = *s++;
yading@11: #if HAVE_BIGENDIAN
yading@11:         *d++ = 255;
yading@11:         *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
yading@11:         *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9);
yading@11:         *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
yading@11: #else
yading@11:         *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
yading@11:         *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9);
yading@11:         *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
yading@11:         *d++ = 255;
yading@11: #endif
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void shuffle_bytes_2103_c(const uint8_t *src, uint8_t *dst,
yading@11:                                         int src_size)
yading@11: {
yading@11:     int idx          = 15  - src_size;
yading@11:     const uint8_t *s = src - idx;
yading@11:     uint8_t *d       = dst - idx;
yading@11: 
yading@11:     for (; idx < 15; idx += 4) {
yading@11:         register int v        = *(const uint32_t *)&s[idx], g = v & 0xff00ff00;
yading@11:         v                    &= 0xff00ff;
yading@11:         *(uint32_t *)&d[idx]  = (v >> 16) + g + (v << 16);
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void rgb24tobgr24_c(const uint8_t *src, uint8_t *dst, int src_size)
yading@11: {
yading@11:     unsigned i;
yading@11: 
yading@11:     for (i = 0; i < src_size; i += 3) {
yading@11:         register uint8_t x = src[i + 2];
yading@11:         dst[i + 1]         = src[i + 1];
yading@11:         dst[i + 2]         = src[i + 0];
yading@11:         dst[i + 0]         = x;
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void yuvPlanartoyuy2_c(const uint8_t *ysrc, const uint8_t *usrc,
yading@11:                                      const uint8_t *vsrc, uint8_t *dst,
yading@11:                                      int width, int height,
yading@11:                                      int lumStride, int chromStride,
yading@11:                                      int dstStride, int vertLumPerChroma)
yading@11: {
yading@11:     int y, i;
yading@11:     const int chromWidth = width >> 1;
yading@11: 
yading@11:     for (y = 0; y < height; y++) {
yading@11: #if HAVE_FAST_64BIT
yading@11:         uint64_t *ldst = (uint64_t *)dst;
yading@11:         const uint8_t *yc = ysrc, *uc = usrc, *vc = vsrc;
yading@11:         for (i = 0; i < chromWidth; i += 2) {
yading@11:             uint64_t k = yc[0] + (uc[0] << 8) +
yading@11:                          (yc[1] << 16) + (unsigned)(vc[0] << 24);
yading@11:             uint64_t l = yc[2] + (uc[1] << 8) +
yading@11:                          (yc[3] << 16) + (unsigned)(vc[1] << 24);
yading@11:             *ldst++ = k + (l << 32);
yading@11:             yc     += 4;
yading@11:             uc     += 2;
yading@11:             vc     += 2;
yading@11:         }
yading@11: 
yading@11: #else
yading@11:         int *idst = (int32_t *)dst;
yading@11:         const uint8_t *yc = ysrc, *uc = usrc, *vc = vsrc;
yading@11: 
yading@11:         for (i = 0; i < chromWidth; i++) {
yading@11: #if HAVE_BIGENDIAN
yading@11:             *idst++ = (yc[0] << 24) + (uc[0] << 16) +
yading@11:                       (yc[1] <<  8) + (vc[0] <<  0);
yading@11: #else
yading@11:             *idst++ = yc[0] + (uc[0] << 8) +
yading@11:                       (yc[1] << 16) + (vc[0] << 24);
yading@11: #endif
yading@11:             yc += 2;
yading@11:             uc++;
yading@11:             vc++;
yading@11:         }
yading@11: #endif
yading@11:         if ((y & (vertLumPerChroma - 1)) == vertLumPerChroma - 1) {
yading@11:             usrc += chromStride;
yading@11:             vsrc += chromStride;
yading@11:         }
yading@11:         ysrc += lumStride;
yading@11:         dst  += dstStride;
yading@11:     }
yading@11: }
yading@11: 
yading@11: /**
yading@11:  * Height should be a multiple of 2 and width should be a multiple of 16.
yading@11:  * (If this is a problem for anyone then tell me, and I will fix it.)
