/*
- * yuv2rgb.c, Software YUV to RGB converter
+ * software YUV to RGB converter
*
- * Copyright (C) 1999, Aaron Holtzman <aholtzma@ess.engr.uvic.ca>
+ * Copyright (C) 2009 Konstantin Shishkov
*
- * Functions broken out from display_x11.c and several new modes
- * added by HÃ¥kan Hjort <d95hjort@dtek.chalmers.se>
+ * 1,4,8bpp support and context / deglobalize stuff
+ * by Michael Niedermayer (michaelni@gmx.at)
*
- * 15 & 16 bpp support by Franck Sicard <Franck.Sicard@solsoft.fr>
+ * This file is part of Libav.
*
- * MMX/MMX2 template stuff (needed for fast movntq support),
- * 1,4,8bpp support and context / deglobalize stuff
- * by Michael Niedermayer (michaelni@gmx.at)
+ * Libav 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.
*
- * This file is part of mpeg2dec, a free MPEG-2 video decoder
+ * Libav 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.
*
- * mpeg2dec is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2, or (at your option)
- * any later version.
- *
- * mpeg2dec 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 General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with mpeg2dec; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with Libav; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include <stdio.h>
#include <inttypes.h>
#include <assert.h>
+#include "libavutil/cpu.h"
+#include "libavutil/bswap.h"
#include "config.h"
#include "rgb2rgb.h"
#include "swscale.h"
#include "swscale_internal.h"
-#define DITHER1XBPP // only for MMX
-
-const uint8_t __attribute__((aligned(8))) dither_2x2_4[2][8]={
-{ 1, 3, 1, 3, 1, 3, 1, 3, },
-{ 2, 0, 2, 0, 2, 0, 2, 0, },
-};
-
-const uint8_t __attribute__((aligned(8))) dither_2x2_8[2][8]={
-{ 6, 2, 6, 2, 6, 2, 6, 2, },
-{ 0, 4, 0, 4, 0, 4, 0, 4, },
+extern const uint8_t dither_4x4_16[4][8];
+extern const uint8_t dither_8x8_32[8][8];
+extern const uint8_t dither_8x8_73[8][8];
+extern const uint8_t dither_8x8_220[8][8];
+
+const int32_t ff_yuv2rgb_coeffs[8][4] = {
+ { 117504, 138453, 13954, 34903 }, /* no sequence_display_extension */
+ { 117504, 138453, 13954, 34903 }, /* ITU-R Rec. 709 (1990) */
+ { 104597, 132201, 25675, 53279 }, /* unspecified */
+ { 104597, 132201, 25675, 53279 }, /* reserved */
+ { 104448, 132798, 24759, 53109 }, /* FCC */
+ { 104597, 132201, 25675, 53279 }, /* ITU-R Rec. 624-4 System B, G */
+ { 104597, 132201, 25675, 53279 }, /* SMPTE 170M */
+ { 117579, 136230, 16907, 35559 } /* SMPTE 240M (1987) */
};
-const uint8_t __attribute__((aligned(8))) dither_8x8_32[8][8]={
-{ 17, 9, 23, 15, 16, 8, 22, 14, },
-{ 5, 29, 3, 27, 4, 28, 2, 26, },
-{ 21, 13, 19, 11, 20, 12, 18, 10, },
-{ 0, 24, 6, 30, 1, 25, 7, 31, },
-{ 16, 8, 22, 14, 17, 9, 23, 15, },
-{ 4, 28, 2, 26, 5, 29, 3, 27, },
-{ 20, 12, 18, 10, 21, 13, 19, 11, },
-{ 1, 25, 7, 31, 0, 24, 6, 30, },
-};
-
-#if 0
-const uint8_t __attribute__((aligned(8))) dither_8x8_64[8][8]={
-{ 0, 48, 12, 60, 3, 51, 15, 63, },
-{ 32, 16, 44, 28, 35, 19, 47, 31, },
-{ 8, 56, 4, 52, 11, 59, 7, 55, },
-{ 40, 24, 36, 20, 43, 27, 39, 23, },
-{ 2, 50, 14, 62, 1, 49, 13, 61, },
-{ 34, 18, 46, 30, 33, 17, 45, 29, },
-{ 10, 58, 6, 54, 9, 57, 5, 53, },
-{ 42, 26, 38, 22, 41, 25, 37, 21, },
-};
-#endif
-
-const uint8_t __attribute__((aligned(8))) dither_8x8_73[8][8]={
-{ 0, 55, 14, 68, 3, 58, 17, 72, },
-{ 37, 18, 50, 32, 40, 22, 54, 35, },
-{ 9, 64, 5, 59, 13, 67, 8, 63, },
-{ 46, 27, 41, 23, 49, 31, 44, 26, },
-{ 2, 57, 16, 71, 1, 56, 15, 70, },
-{ 39, 21, 52, 34, 38, 19, 51, 33, },
-{ 11, 66, 7, 62, 10, 65, 6, 60, },
-{ 48, 30, 43, 25, 47, 29, 42, 24, },
-};
-
-#if 0
-const uint8_t __attribute__((aligned(8))) dither_8x8_128[8][8]={
-{ 68, 36, 92, 60, 66, 34, 90, 58, },
-{ 20, 116, 12, 108, 18, 114, 10, 106, },
-{ 84, 52, 76, 44, 82, 50, 74, 42, },
-{ 0, 96, 24, 120, 6, 102, 30, 126, },
-{ 64, 32, 88, 56, 70, 38, 94, 62, },
-{ 16, 112, 8, 104, 22, 118, 14, 110, },
-{ 80, 48, 72, 40, 86, 54, 78, 46, },
-{ 4, 100, 28, 124, 2, 98, 26, 122, },
-};
-#endif
-
-#if 1
-const uint8_t __attribute__((aligned(8))) dither_8x8_220[8][8]={
-{117, 62, 158, 103, 113, 58, 155, 100, },
-{ 34, 199, 21, 186, 31, 196, 17, 182, },
-{144, 89, 131, 76, 141, 86, 127, 72, },
-{ 0, 165, 41, 206, 10, 175, 52, 217, },
-{110, 55, 151, 96, 120, 65, 162, 107, },
-{ 28, 193, 14, 179, 38, 203, 24, 189, },
-{138, 83, 124, 69, 148, 93, 134, 79, },
-{ 7, 172, 48, 213, 3, 168, 45, 210, },
-};
-#elif 1
-// tries to correct a gamma of 1.5
-const uint8_t __attribute__((aligned(8))) dither_8x8_220[8][8]={
-{ 0, 143, 18, 200, 2, 156, 25, 215, },
-{ 78, 28, 125, 64, 89, 36, 138, 74, },
-{ 10, 180, 3, 161, 16, 195, 8, 175, },
-{109, 51, 93, 38, 121, 60, 105, 47, },
-{ 1, 152, 23, 210, 0, 147, 20, 205, },
-{ 85, 33, 134, 71, 81, 30, 130, 67, },
-{ 14, 190, 6, 171, 12, 185, 5, 166, },
-{117, 57, 101, 44, 113, 54, 97, 41, },
-};
-#elif 1
-// tries to correct a gamma of 2.0
-const uint8_t __attribute__((aligned(8))) dither_8x8_220[8][8]={
-{ 0, 124, 8, 193, 0, 140, 12, 213, },
-{ 55, 14, 104, 42, 66, 19, 119, 52, },
-{ 3, 168, 1, 145, 6, 187, 3, 162, },
-{ 86, 31, 70, 21, 99, 39, 82, 28, },
-{ 0, 134, 11, 206, 0, 129, 9, 200, },
-{ 62, 17, 114, 48, 58, 16, 109, 45, },
-{ 5, 181, 2, 157, 4, 175, 1, 151, },
-{ 95, 36, 78, 26, 90, 34, 74, 24, },
-};
-#else
-// tries to correct a gamma of 2.5
-const uint8_t __attribute__((aligned(8))) dither_8x8_220[8][8]={
-{ 0, 107, 3, 187, 0, 125, 6, 212, },
-{ 39, 7, 86, 28, 49, 11, 102, 36, },
-{ 1, 158, 0, 131, 3, 180, 1, 151, },
-{ 68, 19, 52, 12, 81, 25, 64, 17, },
-{ 0, 119, 5, 203, 0, 113, 4, 195, },
-{ 45, 9, 96, 33, 42, 8, 91, 30, },
-{ 2, 172, 1, 144, 2, 165, 0, 137, },
-{ 77, 23, 60, 15, 72, 21, 56, 14, },
-};
-#endif
-
-#ifdef HAVE_MMX
-
-/* hope these constant values are cache line aligned */
-DECLARE_ASM_CONST(8, uint64_t, mmx_00ffw) = 0x00ff00ff00ff00ffULL;
-DECLARE_ASM_CONST(8, uint64_t, mmx_redmask) = 0xf8f8f8f8f8f8f8f8ULL;
-DECLARE_ASM_CONST(8, uint64_t, mmx_grnmask) = 0xfcfcfcfcfcfcfcfcULL;
-
-// The volatile is required because gcc otherwise optimizes some writes away
-// not knowing that these are read in the ASM block.