yading@11:  */
yading@11: static inline void yv12toyuy2_c(const uint8_t *ysrc, const uint8_t *usrc,
yading@11:                                 const uint8_t *vsrc, uint8_t *dst,
yading@11:                                 int width, int height, int lumStride,
yading@11:                                 int chromStride, int dstStride)
yading@11: {
yading@11:     //FIXME interpolate chroma
yading@11:     yuvPlanartoyuy2_c(ysrc, usrc, vsrc, dst, width, height, lumStride,
yading@11:                       chromStride, dstStride, 2);
yading@11: }
yading@11: 
yading@11: static inline void yuvPlanartouyvy_c(const uint8_t *ysrc, const uint8_t *usrc,
yading@11:                                      const uint8_t *vsrc, uint8_t *dst,
yading@11:                                      int width, int height,
yading@11:                                      int lumStride, int chromStride,
yading@11:                                      int dstStride, int vertLumPerChroma)
yading@11: {
yading@11:     int y, i;
yading@11:     const int chromWidth = width >> 1;
yading@11: 
yading@11:     for (y = 0; y < height; y++) {
yading@11: #if HAVE_FAST_64BIT
yading@11:         uint64_t *ldst = (uint64_t *)dst;
yading@11:         const uint8_t *yc = ysrc, *uc = usrc, *vc = vsrc;
yading@11:         for (i = 0; i < chromWidth; i += 2) {
yading@11:             uint64_t k = uc[0] + (yc[0] << 8) +
yading@11:                          (vc[0] << 16) + (unsigned)(yc[1] << 24);
yading@11:             uint64_t l = uc[1] + (yc[2] << 8) +
yading@11:                          (vc[1] << 16) + (unsigned)(yc[3] << 24);
yading@11:             *ldst++ = k + (l << 32);
yading@11:             yc     += 4;
yading@11:             uc     += 2;
yading@11:             vc     += 2;
yading@11:         }
yading@11: 
yading@11: #else
yading@11:         int *idst = (int32_t *)dst;
yading@11:         const uint8_t *yc = ysrc, *uc = usrc, *vc = vsrc;
yading@11: 
yading@11:         for (i = 0; i < chromWidth; i++) {
yading@11: #if HAVE_BIGENDIAN
yading@11:             *idst++ = (uc[0] << 24) + (yc[0] << 16) +
yading@11:                       (vc[0] <<  8) + (yc[1] <<  0);
yading@11: #else
yading@11:             *idst++ = uc[0] + (yc[0] << 8) +
yading@11:                       (vc[0] << 16) + (yc[1] << 24);
yading@11: #endif
yading@11:             yc += 2;
yading@11:             uc++;
yading@11:             vc++;
yading@11:         }
yading@11: #endif
yading@11:         if ((y & (vertLumPerChroma - 1)) == vertLumPerChroma - 1) {
yading@11:             usrc += chromStride;
yading@11:             vsrc += chromStride;
yading@11:         }
yading@11:         ysrc += lumStride;
yading@11:         dst  += dstStride;
yading@11:     }
yading@11: }
yading@11: 
yading@11: /**
yading@11:  * Height should be a multiple of 2 and width should be a multiple of 16
yading@11:  * (If this is a problem for anyone then tell me, and I will fix it.)
yading@11:  */
yading@11: static inline void yv12touyvy_c(const uint8_t *ysrc, const uint8_t *usrc,
yading@11:                                 const uint8_t *vsrc, uint8_t *dst,
yading@11:                                 int width, int height, int lumStride,
yading@11:                                 int chromStride, int dstStride)
yading@11: {
yading@11:     //FIXME interpolate chroma
yading@11:     yuvPlanartouyvy_c(ysrc, usrc, vsrc, dst, width, height, lumStride,
yading@11:                       chromStride, dstStride, 2);
yading@11: }
yading@11: 
yading@11: /**
yading@11:  * Width should be a multiple of 16.
yading@11:  */
yading@11: static inline void yuv422ptouyvy_c(const uint8_t *ysrc, const uint8_t *usrc,
yading@11:                                    const uint8_t *vsrc, uint8_t *dst,
yading@11:                                    int width, int height, int lumStride,
yading@11:                                    int chromStride, int dstStride)
yading@11: {
yading@11:     yuvPlanartouyvy_c(ysrc, usrc, vsrc, dst, width, height, lumStride,
yading@11:                       chromStride, dstStride, 1);
yading@11: }
yading@11: 
yading@11: /**
yading@11:  * Width should be a multiple of 16.
yading@11:  */
yading@11: static inline void yuv422ptoyuy2_c(const uint8_t *ysrc, const uint8_t *usrc,
yading@11:                                    const uint8_t *vsrc, uint8_t *dst,
yading@11:                                    int width, int height, int lumStride,
yading@11:                                    int chromStride, int dstStride)
yading@11: {
yading@11:     yuvPlanartoyuy2_c(ysrc, usrc, vsrc, dst, width, height, lumStride,
yading@11:                       chromStride, dstStride, 1);
yading@11: }
yading@11: 
yading@11: /**
yading@11:  * Height should be a multiple of 2 and width should be a multiple of 16.
yading@11:  * (If this is a problem for anyone then tell me, and I will fix it.)
yading@11:  */
yading@11: static inline void yuy2toyv12_c(const uint8_t *src, uint8_t *ydst,
yading@11:                                 uint8_t *udst, uint8_t *vdst,
yading@11:                                 int width, int height, int lumStride,
yading@11:                                 int chromStride, int srcStride)
yading@11: {
yading@11:     int y;
yading@11:     const int chromWidth = width >> 1;
yading@11: 
yading@11:     for (y = 0; y < height; y += 2) {
yading@11:         int i;
yading@11:         for (i = 0; i < chromWidth; i++) {
yading@11:             ydst[2 * i + 0] = src[4 * i + 0];
yading@11:             udst[i]         = src[4 * i + 1];
yading@11:             ydst[2 * i + 1] = src[4 * i + 2];
yading@11:             vdst[i]         = src[4 * i + 3];
yading@11:         }
yading@11:         ydst += lumStride;
yading@11:         src  += srcStride;
yading@11: 
yading@11:         for (i = 0; i < chromWidth; i++) {
yading@11:             ydst[2 * i + 0] = src[4 * i + 0];
yading@11:             ydst[2 * i + 1] = src[4 * i + 2];
yading@11:         }
yading@11:         udst += chromStride;
yading@11:         vdst += chromStride;
yading@11:         ydst += lumStride;
yading@11:         src  += srcStride;
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void planar2x_c(const uint8_t *src, uint8_t *dst, int srcWidth,
yading@11:                               int srcHeight, int srcStride, int dstStride)
yading@11: {
yading@11:     int x, y;
yading@11: 
yading@11:     dst[0] = src[0];
yading@11: 
yading@11:     // first line
yading@11:     for (x = 0; x < srcWidth - 1; x++) {
yading@11:         dst[2 * x + 1] = (3 * src[x] + src[x + 1]) >> 2;
yading@11:         dst[2 * x + 2] = (src[x] + 3 * src[x + 1]) >> 2;
yading@11:     }
yading@11:     dst[2 * srcWidth - 1] = src[srcWidth - 1];
yading@11: 
yading@11:     dst += dstStride;
yading@11: 
yading@11:     for (y = 1; y < srcHeight; y++) {
yading@11:         const int mmxSize = 1;
yading@11: 
yading@11:         dst[0]         = (src[0] * 3 + src[srcStride]) >> 2;
yading@11:         dst[dstStride] = (src[0] + 3 * src[srcStride]) >> 2;
yading@11: 
yading@11:         for (x = mmxSize - 1; x < srcWidth - 1; x++) {
yading@11:             dst[2 * x + 1]             = (src[x + 0] * 3 + src[x + srcStride + 1]) >> 2;
yading@11:             dst[2 * x + dstStride + 2] = (src[x + 0] + 3 * src[x + srcStride + 1]) >> 2;
yading@11:             dst[2 * x + dstStride + 1] = (src[x + 1] + 3 * src[x + srcStride])     >> 2;
yading@11:             dst[2 * x + 2]             = (src[x + 1] * 3 + src[x + srcStride])     >> 2;
yading@11:         }
yading@11:         dst[srcWidth * 2 - 1]             = (src[srcWidth - 1] * 3 + src[srcWidth - 1 + srcStride]) >> 2;
yading@11:         dst[srcWidth * 2 - 1 + dstStride] = (src[srcWidth - 1] + 3 * src[srcWidth - 1 + srcStride]) >> 2;
yading@11: 
yading@11:         dst += dstStride * 2;
yading@11:         src += srcStride;
yading@11:     }
yading@11: 
yading@11:     // last line
yading@11:     dst[0] = src[0];
yading@11: 
yading@11:     for (x = 0; x < srcWidth - 1; x++) {
yading@11:         dst[2 * x + 1] = (src[x] * 3 + src[x + 1]) >> 2;
yading@11:         dst[2 * x + 2] = (src[x] + 3 * src[x + 1]) >> 2;
yading@11:     }
yading@11:     dst[2 * srcWidth - 1] = src[srcWidth - 1];
yading@11: }
yading@11: 
yading@11: /**
yading@11:  * Height should be a multiple of 2 and width should be a multiple of 16.