-static volatile uint64_t attribute_used __attribute__((aligned(8))) b5Dither;
-static volatile uint64_t attribute_used __attribute__((aligned(8))) g5Dither;
-static volatile uint64_t attribute_used __attribute__((aligned(8))) g6Dither;
-static volatile uint64_t attribute_used __attribute__((aligned(8))) r5Dither;
-
-#undef HAVE_MMX
-
-//MMX versions
-#undef RENAME
-#define HAVE_MMX
-#undef HAVE_MMX2
-#undef HAVE_3DNOW
-#define RENAME(a) a ## _MMX
-#include "yuv2rgb_template.c"
-
-//MMX2 versions
-#undef RENAME
-#define HAVE_MMX
-#define HAVE_MMX2
-#undef HAVE_3DNOW
-#define RENAME(a) a ## _MMX2
-#include "yuv2rgb_template.c"
-
-#endif /* HAVE_MMX */
-
-const int32_t Inverse_Table_6_9[8][4] = {
- {117504, 138453, 13954, 34903}, /* no sequence_display_extension */
- {117504, 138453, 13954, 34903}, /* ITU-R Rec. 709 (1990) */
- {104597, 132201, 25675, 53279}, /* unspecified */
- {104597, 132201, 25675, 53279}, /* reserved */
- {104448, 132798, 24759, 53109}, /* FCC */
- {104597, 132201, 25675, 53279}, /* ITU-R Rec. 624-4 System B, G */
- {104597, 132201, 25675, 53279}, /* SMPTE 170M */
- {117579, 136230, 16907, 35559} /* SMPTE 240M (1987) */
-};
+const int *sws_getCoefficients(int colorspace)
+{
+ if (colorspace > 7 || colorspace < 0)
+ colorspace = SWS_CS_DEFAULT;
+ return ff_yuv2rgb_coeffs[colorspace];
+}
-#define RGB(i) \
+#define LOADCHROMA(i) \
U = pu[i]; \
V = pv[i]; \
r = (void *)c->table_rV[V]; \
g = (void *)(c->table_gU[U] + c->table_gV[V]); \
b = (void *)c->table_bU[U];
-#define DST1(i) \
- Y = py_1[2*i]; \
- dst_1[2*i] = r[Y] + g[Y] + b[Y]; \
- Y = py_1[2*i+1]; \
- dst_1[2*i+1] = r[Y] + g[Y] + b[Y];
-
-#define DST2(i) \
- Y = py_2[2*i]; \
- dst_2[2*i] = r[Y] + g[Y] + b[Y]; \
- Y = py_2[2*i+1]; \
- dst_2[2*i+1] = r[Y] + g[Y] + b[Y];
-
-#define DST1RGB(i) \
- Y = py_1[2*i]; \
- dst_1[6*i] = r[Y]; dst_1[6*i+1] = g[Y]; dst_1[6*i+2] = b[Y]; \
- Y = py_1[2*i+1]; \
- dst_1[6*i+3] = r[Y]; dst_1[6*i+4] = g[Y]; dst_1[6*i+5] = b[Y];
-
-#define DST2RGB(i) \
- Y = py_2[2*i]; \
- dst_2[6*i] = r[Y]; dst_2[6*i+1] = g[Y]; dst_2[6*i+2] = b[Y]; \
- Y = py_2[2*i+1]; \
- dst_2[6*i+3] = r[Y]; dst_2[6*i+4] = g[Y]; dst_2[6*i+5] = b[Y];
-
-#define DST1BGR(i) \
- Y = py_1[2*i]; \
- dst_1[6*i] = b[Y]; dst_1[6*i+1] = g[Y]; dst_1[6*i+2] = r[Y]; \
- Y = py_1[2*i+1]; \
- dst_1[6*i+3] = b[Y]; dst_1[6*i+4] = g[Y]; dst_1[6*i+5] = r[Y];
-
-#define DST2BGR(i) \
- Y = py_2[2*i]; \
- dst_2[6*i] = b[Y]; dst_2[6*i+1] = g[Y]; dst_2[6*i+2] = r[Y]; \
- Y = py_2[2*i+1]; \
- dst_2[6*i+3] = b[Y]; dst_2[6*i+4] = g[Y]; dst_2[6*i+5] = r[Y];
-
-#define PROLOG(func_name, dst_type) \
-static int func_name(SwsContext *c, uint8_t* src[], int srcStride[], int srcSliceY, \
- int srcSliceH, uint8_t* dst[], int dstStride[]){\
- int y;\
-\
- if (c->srcFormat == PIX_FMT_YUV422P){\
- srcStride[1] *= 2;\
- srcStride[2] *= 2;\
- }\
- for (y=0; y<srcSliceH; y+=2){\
- dst_type *dst_1= (dst_type*)(dst[0] + (y+srcSliceY )*dstStride[0]);\
- dst_type *dst_2= (dst_type*)(dst[0] + (y+srcSliceY+1)*dstStride[0]);\
- dst_type av_unused *r, *b;\
- dst_type *g;\
- uint8_t *py_1= src[0] + y*srcStride[0];\
- uint8_t *py_2= py_1 + srcStride[0];\
- uint8_t *pu= src[1] + (y>>1)*srcStride[1];\
- uint8_t *pv= src[2] + (y>>1)*srcStride[2];\
- unsigned int h_size= c->dstW>>3;\
- while (h_size--) {\
- int av_unused U, V;\
- int Y;\
-
-#define EPILOG1(dst_delta)\
- pu += 4;\
- pv += 4;\
- py_1 += 8;\
- py_2 += 8;\
- dst_1 += dst_delta;\
- dst_2 += dst_delta;\
- }\
- if (c->dstW & 4) {\
- int av_unused Y, U, V;\
-
-#define EPILOG2()\
- }\
- }\
- return srcSliceH;\
-}
+#define PUTRGB(dst, src, i) \
+ Y = src[2 * i]; \
+ dst[2 * i] = r[Y] + g[Y] + b[Y]; \
+ Y = src[2 * i + 1]; \
+ dst[2 * i + 1] = r[Y] + g[Y] + b[Y];
+
+#define PUTRGB24(dst, src, i) \
+ Y = src[2 * i]; \
+ dst[6 * i + 0] = r[Y]; \
+ dst[6 * i + 1] = g[Y]; \
+ dst[6 * i + 2] = b[Y]; \
+ Y = src[2 * i + 1]; \
+ dst[6 * i + 3] = r[Y]; \
+ dst[6 * i + 4] = g[Y]; \
+ dst[6 * i + 5] = b[Y];
+
+#define PUTBGR24(dst, src, i) \
+ Y = src[2 * i]; \
+ dst[6 * i + 0] = b[Y]; \
+ dst[6 * i + 1] = g[Y]; \
+ dst[6 * i + 2] = r[Y]; \
+ Y = src[2 * i + 1]; \
+ dst[6 * i + 3] = b[Y]; \
+ dst[6 * i + 4] = g[Y]; \
+ dst[6 * i + 5] = r[Y];
+
+#define PUTRGBA(dst, ysrc, asrc, i, s) \
+ Y = ysrc[2 * i]; \
+ dst[2 * i] = r[Y] + g[Y] + b[Y] + (asrc[2 * i] << s); \
+ Y = ysrc[2 * i + 1]; \
+ dst[2 * i + 1] = r[Y] + g[Y] + b[Y] + (asrc[2 * i + 1] << s);
+
+#define PUTRGB48(dst, src, i) \
+ Y = src[ 2 * i]; \
+ dst[12 * i + 0] = dst[12 * i + 1] = r[Y]; \
+ dst[12 * i + 2] = dst[12 * i + 3] = g[Y]; \
+ dst[12 * i + 4] = dst[12 * i + 5] = b[Y]; \
+ Y = src[ 2 * i + 1]; \
+ dst[12 * i + 6] = dst[12 * i + 7] = r[Y]; \
+ dst[12 * i + 8] = dst[12 * i + 9] = g[Y]; \
+ dst[12 * i + 10] = dst[12 * i + 11] = b[Y];
+
+#define PUTBGR48(dst, src, i) \
+ Y = src[2 * i]; \
+ dst[12 * i + 0] = dst[12 * i + 1] = b[Y]; \
+ dst[12 * i + 2] = dst[12 * i + 3] = g[Y]; \
+ dst[12 * i + 4] = dst[12 * i + 5] = r[Y]; \
+ Y = src[2 * i + 1]; \
+ dst[12 * i + 6] = dst[12 * i + 7] = b[Y]; \
+ dst[12 * i + 8] = dst[12 * i + 9] = g[Y]; \
+ dst[12 * i + 10] = dst[12 * i + 11] = r[Y];
+