yading@11:  * (If this is a problem for anyone then tell me, and I will fix it.)
yading@11:  * Chrominance data is only taken from every second line, others are ignored.
yading@11:  * FIXME: Write HQ version.
yading@11:  */
yading@11: static inline void uyvytoyv12_c(const uint8_t *src, uint8_t *ydst,
yading@11:                                 uint8_t *udst, uint8_t *vdst,
yading@11:                                 int width, int height, int lumStride,
yading@11:                                 int chromStride, int srcStride)
yading@11: {
yading@11:     int y;
yading@11:     const int chromWidth = width >> 1;
yading@11: 
yading@11:     for (y = 0; y < height; y += 2) {
yading@11:         int i;
yading@11:         for (i = 0; i < chromWidth; i++) {
yading@11:             udst[i]         = src[4 * i + 0];
yading@11:             ydst[2 * i + 0] = src[4 * i + 1];
yading@11:             vdst[i]         = src[4 * i + 2];
yading@11:             ydst[2 * i + 1] = src[4 * i + 3];
yading@11:         }
yading@11:         ydst += lumStride;
yading@11:         src  += srcStride;
yading@11: 
yading@11:         for (i = 0; i < chromWidth; i++) {
yading@11:             ydst[2 * i + 0] = src[4 * i + 1];
yading@11:             ydst[2 * i + 1] = src[4 * i + 3];
yading@11:         }
yading@11:         udst += chromStride;
yading@11:         vdst += chromStride;
yading@11:         ydst += lumStride;
yading@11:         src  += srcStride;
yading@11:     }
yading@11: }
yading@11: 
yading@11: /**
yading@11:  * Height should be a multiple of 2 and width should be a multiple of 2.
yading@11:  * (If this is a problem for anyone then tell me, and I will fix it.)
yading@11:  * Chrominance data is only taken from every second line,
yading@11:  * others are ignored in the C version.
yading@11:  * FIXME: Write HQ version.
yading@11:  */
yading@11: void ff_rgb24toyv12_c(const uint8_t *src, uint8_t *ydst, uint8_t *udst,
yading@11:                    uint8_t *vdst, int width, int height, int lumStride,
yading@11:                    int chromStride, int srcStride, int32_t *rgb2yuv)
yading@11: {
yading@11:     int32_t ry = rgb2yuv[RY_IDX], gy = rgb2yuv[GY_IDX], by = rgb2yuv[BY_IDX];
yading@11:     int32_t ru = rgb2yuv[RU_IDX], gu = rgb2yuv[GU_IDX], bu = rgb2yuv[BU_IDX];
yading@11:     int32_t rv = rgb2yuv[RV_IDX], gv = rgb2yuv[GV_IDX], bv = rgb2yuv[BV_IDX];
yading@11:     int y;
yading@11:     const int chromWidth = width >> 1;
yading@11: 
yading@11:     for (y = 0; y < height; y += 2) {
yading@11:         int i;
yading@11:         for (i = 0; i < chromWidth; i++) {
yading@11:             unsigned int b = src[6 * i + 0];
yading@11:             unsigned int g = src[6 * i + 1];
yading@11:             unsigned int r = src[6 * i + 2];
yading@11: 
yading@11:             unsigned int Y = ((ry * r + gy * g + by * b) >> RGB2YUV_SHIFT) +  16;
yading@11:             unsigned int V = ((rv * r + gv * g + bv * b) >> RGB2YUV_SHIFT) + 128;
yading@11:             unsigned int U = ((ru * r + gu * g + bu * b) >> RGB2YUV_SHIFT) + 128;
yading@11: 
yading@11:             udst[i]     = U;
yading@11:             vdst[i]     = V;
yading@11:             ydst[2 * i] = Y;
yading@11: 
yading@11:             b = src[6 * i + 3];
yading@11:             g = src[6 * i + 4];
yading@11:             r = src[6 * i + 5];
yading@11: 
yading@11:             Y = ((ry * r + gy * g + by * b) >> RGB2YUV_SHIFT) + 16;
yading@11:             ydst[2 * i + 1] = Y;
yading@11:         }
yading@11:         ydst += lumStride;
yading@11:         src  += srcStride;
yading@11: 
yading@11:         if (y+1 == height)
yading@11:             break;
yading@11: 
yading@11:         for (i = 0; i < chromWidth; i++) {
yading@11:             unsigned int b = src[6 * i + 0];
yading@11:             unsigned int g = src[6 * i + 1];
yading@11:             unsigned int r = src[6 * i + 2];
yading@11: 
yading@11:             unsigned int Y = ((ry * r + gy * g + by * b) >> RGB2YUV_SHIFT) + 16;
yading@11: 
yading@11:             ydst[2 * i] = Y;
yading@11: 
yading@11:             b = src[6 * i + 3];
yading@11:             g = src[6 * i + 4];
yading@11:             r = src[6 * i + 5];
yading@11: 
yading@11:             Y = ((ry * r + gy * g + by * b) >> RGB2YUV_SHIFT) + 16;
yading@11:             ydst[2 * i + 1] = Y;
yading@11:         }
yading@11:         udst += chromStride;
yading@11:         vdst += chromStride;
yading@11:         ydst += lumStride;
yading@11:         src  += srcStride;
yading@11:     }
yading@11: }
yading@11: 
yading@11: static void interleaveBytes_c(const uint8_t *src1, const uint8_t *src2,
yading@11:                               uint8_t *dest, int width, int height,
yading@11:                               int src1Stride, int src2Stride, int dstStride)
yading@11: {
yading@11:     int h;
yading@11: 
yading@11:     for (h = 0; h < height; h++) {
yading@11:         int w;
yading@11:         for (w = 0; w < width; w++) {
yading@11:             dest[2 * w + 0] = src1[w];
yading@11:             dest[2 * w + 1] = src2[w];
yading@11:         }
yading@11:         dest += dstStride;
yading@11:         src1 += src1Stride;
yading@11:         src2 += src2Stride;
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void vu9_to_vu12_c(const uint8_t *src1, const uint8_t *src2,
yading@11:                                  uint8_t *dst1, uint8_t *dst2,
yading@11:                                  int width, int height,
yading@11:                                  int srcStride1, int srcStride2,
yading@11:                                  int dstStride1, int dstStride2)
yading@11: {
yading@11:     int x, y;
yading@11:     int w = width  / 2;
yading@11:     int h = height / 2;
yading@11: 
yading@11:     for (y = 0; y < h; y++) {
yading@11:         const uint8_t *s1 = src1 + srcStride1 * (y >> 1);
yading@11:         uint8_t *d        = dst1 + dstStride1 *  y;
yading@11:         for (x = 0; x < w; x++)
yading@11:             d[2 * x] = d[2 * x + 1] = s1[x];
yading@11:     }
yading@11:     for (y = 0; y < h; y++) {
yading@11:         const uint8_t *s2 = src2 + srcStride2 * (y >> 1);
yading@11:         uint8_t *d        = dst2 + dstStride2 *  y;
yading@11:         for (x = 0; x < w; x++)
yading@11:             d[2 * x] = d[2 * x + 1] = s2[x];
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void yvu9_to_yuy2_c(const uint8_t *src1, const uint8_t *src2,
yading@11:                                   const uint8_t *src3, uint8_t *dst,
yading@11:                                   int width, int height,
yading@11:                                   int srcStride1, int srcStride2,
yading@11:                                   int srcStride3, int dstStride)
yading@11: {
yading@11:     int x, y;
yading@11:     int w = width / 2;
yading@11:     int h = height;
yading@11: 
yading@11:     for (y = 0; y < h; y++) {
yading@11:         const uint8_t *yp = src1 + srcStride1 *  y;
yading@11:         const uint8_t *up = src2 + srcStride2 * (y >> 2);
yading@11:         const uint8_t *vp = src3 + srcStride3 * (y >> 2);
yading@11:         uint8_t *d        = dst  + dstStride  *  y;
yading@11:         for (x = 0; x < w; x++) {
yading@11:             const int x2 = x << 2;
yading@11:             d[8 * x + 0] = yp[x2];
yading@11:             d[8 * x + 1] = up[x];
yading@11:             d[8 * x + 2] = yp[x2 + 1];
yading@11:             d[8 * x + 3] = vp[x];
yading@11:             d[8 * x + 4] = yp[x2 + 2];
yading@11:             d[8 * x + 5] = up[x];
yading@11:             d[8 * x + 6] = yp[x2 + 3];
yading@11:             d[8 * x + 7] = vp[x];