+#define YUV2RGBFUNC(func_name, dst_type, alpha) \
+ static int func_name(SwsContext *c, const uint8_t *src[], \
+ int srcStride[], int srcSliceY, int srcSliceH, \
+ uint8_t *dst[], int dstStride[]) \
+ { \
+ int y; \
+ \
+ if (!alpha && c->srcFormat == PIX_FMT_YUV422P) { \
+ srcStride[1] *= 2; \
+ srcStride[2] *= 2; \
+ } \
+ for (y = 0; y < srcSliceH; y += 2) { \
+ dst_type *dst_1 = \
+ (dst_type *)(dst[0] + (y + srcSliceY) * dstStride[0]); \
+ dst_type *dst_2 = \
+ (dst_type *)(dst[0] + (y + srcSliceY + 1) * dstStride[0]); \
+ dst_type av_unused *r, *g, *b; \
+ const uint8_t *py_1 = src[0] + y * srcStride[0]; \
+ const uint8_t *py_2 = py_1 + srcStride[0]; \
+ const uint8_t *pu = src[1] + (y >> 1) * srcStride[1]; \
+ const uint8_t *pv = src[2] + (y >> 1) * srcStride[2]; \
+ const uint8_t av_unused *pa_1, *pa_2; \
+ unsigned int h_size = c->dstW >> 3; \
+ if (alpha) { \
+ pa_1 = src[3] + y * srcStride[3]; \
+ pa_2 = pa_1 + srcStride[3]; \
+ } \
+ while (h_size--) { \
+ int av_unused U, V, Y; \
+
+#define ENDYUV2RGBLINE(dst_delta) \
+ pu += 4; \
+ pv += 4; \
+ py_1 += 8; \
+ py_2 += 8; \
+ dst_1 += dst_delta; \
+ dst_2 += dst_delta; \
+ } \
+ if (c->dstW & 4) { \
+ int av_unused Y, U, V; \
+
+#define ENDYUV2RGBFUNC() \
+ } \
+ } \
+ return srcSliceH; \
+ }
-#define EPILOG(dst_delta)\
- EPILOG1(dst_delta)\
- EPILOG2()
-
-PROLOG(yuv2rgb_c_32, uint32_t)
- RGB(0);
- DST1(0);
- DST2(0);
-
- RGB(1);
- DST2(1);
- DST1(1);
-
- RGB(2);
- DST1(2);
- DST2(2);
-
- RGB(3);
- DST2(3);
- DST1(3);
-EPILOG1(8)
- RGB(0);
- DST1(0);
- DST2(0);
-
- RGB(1);
- DST2(1);
- DST1(1);
-EPILOG2()
-
-PROLOG(yuv2rgb_c_24_rgb, uint8_t)
- RGB(0);
- DST1RGB(0);
- DST2RGB(0);
-
- RGB(1);
- DST2RGB(1);
- DST1RGB(1);
-
- RGB(2);
- DST1RGB(2);
- DST2RGB(2);
-
- RGB(3);
- DST2RGB(3);
- DST1RGB(3);
-EPILOG1(24)
- RGB(0);
- DST1RGB(0);
- DST2RGB(0);
-
- RGB(1);
- DST2RGB(1);
- DST1RGB(1);
-EPILOG2()
+#define CLOSEYUV2RGBFUNC(dst_delta) \
+ ENDYUV2RGBLINE(dst_delta) \
+ ENDYUV2RGBFUNC()
+
+YUV2RGBFUNC(yuv2rgb_c_48, uint8_t, 0)
+ LOADCHROMA(0);
+ PUTRGB48(dst_1, py_1, 0);
+ PUTRGB48(dst_2, py_2, 0);
+
+ LOADCHROMA(1);
+ PUTRGB48(dst_2, py_2, 1);
+ PUTRGB48(dst_1, py_1, 1);
+
+ LOADCHROMA(2);
+ PUTRGB48(dst_1, py_1, 2);
+ PUTRGB48(dst_2, py_2, 2);
+
+ LOADCHROMA(3);
+ PUTRGB48(dst_2, py_2, 3);
+ PUTRGB48(dst_1, py_1, 3);
+ENDYUV2RGBLINE(48)
+ LOADCHROMA(0);
+ PUTRGB48(dst_1, py_1, 0);
+ PUTRGB48(dst_2, py_2, 0);
+
+ LOADCHROMA(1);
+ PUTRGB48(dst_2, py_2, 1);
+ PUTRGB48(dst_1, py_1, 1);
+ENDYUV2RGBFUNC()
+
+YUV2RGBFUNC(yuv2rgb_c_bgr48, uint8_t, 0)
+ LOADCHROMA(0);
+ PUTBGR48(dst_1, py_1, 0);
+ PUTBGR48(dst_2, py_2, 0);
+
+ LOADCHROMA(1);
+ PUTBGR48(dst_2, py_2, 1);
+ PUTBGR48(dst_1, py_1, 1);
+
+ LOADCHROMA(2);
+ PUTBGR48(dst_1, py_1, 2);
+ PUTBGR48(dst_2, py_2, 2);
+
+ LOADCHROMA(3);
+ PUTBGR48(dst_2, py_2, 3);
+ PUTBGR48(dst_1, py_1, 3);
+ENDYUV2RGBLINE(48)
+ LOADCHROMA(0);
+ PUTBGR48(dst_1, py_1, 0);
+ PUTBGR48(dst_2, py_2, 0);
+
+ LOADCHROMA(1);
+ PUTBGR48(dst_2, py_2, 1);
+ PUTBGR48(dst_1, py_1, 1);
+ENDYUV2RGBFUNC()
+
+YUV2RGBFUNC(yuv2rgb_c_32, uint32_t, 0)
+ LOADCHROMA(0);
+ PUTRGB(dst_1, py_1, 0);
+ PUTRGB(dst_2, py_2, 0);
+
+ LOADCHROMA(1);
+ PUTRGB(dst_2, py_2, 1);
+ PUTRGB(dst_1, py_1, 1);
+
+ LOADCHROMA(2);
+ PUTRGB(dst_1, py_1, 2);
+ PUTRGB(dst_2, py_2, 2);
+
+ LOADCHROMA(3);
+ PUTRGB(dst_2, py_2, 3);
+ PUTRGB(dst_1, py_1, 3);
+ENDYUV2RGBLINE(8)
+ LOADCHROMA(0);
+ PUTRGB(dst_1, py_1, 0);
+ PUTRGB(dst_2, py_2, 0);
+
+ LOADCHROMA(1);
+ PUTRGB(dst_2, py_2, 1);
+ PUTRGB(dst_1, py_1, 1);
+ENDYUV2RGBFUNC()
+
+YUV2RGBFUNC(yuva2rgba_c, uint32_t, 1)
+ LOADCHROMA(0);
+ PUTRGBA(dst_1, py_1, pa_1, 0, 24);
+ PUTRGBA(dst_2, py_2, pa_2, 0, 24);
+
+ LOADCHROMA(1);
+ PUTRGBA(dst_2, py_2, pa_1, 1, 24);
+ PUTRGBA(dst_1, py_1, pa_2, 1, 24);
+
+ LOADCHROMA(2);
+ PUTRGBA(dst_1, py_1, pa_1, 2, 24);
+ PUTRGBA(dst_2, py_2, pa_2, 2, 24);
+
+ LOADCHROMA(3);
+ PUTRGBA(dst_2, py_2, pa_1, 3, 24);
+ PUTRGBA(dst_1, py_1, pa_2, 3, 24);
+ pa_1 += 8; \
+ pa_2 += 8; \
+ENDYUV2RGBLINE(8)
+ LOADCHROMA(0);
+ PUTRGBA(dst_1, py_1, pa_1, 0, 24);
+ PUTRGBA(dst_2, py_2, pa_2, 0, 24);
+
+ LOADCHROMA(1);
+ PUTRGBA(dst_2, py_2, pa_1, 1, 24);
+ PUTRGBA(dst_1, py_1, pa_2, 1, 24);
+ENDYUV2RGBFUNC()
+
+YUV2RGBFUNC(yuva2argb_c, uint32_t, 1)
+ LOADCHROMA(0);
+ PUTRGBA(dst_1, py_1, pa_1, 0, 0);
+ PUTRGBA(dst_2, py_2, pa_2, 0, 0);
+
+ LOADCHROMA(1);
+ PUTRGBA(dst_2, py_2, pa_2, 1, 0);
+ PUTRGBA(dst_1, py_1, pa_1, 1, 0);
+
+ LOADCHROMA(2);
+ PUTRGBA(dst_1, py_1, pa_1, 2, 0);
+ PUTRGBA(dst_2, py_2, pa_2, 2, 0);
+
+ LOADCHROMA(3);
+ PUTRGBA(dst_2, py_2, pa_2, 3, 0);
+ PUTRGBA(dst_1, py_1, pa_1, 3, 0);
+ pa_1 += 8; \
+ pa_2 += 8; \
+ENDYUV2RGBLINE(8)
+ LOADCHROMA(0);
+ PUTRGBA(dst_1, py_1, pa_1, 0, 0);
+ PUTRGBA(dst_2, py_2, pa_2, 0, 0);
+
+ LOADCHROMA(1);
+ PUTRGBA(dst_2, py_2, pa_2, 1, 0);
+ PUTRGBA(dst_1, py_1, pa_1, 1, 0);
+ENDYUV2RGBFUNC()
+
+YUV2RGBFUNC(yuv2rgb_c_24_rgb, uint8_t, 0)
+ LOADCHROMA(0);
+ PUTRGB24(dst_1, py_1, 0);
+ PUTRGB24(dst_2, py_2, 0);
+
+ LOADCHROMA(1);
+ PUTRGB24(dst_2, py_2, 1);
+ PUTRGB24(dst_1, py_1, 1);
+
+ LOADCHROMA(2);
+ PUTRGB24(dst_1, py_1, 2);
+ PUTRGB24(dst_2, py_2, 2);