yading@11:         }
yading@11:     }
yading@11: }
yading@11: 
yading@11: static void extract_even_c(const uint8_t *src, uint8_t *dst, int count)
yading@11: {
yading@11:     dst   +=  count;
yading@11:     src   +=  count * 2;
yading@11:     count  = -count;
yading@11:     while (count < 0) {
yading@11:         dst[count] = src[2 * count];
yading@11:         count++;
yading@11:     }
yading@11: }
yading@11: 
yading@11: static void extract_even2_c(const uint8_t *src, uint8_t *dst0, uint8_t *dst1,
yading@11:                             int count)
yading@11: {
yading@11:     dst0  +=  count;
yading@11:     dst1  +=  count;
yading@11:     src   +=  count * 4;
yading@11:     count  = -count;
yading@11:     while (count < 0) {
yading@11:         dst0[count] = src[4 * count + 0];
yading@11:         dst1[count] = src[4 * count + 2];
yading@11:         count++;
yading@11:     }
yading@11: }
yading@11: 
yading@11: static void extract_even2avg_c(const uint8_t *src0, const uint8_t *src1,
yading@11:                                uint8_t *dst0, uint8_t *dst1, int count)
yading@11: {
yading@11:     dst0  +=  count;
yading@11:     dst1  +=  count;
yading@11:     src0  +=  count * 4;
yading@11:     src1  +=  count * 4;
yading@11:     count  = -count;
yading@11:     while (count < 0) {
yading@11:         dst0[count] = (src0[4 * count + 0] + src1[4 * count + 0]) >> 1;
yading@11:         dst1[count] = (src0[4 * count + 2] + src1[4 * count + 2]) >> 1;
yading@11:         count++;
yading@11:     }
yading@11: }
yading@11: 
yading@11: static void extract_odd2_c(const uint8_t *src, uint8_t *dst0, uint8_t *dst1,
yading@11:                            int count)
yading@11: {
yading@11:     dst0  +=  count;
yading@11:     dst1  +=  count;
yading@11:     src   +=  count * 4;
yading@11:     count  = -count;
yading@11:     src++;
yading@11:     while (count < 0) {
yading@11:         dst0[count] = src[4 * count + 0];
yading@11:         dst1[count] = src[4 * count + 2];
yading@11:         count++;
yading@11:     }
yading@11: }
yading@11: 
yading@11: static void extract_odd2avg_c(const uint8_t *src0, const uint8_t *src1,
yading@11:                               uint8_t *dst0, uint8_t *dst1, int count)
yading@11: {
yading@11:     dst0  +=  count;
yading@11:     dst1  +=  count;
yading@11:     src0  +=  count * 4;
yading@11:     src1  +=  count * 4;
yading@11:     count  = -count;
yading@11:     src0++;
yading@11:     src1++;
yading@11:     while (count < 0) {
yading@11:         dst0[count] = (src0[4 * count + 0] + src1[4 * count + 0]) >> 1;
yading@11:         dst1[count] = (src0[4 * count + 2] + src1[4 * count + 2]) >> 1;
yading@11:         count++;
yading@11:     }
yading@11: }
yading@11: 
yading@11: static void yuyvtoyuv420_c(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
yading@11:                            const uint8_t *src, int width, int height,
yading@11:                            int lumStride, int chromStride, int srcStride)
yading@11: {
yading@11:     int y;
yading@11:     const int chromWidth = -((-width) >> 1);
yading@11: 
yading@11:     for (y = 0; y < height; y++) {
yading@11:         extract_even_c(src, ydst, width);
yading@11:         if (y & 1) {
yading@11:             extract_odd2avg_c(src - srcStride, src, udst, vdst, chromWidth);
yading@11:             udst += chromStride;
yading@11:             vdst += chromStride;
yading@11:         }
yading@11: 
yading@11:         src  += srcStride;
yading@11:         ydst += lumStride;
yading@11:     }
yading@11: }
yading@11: 
yading@11: static void yuyvtoyuv422_c(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
yading@11:                            const uint8_t *src, int width, int height,
yading@11:                            int lumStride, int chromStride, int srcStride)
yading@11: {
yading@11:     int y;
yading@11:     const int chromWidth = -((-width) >> 1);
yading@11: 
yading@11:     for (y = 0; y < height; y++) {
yading@11:         extract_even_c(src, ydst, width);
yading@11:         extract_odd2_c(src, udst, vdst, chromWidth);
yading@11: 
yading@11:         src  += srcStride;
yading@11:         ydst += lumStride;
yading@11:         udst += chromStride;
yading@11:         vdst += chromStride;
yading@11:     }
yading@11: }
yading@11: 
yading@11: static void uyvytoyuv420_c(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
yading@11:                            const uint8_t *src, int width, int height,
yading@11:                            int lumStride, int chromStride, int srcStride)
yading@11: {
yading@11:     int y;
yading@11:     const int chromWidth = -((-width) >> 1);
yading@11: 
yading@11:     for (y = 0; y < height; y++) {
yading@11:         extract_even_c(src + 1, ydst, width);
yading@11:         if (y & 1) {
yading@11:             extract_even2avg_c(src - srcStride, src, udst, vdst, chromWidth);
yading@11:             udst += chromStride;
yading@11:             vdst += chromStride;
yading@11:         }
yading@11: 
yading@11:         src  += srcStride;
yading@11:         ydst += lumStride;
yading@11:     }
yading@11: }
yading@11: 
yading@11: static void uyvytoyuv422_c(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
yading@11:                            const uint8_t *src, int width, int height,
yading@11:                            int lumStride, int chromStride, int srcStride)
yading@11: {
yading@11:     int y;
yading@11:     const int chromWidth = -((-width) >> 1);
yading@11: 
yading@11:     for (y = 0; y < height; y++) {
yading@11:         extract_even_c(src + 1, ydst, width);
yading@11:         extract_even2_c(src, udst, vdst, chromWidth);
yading@11: 
yading@11:         src  += srcStride;
yading@11:         ydst += lumStride;
yading@11:         udst += chromStride;
yading@11:         vdst += chromStride;
yading@11:     }
yading@11: }
yading@11: 
yading@11: static inline void rgb2rgb_init_c(void)
yading@11: {
yading@11:     rgb15to16          = rgb15to16_c;
yading@11:     rgb15tobgr24       = rgb15tobgr24_c;
yading@11:     rgb15to32          = rgb15to32_c;
yading@11:     rgb16tobgr24       = rgb16tobgr24_c;
yading@11:     rgb16to32          = rgb16to32_c;
yading@11:     rgb16to15          = rgb16to15_c;
yading@11:     rgb24tobgr16       = rgb24tobgr16_c;
yading@11:     rgb24tobgr15       = rgb24tobgr15_c;
yading@11:     rgb24tobgr32       = rgb24tobgr32_c;
yading@11:     rgb32to16          = rgb32to16_c;
yading@11:     rgb32to15          = rgb32to15_c;
yading@11:     rgb32tobgr24       = rgb32tobgr24_c;
yading@11:     rgb24to15          = rgb24to15_c;
yading@11:     rgb24to16          = rgb24to16_c;
yading@11:     rgb24tobgr24       = rgb24tobgr24_c;
yading@11:     shuffle_bytes_2103 = shuffle_bytes_2103_c;
yading@11:     rgb32tobgr16       = rgb32tobgr16_c;
yading@11:     rgb32tobgr15       = rgb32tobgr15_c;
yading@11:     yv12toyuy2         = yv12toyuy2_c;
yading@11:     yv12touyvy         = yv12touyvy_c;
yading@11:     yuv422ptoyuy2      = yuv422ptoyuy2_c;
yading@11:     yuv422ptouyvy      = yuv422ptouyvy_c;
yading@11:     yuy2toyv12         = yuy2toyv12_c;
yading@11:     planar2x           = planar2x_c;
yading@11:     ff_rgb24toyv12     = ff_rgb24toyv12_c;
yading@11:     interleaveBytes    = interleaveBytes_c;
yading@11:     vu9_to_vu12        = vu9_to_vu12_c;
yading@11:     yvu9_to_yuy2       = yvu9_to_yuy2_c;
yading@11: 
yading@11:     uyvytoyuv420       = uyvytoyuv420_c;
yading@11:     uyvytoyuv422       = uyvytoyuv422_c;
yading@11:     yuyvtoyuv420       = yuyvtoyuv420_c;
yading@11:     yuyvtoyuv422       = yuyvtoyuv422_c;
yading@11: }