+
+ LOADCHROMA(3);
+ PUTRGB24(dst_2, py_2, 3);
+ PUTRGB24(dst_1, py_1, 3);
+ENDYUV2RGBLINE(24)
+ LOADCHROMA(0);
+ PUTRGB24(dst_1, py_1, 0);
+ PUTRGB24(dst_2, py_2, 0);
+
+ LOADCHROMA(1);
+ PUTRGB24(dst_2, py_2, 1);
+ PUTRGB24(dst_1, py_1, 1);
+ENDYUV2RGBFUNC()
// only trivial mods from yuv2rgb_c_24_rgb
-PROLOG(yuv2rgb_c_24_bgr, uint8_t)
- RGB(0);
- DST1BGR(0);
- DST2BGR(0);
-
- RGB(1);
- DST2BGR(1);
- DST1BGR(1);
-
- RGB(2);
- DST1BGR(2);
- DST2BGR(2);
-
- RGB(3);
- DST2BGR(3);
- DST1BGR(3);
-EPILOG1(24)
- RGB(0);
- DST1BGR(0);
- DST2BGR(0);
-
- RGB(1);
- DST2BGR(1);
- DST1BGR(1);
-EPILOG2()
+YUV2RGBFUNC(yuv2rgb_c_24_bgr, uint8_t, 0)
+ LOADCHROMA(0);
+ PUTBGR24(dst_1, py_1, 0);
+ PUTBGR24(dst_2, py_2, 0);
+
+ LOADCHROMA(1);
+ PUTBGR24(dst_2, py_2, 1);
+ PUTBGR24(dst_1, py_1, 1);
+
+ LOADCHROMA(2);
+ PUTBGR24(dst_1, py_1, 2);
+ PUTBGR24(dst_2, py_2, 2);
+
+ LOADCHROMA(3);
+ PUTBGR24(dst_2, py_2, 3);
+ PUTBGR24(dst_1, py_1, 3);
+ENDYUV2RGBLINE(24)
+ LOADCHROMA(0);
+ PUTBGR24(dst_1, py_1, 0);
+ PUTBGR24(dst_2, py_2, 0);
+
+ LOADCHROMA(1);
+ PUTBGR24(dst_2, py_2, 1);
+ PUTBGR24(dst_1, py_1, 1);
+ENDYUV2RGBFUNC()
// This is exactly the same code as yuv2rgb_c_32 except for the types of
// r, g, b, dst_1, dst_2
-PROLOG(yuv2rgb_c_16, uint16_t)
- RGB(0);
- DST1(0);
- DST2(0);
+YUV2RGBFUNC(yuv2rgb_c_16, uint16_t, 0)
+ LOADCHROMA(0);
+ PUTRGB(dst_1, py_1, 0);
+ PUTRGB(dst_2, py_2, 0);
- RGB(1);
- DST2(1);
- DST1(1);
+ LOADCHROMA(1);
+ PUTRGB(dst_2, py_2, 1);
+ PUTRGB(dst_1, py_1, 1);
- RGB(2);
- DST1(2);
- DST2(2);
+ LOADCHROMA(2);
+ PUTRGB(dst_1, py_1, 2);
+ PUTRGB(dst_2, py_2, 2);
- RGB(3);
- DST2(3);
- DST1(3);
-EPILOG(8)
+ LOADCHROMA(3);
+ PUTRGB(dst_2, py_2, 3);
+ PUTRGB(dst_1, py_1, 3);
+CLOSEYUV2RGBFUNC(8)
-// This is exactly the same code as yuv2rgb_c_32 except for the types of
// r, g, b, dst_1, dst_2
-PROLOG(yuv2rgb_c_8, uint8_t)
- RGB(0);
- DST1(0);
- DST2(0);
-
- RGB(1);
- DST2(1);
- DST1(1);
-
- RGB(2);
- DST1(2);
- DST2(2);
-
- RGB(3);
- DST2(3);
- DST1(3);
-EPILOG(8)
+YUV2RGBFUNC(yuv2rgb_c_12_ordered_dither, uint16_t, 0)
+ const uint8_t *d16 = dither_4x4_16[y & 3];
+
+#define PUTRGB12(dst, src, i, o) \
+ Y = src[2 * i]; \
+ dst[2 * i] = r[Y + d16[0 + o]] + \
+ g[Y + d16[0 + o]] + \
+ b[Y + d16[0 + o]]; \
+ Y = src[2 * i + 1]; \
+ dst[2 * i + 1] = r[Y + d16[1 + o]] + \
+ g[Y + d16[1 + o]] + \
+ b[Y + d16[1 + o]];
+
+ LOADCHROMA(0);
+ PUTRGB12(dst_1, py_1, 0, 0);
+ PUTRGB12(dst_2, py_2, 0, 0 + 8);
+
+ LOADCHROMA(1);
+ PUTRGB12(dst_2, py_2, 1, 2 + 8);
+ PUTRGB12(dst_1, py_1, 1, 2);
+
+ LOADCHROMA(2);
+ PUTRGB12(dst_1, py_1, 2, 4);
+ PUTRGB12(dst_2, py_2, 2, 4 + 8);
+
+ LOADCHROMA(3);
+ PUTRGB12(dst_2, py_2, 3, 6 + 8);
+ PUTRGB12(dst_1, py_1, 3, 6);
+CLOSEYUV2RGBFUNC(8)
// r, g, b, dst_1, dst_2
-PROLOG(yuv2rgb_c_8_ordered_dither, uint8_t)
- const uint8_t *d32= dither_8x8_32[y&7];
- const uint8_t *d64= dither_8x8_73[y&7];
-#define DST1bpp8(i,o) \
- Y = py_1[2*i]; \
- dst_1[2*i] = r[Y+d32[0+o]] + g[Y+d32[0+o]] + b[Y+d64[0+o]]; \
- Y = py_1[2*i+1]; \
- dst_1[2*i+1] = r[Y+d32[1+o]] + g[Y+d32[1+o]] + b[Y+d64[1+o]];
-
-#define DST2bpp8(i,o) \
- Y = py_2[2*i]; \
- dst_2[2*i] = r[Y+d32[8+o]] + g[Y+d32[8+o]] + b[Y+d64[8+o]]; \
- Y = py_2[2*i+1]; \
- dst_2[2*i+1] = r[Y+d32[9+o]] + g[Y+d32[9+o]] + b[Y+d64[9+o]];
-
-
- RGB(0);
- DST1bpp8(0,0);
- DST2bpp8(0,0);
-
- RGB(1);
- DST2bpp8(1,2);
- DST1bpp8(1,2);
-
- RGB(2);
- DST1bpp8(2,4);
- DST2bpp8(2,4);
-
- RGB(3);
- DST2bpp8(3,6);
- DST1bpp8(3,6);
-EPILOG(8)
-
-
-// This is exactly the same code as yuv2rgb_c_32 except for the types of
-// r, g, b, dst_1, dst_2
-PROLOG(yuv2rgb_c_4, uint8_t)
- int acc;
-#define DST1_4(i) \
- Y = py_1[2*i]; \
- acc = r[Y] + g[Y] + b[Y]; \
- Y = py_1[2*i+1]; \
- acc |= (r[Y] + g[Y] + b[Y])<<4; \
- dst_1[i] = acc;
-
-#define DST2_4(i) \
- Y = py_2[2*i]; \
- acc = r[Y] + g[Y] + b[Y]; \
- Y = py_2[2*i+1]; \
- acc |= (r[Y] + g[Y] + b[Y])<<4; \
- dst_2[i] = acc;
-
- RGB(0);
- DST1_4(0);
- DST2_4(0);
-
- RGB(1);
- DST2_4(1);
- DST1_4(1);
-
- RGB(2);
- DST1_4(2);
- DST2_4(2);
-
- RGB(3);
- DST2_4(3);
- DST1_4(3);
-EPILOG(4)
-
-PROLOG(yuv2rgb_c_4_ordered_dither, uint8_t)
- const uint8_t *d64= dither_8x8_73[y&7];
- const uint8_t *d128=dither_8x8_220[y&7];
+YUV2RGBFUNC(yuv2rgb_c_8_ordered_dither, uint8_t, 0)
+ const uint8_t *d32 = dither_8x8_32[y & 7];
+ const uint8_t *d64 = dither_8x8_73[y & 7];
+
+#define PUTRGB8(dst, src, i, o) \
+ Y = src[2 * i]; \
+ dst[2 * i] = r[Y + d32[0 + o]] + \
+ g[Y + d32[0 + o]] + \
+ b[Y + d64[0 + o]]; \
+ Y = src[2 * i + 1]; \
+ dst[2 * i + 1] = r[Y + d32[1 + o]] + \
+ g[Y + d32[1 + o]] + \
+ b[Y + d64[1 + o]];
+
+ LOADCHROMA(0);
+ PUTRGB8(dst_1, py_1, 0, 0);
+ PUTRGB8(dst_2, py_2, 0, 0 + 8);
+
+ LOADCHROMA(1);
+ PUTRGB8(dst_2, py_2, 1, 2 + 8);
+ PUTRGB8(dst_1, py_1, 1, 2);
+
+ LOADCHROMA(2);
+ PUTRGB8(dst_1, py_1, 2, 4);
+ PUTRGB8(dst_2, py_2, 2, 4 + 8);
+
+ LOADCHROMA(3);
+ PUTRGB8(dst_2, py_2, 3, 6 + 8);
+ PUTRGB8(dst_1, py_1, 3, 6);
+CLOSEYUV2RGBFUNC(8)
+
+YUV2RGBFUNC(yuv2rgb_c_4_ordered_dither, uint8_t, 0)
+ const uint8_t * d64 = dither_8x8_73[y & 7];
+ const uint8_t *d128 = dither_8x8_220[y & 7];
int acc;
-#define DST1bpp4(i,o) \
- Y = py_1[2*i]; \
- acc = r[Y+d128[0+o]] + g[Y+d64[0+o]] + b[Y+d128[0+o]]; \
- Y = py_1[2*i+1]; \
- acc |= (r[Y+d128[1+o]] + g[Y+d64[1+o]] + b[Y+d128[1+o]])<<4; \
- dst_1[i]= acc;
-
-#define DST2bpp4(i,o) \
- Y = py_2[2*i]; \
- acc = r[Y+d128[8+o]] + g[Y+d64[8+o]] + b[Y+d128[8+o]]; \
- Y = py_2[2*i+1]; \
- acc |= (r[Y+d128[9+o]] + g[Y+d64[9+o]] + b[Y+d128[9+o]])<<4; \
- dst_2[i]= acc;
-
-
- RGB(0);
- DST1bpp4(0,0);
- DST2bpp4(0,0);
-
- RGB(1);
- DST2bpp4(1,2);
- DST1bpp4(1,2);
-
- RGB(2);
- DST1bpp4(2,4);
- DST2bpp4(2,4);
-
- RGB(3);
- DST2bpp4(3,6);
- DST1bpp4(3,6);
-EPILOG(4)
-
-// This is exactly the same code as yuv2rgb_c_32 except for the types of
-// r, g, b, dst_1, dst_2
-PROLOG(yuv2rgb_c_4b, uint8_t)
- RGB(0);
- DST1(0);
- DST2(0);
-
- RGB(1);
- DST2(1);
- DST1(1);
-
- RGB(2);
- DST1(2);
- DST2(2);
-
- RGB(3);
- DST2(3);
- DST1(3);
-EPILOG(8)
-
-PROLOG(yuv2rgb_c_4b_ordered_dither, uint8_t)
- const uint8_t *d64= dither_8x8_73[y&7];
- const uint8_t *d128=dither_8x8_220[y&7];
-
-#define DST1bpp4b(i,o) \
- Y = py_1[2*i]; \
- dst_1[2*i] = r[Y+d128[0+o]] + g[Y+d64[0+o]] + b[Y+d128[0+o]]; \
- Y = py_1[2*i+1]; \
- dst_1[2*i+1] = r[Y+d128[1+o]] + g[Y+d64[1+o]] + b[Y+d128[1+o]];
-
-#define DST2bpp4b(i,o) \
- Y = py_2[2*i]; \
- dst_2[2*i] = r[Y+d128[8+o]] + g[Y+d64[8+o]] + b[Y+d128[8+o]]; \
- Y = py_2[2*i+1]; \
- dst_2[2*i+1] = r[Y+d128[9+o]] + g[Y+d64[9+o]] + b[Y+d128[9+o]];
-
-
- RGB(0);
- DST1bpp4b(0,0);
- DST2bpp4b(0,0);
-
- RGB(1);
- DST2bpp4b(1,2);
- DST1bpp4b(1,2);
-
- RGB(2);
- DST1bpp4b(2,4);
- DST2bpp4b(2,4);
-
- RGB(3);
- DST2bpp4b(3,6);
- DST1bpp4b(3,6);
-EPILOG(8)
-
-PROLOG(yuv2rgb_c_1_ordered_dither, uint8_t)
- const uint8_t *d128=dither_8x8_220[y&7];
- char out_1=0, out_2=0;
- g= c->table_gU[128] + c->table_gV[128];
-
-#define DST1bpp1(i,o) \
- Y = py_1[2*i]; \
- out_1+= out_1 + g[Y+d128[0+o]]; \
- Y = py_1[2*i+1]; \
- out_1+= out_1 + g[Y+d128[1+o]];
-
-#define DST2bpp1(i,o) \
- Y = py_2[2*i]; \
- out_2+= out_2 + g[Y+d128[8+o]]; \
- Y = py_2[2*i+1]; \
- out_2+= out_2 + g[Y+d128[9+o]];
-
- DST1bpp1(0,0);
- DST2bpp1(0,0);
-
- DST2bpp1(1,2);
- DST1bpp1(1,2);
-
- DST1bpp1(2,4);
- DST2bpp1(2,4);
-
- DST2bpp1(3,6);
- DST1bpp1(3,6);
-
- dst_1[0]= out_1;
- dst_2[0]= out_2;
-EPILOG(1)
-
-SwsFunc yuv2rgb_get_func_ptr (SwsContext *c)
+#define PUTRGB4D(dst, src, i, o) \
+ Y = src[2 * i]; \
+ acc = r[Y + d128[0 + o]] + \
+ g[Y + d64[0 + o]] + \
+ b[Y + d128[0 + o]]; \
+ Y = src[2 * i + 1]; \
+ acc |= (r[Y + d128[1 + o]] + \
+ g[Y + d64[1 + o]] + \
+ b[Y + d128[1 + o]]) << 4; \
+ dst[i] = acc;
+
+ LOADCHROMA(0);
+ PUTRGB4D(dst_1, py_1, 0, 0);
+ PUTRGB4D(dst_2, py_2, 0, 0 + 8);
+
+ LOADCHROMA(1);
+ PUTRGB4D(dst_2, py_2, 1, 2 + 8);
+ PUTRGB4D(dst_1, py_1, 1, 2);
+
+ LOADCHROMA(2);
+ PUTRGB4D(dst_1, py_1, 2, 4);
+ PUTRGB4D(dst_2, py_2, 2, 4 + 8);
+
+ LOADCHROMA(3);
+ PUTRGB4D(dst_2, py_2, 3, 6 + 8);
+ PUTRGB4D(dst_1, py_1, 3, 6);
+CLOSEYUV2RGBFUNC(4)
+
+YUV2RGBFUNC(yuv2rgb_c_4b_ordered_dither, uint8_t, 0)
+ const uint8_t *d64 = dither_8x8_73[y & 7];
+ const uint8_t *d128 = dither_8x8_220[y & 7];
+
+#define PUTRGB4DB(dst, src, i, o) \
+ Y = src[2 * i]; \
+ dst[2 * i] = r[Y + d128[0 + o]] + \
+ g[Y + d64[0 + o]] + \
+ b[Y + d128[0 + o]]; \
+ Y = src[2 * i + 1]; \
+ dst[2 * i + 1] = r[Y + d128[1 + o]] + \
+ g[Y + d64[1 + o]] + \
+ b[Y + d128[1 + o]];
+
+ LOADCHROMA(0);
+ PUTRGB4DB(dst_1, py_1, 0, 0);
+ PUTRGB4DB(dst_2, py_2, 0, 0 + 8);
+
+ LOADCHROMA(1);
+ PUTRGB4DB(dst_2, py_2, 1, 2 + 8);
+ PUTRGB4DB(dst_1, py_1, 1, 2);
+
+ LOADCHROMA(2);
+ PUTRGB4DB(dst_1, py_1, 2, 4);
+ PUTRGB4DB(dst_2, py_2, 2, 4 + 8);
+
+ LOADCHROMA(3);
+ PUTRGB4DB(dst_2, py_2, 3, 6 + 8);
+ PUTRGB4DB(dst_1, py_1, 3, 6);
+CLOSEYUV2RGBFUNC(8)
+
+YUV2RGBFUNC(yuv2rgb_c_1_ordered_dither, uint8_t, 0)
+ const uint8_t *d128 = dither_8x8_220[y & 7];
+ char out_1 = 0, out_2 = 0;
+ g = c->table_gU[128] + c->table_gV[128];
+
+#define PUTRGB1(out, src, i, o) \
+ Y = src[2 * i]; \
+ out += out + g[Y + d128[0 + o]]; \
+ Y = src[2 * i + 1]; \
+ out += out + g[Y + d128[1 + o]];
+
+ PUTRGB1(out_1, py_1, 0, 0);
+ PUTRGB1(out_2, py_2, 0, 0 + 8);
+
+ PUTRGB1(out_2, py_2, 1, 2 + 8);
+ PUTRGB1(out_1, py_1, 1, 2);
+
+ PUTRGB1(out_1, py_1, 2, 4);
+ PUTRGB1(out_2, py_2, 2, 4 + 8);
+
+ PUTRGB1(out_2, py_2, 3, 6 + 8);
+ PUTRGB1(out_1, py_1, 3, 6);
+
+ dst_1[0] = out_1;
+ dst_2[0] = out_2;
+CLOSEYUV2RGBFUNC(1)
+
+SwsFunc ff_yuv2rgb_get_func_ptr(SwsContext *c)
{
-#if defined(HAVE_MMX2) || defined(HAVE_MMX)
- if (c->flags & SWS_CPU_CAPS_MMX2){
- switch(c->dstFormat){
- case PIX_FMT_RGB32: return yuv420_rgb32_MMX2;
- case PIX_FMT_BGR24: return yuv420_rgb24_MMX2;
- case PIX_FMT_RGB565: return yuv420_rgb16_MMX2;
- case PIX_FMT_RGB555: return yuv420_rgb15_MMX2;
- }
- }
- if (c->flags & SWS_CPU_CAPS_MMX){
- switch(c->dstFormat){
- case PIX_FMT_RGB32: return yuv420_rgb32_MMX;
- case PIX_FMT_BGR24: return yuv420_rgb24_MMX;
- case PIX_FMT_RGB565: return yuv420_rgb16_MMX;
- case PIX_FMT_RGB555: return yuv420_rgb15_MMX;
- }
- }
-#endif
-#ifdef HAVE_VIS
- {
- SwsFunc t= yuv2rgb_init_vis(c);
- if (t) return t;
- }
-#endif
-#ifdef CONFIG_MLIB
- {
- SwsFunc t= yuv2rgb_init_mlib(c);
- if (t) return t;
- }
-#endif
-#ifdef HAVE_ALTIVEC
- if (c->flags & SWS_CPU_CAPS_ALTIVEC)
- {
- SwsFunc t = yuv2rgb_init_altivec(c);
- if (t) return t;
- }
-#endif
-
-#ifdef ARCH_BFIN
- if (c->flags & SWS_CPU_CAPS_BFIN)
- {
- SwsFunc t = ff_bfin_yuv2rgb_get_func_ptr (c);
- if (t) return t;
- }
-#endif
-
- av_log(c, AV_LOG_WARNING, "No accelerated colorspace conversion found.\n");
-
- switch(c->dstFormat){
- case PIX_FMT_BGR32_1:
- case PIX_FMT_RGB32_1:
- case PIX_FMT_BGR32:
- case PIX_FMT_RGB32: return yuv2rgb_c_32;
- case PIX_FMT_RGB24: return yuv2rgb_c_24_rgb;
- case PIX_FMT_BGR24: return yuv2rgb_c_24_bgr;
+ SwsFunc t = NULL;
+
+ if (HAVE_MMX)
+ t = ff_yuv2rgb_init_mmx(c);
+ else if (HAVE_VIS)
+ t = ff_yuv2rgb_init_vis(c);
+ else if (HAVE_ALTIVEC)
+ t = ff_yuv2rgb_init_altivec(c);
+ else if (ARCH_BFIN)
+ t = ff_yuv2rgb_get_func_ptr_bfin(c);
+
+ if (t)
+ return t;
+
+ av_log(c, AV_LOG_WARNING,
+ "No accelerated colorspace conversion found from %s to %s.\n",
+ sws_format_name(c->srcFormat), sws_format_name(c->dstFormat));
+
+ switch (c->dstFormat) {
+ case PIX_FMT_BGR48BE:
+ case PIX_FMT_BGR48LE:
+ return yuv2rgb_c_bgr48;
+ case PIX_FMT_RGB48BE:
+ case PIX_FMT_RGB48LE:
+ return yuv2rgb_c_48;
+ case PIX_FMT_ARGB:
+ case PIX_FMT_ABGR:
+ if (CONFIG_SWSCALE_ALPHA && c->srcFormat == PIX_FMT_YUVA420P)
+ return yuva2argb_c;
+ case PIX_FMT_RGBA:
+ case PIX_FMT_BGRA:
+ if (CONFIG_SWSCALE_ALPHA && c->srcFormat == PIX_FMT_YUVA420P)
+ return yuva2rgba_c;
+ else
+ return yuv2rgb_c_32;
+ case PIX_FMT_RGB24:
+ return yuv2rgb_c_24_rgb;
+ case PIX_FMT_BGR24:
+ return yuv2rgb_c_24_bgr;
case PIX_FMT_RGB565:
case PIX_FMT_BGR565:
case PIX_FMT_RGB555:
- case PIX_FMT_BGR555: return yuv2rgb_c_16;
+ case PIX_FMT_BGR555:
+ return yuv2rgb_c_16;
+ case PIX_FMT_RGB444:
+ case PIX_FMT_BGR444:
+ return yuv2rgb_c_12_ordered_dither;
case PIX_FMT_RGB8:
- case PIX_FMT_BGR8: return yuv2rgb_c_8_ordered_dither;
+ case PIX_FMT_BGR8:
+ return yuv2rgb_c_8_ordered_dither;
case PIX_FMT_RGB4:
- case PIX_FMT_BGR4: return yuv2rgb_c_4_ordered_dither;
+ case PIX_FMT_BGR4:
+ return yuv2rgb_c_4_ordered_dither;
case PIX_FMT_RGB4_BYTE:
- case PIX_FMT_BGR4_BYTE: return yuv2rgb_c_4b_ordered_dither;
- case PIX_FMT_MONOBLACK: return yuv2rgb_c_1_ordered_dither;
+ case PIX_FMT_BGR4_BYTE:
+ return yuv2rgb_c_4b_ordered_dither;
+ case PIX_FMT_MONOBLACK:
+ return yuv2rgb_c_1_ordered_dither;
default:
assert(0);
}
return NULL;
}
-static int div_round (int dividend, int divisor)
+static void fill_table(uint8_t *table[256], const int elemsize,
+ const int inc, void *y_tab)
+{
+ int i;
+ int64_t cb = 0;
+ uint8_t *y_table = y_tab;
+
+ y_table -= elemsize * (inc >> 9);
+
+ for (i = 0; i < 256; i++) {
+ table[i] = y_table + elemsize * (cb >> 16);
+ cb += inc;
+ }
+}
+
+static void fill_gv_table(int table[256], const int elemsize, const int inc)
+{
+ int i;
+ int64_t cb = 0;
+ int off = -(inc >> 9);
+
+ for (i = 0; i < 256; i++) {
+ table[i] = elemsize * (off + (cb >> 16));
+ cb += inc;
+ }
+}
+
+static uint16_t roundToInt16(int64_t f)
{
- if (dividend > 0)
- return (dividend + (divisor>>1)) / divisor;
+ int r = (f + (1 << 15)) >> 16;
+
+ if (r < -0x7FFF)
+ return 0x8000;
+ else if (r > 0x7FFF)
+ return 0x7FFF;
else
- return -((-dividend + (divisor>>1)) / divisor);
+ return r;
}
-int yuv2rgb_c_init_tables (SwsContext *c, const int inv_table[4], int fullRange, int brightness, int contrast, int saturation)
+av_cold int ff_yuv2rgb_c_init_tables(SwsContext *c, const int inv_table[4],
+ int fullRange, int brightness,
+ int contrast, int saturation)
{
- const int isRgb = c->dstFormat==PIX_FMT_RGB32
- || c->dstFormat==PIX_FMT_RGB32_1
- || c->dstFormat==PIX_FMT_BGR24
- || c->dstFormat==PIX_FMT_RGB565
- || c->dstFormat==PIX_FMT_RGB555
- || c->dstFormat==PIX_FMT_RGB8
- || c->dstFormat==PIX_FMT_RGB4
- || c->dstFormat==PIX_FMT_RGB4_BYTE
- || c->dstFormat==PIX_FMT_MONOBLACK;
- const int bpp = fmt_depth(c->dstFormat);
- int i, base;
- uint8_t table_Y[1024];
- uint32_t *table_32 = 0;
- uint16_t *table_16 = 0;
- uint8_t *table_8 = 0;
- uint8_t *table_332 = 0;
- uint8_t *table_121 = 0;
- uint8_t *table_1 = 0;
- int entry_size = 0;
- void *table_r = 0, *table_g = 0, *table_b = 0;
- void *table_start;
+ const int isRgb = c->dstFormat == PIX_FMT_RGB32 ||
+ c->dstFormat == PIX_FMT_RGB32_1 ||
+ c->dstFormat == PIX_FMT_BGR24 ||
+ c->dstFormat == PIX_FMT_RGB565BE ||
+ c->dstFormat == PIX_FMT_RGB565LE ||
+ c->dstFormat == PIX_FMT_RGB555BE ||
+ c->dstFormat == PIX_FMT_RGB555LE ||
+ c->dstFormat == PIX_FMT_RGB444BE ||
+ c->dstFormat == PIX_FMT_RGB444LE ||
+ c->dstFormat == PIX_FMT_RGB8 ||
+ c->dstFormat == PIX_FMT_RGB4 ||
+ c->dstFormat == PIX_FMT_RGB4_BYTE ||
+ c->dstFormat == PIX_FMT_MONOBLACK;
+ const int isNotNe = c->dstFormat == PIX_FMT_NE(RGB565LE, RGB565BE) ||
+ c->dstFormat == PIX_FMT_NE(RGB555LE, RGB555BE) ||
+ c->dstFormat == PIX_FMT_NE(RGB444LE, RGB444BE) ||
+ c->dstFormat == PIX_FMT_NE(BGR565LE, BGR565BE) ||
+ c->dstFormat == PIX_FMT_NE(BGR555LE, BGR555BE) ||
+ c->dstFormat == PIX_FMT_NE(BGR444LE, BGR444BE);
+ const int bpp = c->dstFormatBpp;
+ uint8_t *y_table;
+ uint16_t *y_table16;
+ uint32_t *y_table32;
+ int i, base, rbase, gbase, bbase, abase, needAlpha;
+ const int yoffs = fullRange ? 384 : 326;
int64_t crv = inv_table[0];
int64_t cbu = inv_table[1];
int64_t cgu = -inv_table[2];
int64_t cgv = -inv_table[3];
- int64_t cy = 1<<16;
+ int64_t cy = 1 << 16;
int64_t oy = 0;
-
-//printf("%lld %lld %lld %lld %lld\n", cy, crv, cbu, cgu, cgv);
- if (!fullRange){
- cy= (cy*255) / 219;
- oy= 16<<16;
- }else{
- crv= (crv*224) / 255;
- cbu= (cbu*224) / 255;
- cgu= (cgu*224) / 255;
- cgv= (cgv*224) / 255;
+ int64_t yb = 0;
+
+ if (!fullRange) {
+ cy = (cy * 255) / 219;
+ oy = 16 << 16;
+ } else {
+ crv = (crv * 224) / 255;
+ cbu = (cbu * 224) / 255;
+ cgu = (cgu * 224) / 255;
+ cgv = (cgv * 224) / 255;
}
- cy = (cy *contrast )>>16;
- crv= (crv*contrast * saturation)>>32;
- cbu= (cbu*contrast * saturation)>>32;
- cgu= (cgu*contrast * saturation)>>32;
- cgv= (cgv*contrast * saturation)>>32;
-//printf("%lld %lld %lld %lld %lld\n", cy, crv, cbu, cgu, cgv);
- oy -= 256*brightness;
+ cy = (cy * contrast) >> 16;
+ crv = (crv * contrast * saturation) >> 32;
+ cbu = (cbu * contrast * saturation) >> 32;
+ cgu = (cgu * contrast * saturation) >> 32;
+ cgv = (cgv * contrast * saturation) >> 32;
+ oy -= 256 * brightness;
+
+ c->uOffset = 0x0400040004000400LL;
+ c->vOffset = 0x0400040004000400LL;
+ c->yCoeff = roundToInt16(cy * 8192) * 0x0001000100010001ULL;
+ c->vrCoeff = roundToInt16(crv * 8192) * 0x0001000100010001ULL;
+ c->ubCoeff = roundToInt16(cbu * 8192) * 0x0001000100010001ULL;
+ c->vgCoeff = roundToInt16(cgv * 8192) * 0x0001000100010001ULL;
+ c->ugCoeff = roundToInt16(cgu * 8192) * 0x0001000100010001ULL;
+ c->yOffset = roundToInt16(oy * 8) * 0x0001000100010001ULL;
+
+ c->yuv2rgb_y_coeff = (int16_t)roundToInt16(cy << 13);
+ c->yuv2rgb_y_offset = (int16_t)roundToInt16(oy << 9);
+ c->yuv2rgb_v2r_coeff = (int16_t)roundToInt16(crv << 13);
+ c->yuv2rgb_v2g_coeff = (int16_t)roundToInt16(cgv << 13);
+ c->yuv2rgb_u2g_coeff = (int16_t)roundToInt16(cgu << 13);
+ c->yuv2rgb_u2b_coeff = (int16_t)roundToInt16(cbu << 13);
+
+ //scale coefficients by cy
+ crv = ((crv << 16) + 0x8000) / cy;
+ cbu = ((cbu << 16) + 0x8000) / cy;
+ cgu = ((cgu << 16) + 0x8000) / cy;
+ cgv = ((cgv << 16) + 0x8000) / cy;
- for (i = 0; i < 1024; i++) {
- int j;
-
- j= (cy*(((i - 384)<<16) - oy) + (1<<31))>>32;
- j = (j < 0) ? 0 : ((j > 255) ? 255 : j);
- table_Y[i] = j;
- }
+ av_free(c->yuvTable);
switch (bpp) {
- case 32:
- table_start= table_32 = av_malloc ((197 + 2*682 + 256 + 132) * sizeof (uint32_t));
- base= (c->dstFormat == PIX_FMT_RGB32_1 || c->dstFormat == PIX_FMT_BGR32_1) ? 8 : 0;
-
- entry_size = sizeof (uint32_t);
- table_r = table_32 + 197;
- table_b = table_32 + 197 + 685;
- table_g = table_32 + 197 + 2*682;
-
- for (i = -197; i < 256+197; i++)
- ((uint32_t *)table_r)[i] = table_Y[i+384] << ((isRgb ? 16 : 0) + base);
- for (i = -132; i < 256+132; i++)
- ((uint32_t *)table_g)[i] = table_Y[i+384] << (8 + base);
- for (i = -232; i < 256+232; i++)
- ((uint32_t *)table_b)[i] = table_Y[i+384] << ((isRgb ? 0 : 16) + base);
- break;
-
- case 24:
- table_start= table_8 = av_malloc ((256 + 2*232) * sizeof (uint8_t));
-
- entry_size = sizeof (uint8_t);
- table_r = table_g = table_b = table_8 + 232;
-
- for (i = -232; i < 256+232; i++)
- ((uint8_t * )table_b)[i] = table_Y[i+384];
- break;
-
- case 15:
- case 16:
- table_start= table_16 = av_malloc ((197 + 2*682 + 256 + 132) * sizeof (uint16_t));
-
- entry_size = sizeof (uint16_t);
- table_r = table_16 + 197;
- table_b = table_16 + 197 + 685;
- table_g = table_16 + 197 + 2*682;
-
- for (i = -197; i < 256+197; i++) {
- int j = table_Y[i+384] >> 3;
-
- if (isRgb)
- j <<= ((bpp==16) ? 11 : 10);
-
- ((uint16_t *)table_r)[i] = j;
- }
- for (i = -132; i < 256+132; i++) {
- int j = table_Y[i+384] >> ((bpp==16) ? 2 : 3);
-
- ((uint16_t *)table_g)[i] = j << 5;
+ case 1:
+ c->yuvTable = av_malloc(1024);
+ y_table = c->yuvTable;
+ yb = -(384 << 16) - oy;
+ for (i = 0; i < 1024 - 110; i++) {
+ y_table[i + 110] = av_clip_uint8((yb + 0x8000) >> 16) >> 7;
+ yb += cy;
}
- for (i = -232; i < 256+232; i++) {
- int j = table_Y[i+384] >> 3;
-
- if (!isRgb)
- j <<= ((bpp==16) ? 11 : 10);
-
- ((uint16_t *)table_b)[i] = j;
+ fill_table(c->table_gU, 1, cgu, y_table + yoffs);
+ fill_gv_table(c->table_gV, 1, cgv);
+ break;
+ case 4:
+ case 4 | 128:
+ rbase = isRgb ? 3 : 0;
+ gbase = 1;
+ bbase = isRgb ? 0 : 3;
+ c->yuvTable = av_malloc(1024 * 3);
+ y_table = c->yuvTable;
+ yb = -(384 << 16) - oy;
+ for (i = 0; i < 1024 - 110; i++) {
+ int yval = av_clip_uint8((yb + 0x8000) >> 16);
+ y_table[i + 110] = (yval >> 7) << rbase;
+ y_table[i + 37 + 1024] = ((yval + 43) / 85) << gbase;
+ y_table[i + 110 + 2048] = (yval >> 7) << bbase;
+ yb += cy;
}
+ fill_table(c->table_rV, 1, crv, y_table + yoffs);
+ fill_table(c->table_gU, 1, cgu, y_table + yoffs + 1024);
+ fill_table(c->table_bU, 1, cbu, y_table + yoffs + 2048);
+ fill_gv_table(c->table_gV, 1, cgv);
break;
-
case 8:
- table_start= table_332 = av_malloc ((197 + 2*682 + 256 + 132) * sizeof (uint8_t));
-
- entry_size = sizeof (uint8_t);
- table_r = table_332 + 197;
- table_b = table_332 + 197 + 685;
- table_g = table_332 + 197 + 2*682;
-
- for (i = -197; i < 256+197; i++) {
- int j = (table_Y[i+384 - 16] + 18)/36;
-
- if (isRgb)
- j <<= 5;
-
- ((uint8_t *)table_r)[i] = j;
- }
- for (i = -132; i < 256+132; i++) {
- int j = (table_Y[i+384 - 16] + 18)/36;
-
- if (!isRgb)
- j <<= 1;
-
- ((uint8_t *)table_g)[i] = j << 2;
- }
- for (i = -232; i < 256+232; i++) {
- int j = (table_Y[i+384 - 37] + 43)/85;
-
- if (!isRgb)
- j <<= 6;
-
- ((uint8_t *)table_b)[i] = j;
+ rbase = isRgb ? 5 : 0;
+ gbase = isRgb ? 2 : 3;
+ bbase = isRgb ? 0 : 6;
+ c->yuvTable = av_malloc(1024 * 3);
+ y_table = c->yuvTable;
+ yb = -(384 << 16) - oy;
+ for (i = 0; i < 1024 - 38; i++) {
+ int yval = av_clip_uint8((yb + 0x8000) >> 16);
+ y_table[i + 16] = ((yval + 18) / 36) << rbase;
+ y_table[i + 16 + 1024] = ((yval + 18) / 36) << gbase;
+ y_table[i + 37 + 2048] = ((yval + 43) / 85) << bbase;
+ yb += cy;
}
+ fill_table(c->table_rV, 1, crv, y_table + yoffs);
+ fill_table(c->table_gU, 1, cgu, y_table + yoffs + 1024);
+ fill_table(c->table_bU, 1, cbu, y_table + yoffs + 2048);
+ fill_gv_table(c->table_gV, 1, cgv);
break;
- case 4:
- case 4|128:
- table_start= table_121 = av_malloc ((197 + 2*682 + 256 + 132) * sizeof (uint8_t));
-
- entry_size = sizeof (uint8_t);
- table_r = table_121 + 197;
- table_b = table_121 + 197 + 685;
- table_g = table_121 + 197 + 2*682;
-
- for (i = -197; i < 256+197; i++) {
- int j = table_Y[i+384 - 110] >> 7;
-
- if (isRgb)
- j <<= 3;
-
- ((uint8_t *)table_r)[i] = j;
+ case 12:
+ rbase = isRgb ? 8 : 0;
+ gbase = 4;
+ bbase = isRgb ? 0 : 8;
+ c->yuvTable = av_malloc(1024 * 3 * 2);
+ y_table16 = c->yuvTable;
+ yb = -(384 << 16) - oy;
+ for (i = 0; i < 1024; i++) {
+ uint8_t yval = av_clip_uint8((yb + 0x8000) >> 16);
+ y_table16[i] = (yval >> 4) << rbase;
+ y_table16[i + 1024] = (yval >> 4) << gbase;
+ y_table16[i + 2048] = (yval >> 4) << bbase;
+ yb += cy;
}
- for (i = -132; i < 256+132; i++) {
- int j = (table_Y[i+384 - 37]+ 43)/85;
-
- ((uint8_t *)table_g)[i] = j << 1;
+ if (isNotNe)
+ for (i = 0; i < 1024 * 3; i++)
+ y_table16[i] = av_bswap16(y_table16[i]);
+ fill_table(c->table_rV, 2, crv, y_table16 + yoffs);
+ fill_table(c->table_gU, 2, cgu, y_table16 + yoffs + 1024);
+ fill_table(c->table_bU, 2, cbu, y_table16 + yoffs + 2048);
+ fill_gv_table(c->table_gV, 2, cgv);
+ break;
+ case 15:
+ case 16:
+ rbase = isRgb ? bpp - 5 : 0;
+ gbase = 5;
+ bbase = isRgb ? 0 : (bpp - 5);
+ c->yuvTable = av_malloc(1024 * 3 * 2);
+ y_table16 = c->yuvTable;
+ yb = -(384 << 16) - oy;
+ for (i = 0; i < 1024; i++) {
+ uint8_t yval = av_clip_uint8((yb + 0x8000) >> 16);
+ y_table16[i] = (yval >> 3) << rbase;
+ y_table16[i + 1024] = (yval >> (18 - bpp)) << gbase;
+ y_table16[i + 2048] = (yval >> 3) << bbase;
+ yb += cy;
}
- for (i = -232; i < 256+232; i++) {
- int j =table_Y[i+384 - 110] >> 7;
-
- if (!isRgb)
- j <<= 3;
-
- ((uint8_t *)table_b)[i] = j;
+ if (isNotNe)
+ for (i = 0; i < 1024 * 3; i++)
+ y_table16[i] = av_bswap16(y_table16[i]);
+ fill_table(c->table_rV, 2, crv, y_table16 + yoffs);
+ fill_table(c->table_gU, 2, cgu, y_table16 + yoffs + 1024);
+ fill_table(c->table_bU, 2, cbu, y_table16 + yoffs + 2048);
+ fill_gv_table(c->table_gV, 2, cgv);
+ break;
+ case 24:
+ case 48:
+ c->yuvTable = av_malloc(1024);
+ y_table = c->yuvTable;
+ yb = -(384 << 16) - oy;
+ for (i = 0; i < 1024; i++) {
+ y_table[i] = av_clip_uint8((yb + 0x8000) >> 16);
+ yb += cy;
}
+ fill_table(c->table_rV, 1, crv, y_table + yoffs);
+ fill_table(c->table_gU, 1, cgu, y_table + yoffs);
+ fill_table(c->table_bU, 1, cbu, y_table + yoffs);
+ fill_gv_table(c->table_gV, 1, cgv);
break;
-
- case 1:
- table_start= table_1 = av_malloc (256*2 * sizeof (uint8_t));
-
- entry_size = sizeof (uint8_t);
- table_g = table_1;
- table_r = table_b = NULL;
-
- for (i = 0; i < 256+256; i++) {
- int j = table_Y[i + 384 - 110]>>7;
-
- ((uint8_t *)table_g)[i] = j;
+ case 32:
+ base = (c->dstFormat == PIX_FMT_RGB32_1 ||
+ c->dstFormat == PIX_FMT_BGR32_1) ? 8 : 0;
+ rbase = base + (isRgb ? 16 : 0);
+ gbase = base + 8;
+ bbase = base + (isRgb ? 0 : 16);
+ needAlpha = CONFIG_SWSCALE_ALPHA && isALPHA(c->srcFormat);
+ if (!needAlpha)
+ abase = (base + 24) & 31;
+ c->yuvTable = av_malloc(1024 * 3 * 4);
+ y_table32 = c->yuvTable;
+ yb = -(384 << 16) - oy;
+ for (i = 0; i < 1024; i++) {
+ unsigned yval = av_clip_uint8((yb + 0x8000) >> 16);
+ y_table32[i] = (yval << rbase) +
+ (needAlpha ? 0 : (255u << abase));
+ y_table32[i + 1024] = yval << gbase;
+ y_table32[i + 2048] = yval << bbase;
+ yb += cy;
}
+ fill_table(c->table_rV, 4, crv, y_table32 + yoffs);
+ fill_table(c->table_gU, 4, cgu, y_table32 + yoffs + 1024);
+ fill_table(c->table_bU, 4, cbu, y_table32 + yoffs + 2048);
+ fill_gv_table(c->table_gV, 4, cgv);
break;
-
default:
- table_start= NULL;
+ c->yuvTable = NULL;
av_log(c, AV_LOG_ERROR, "%ibpp not supported by yuv2rgb\n", bpp);
- //free mem?
return -1;
}
-
- for (i = 0; i < 256; i++) {
- c->table_rV[i] = (uint8_t *)table_r + entry_size * div_round (crv * (i-128), 76309);
- c->table_gU[i] = (uint8_t *)table_g + entry_size * div_round (cgu * (i-128), 76309);
- c->table_gV[i] = entry_size * div_round (cgv * (i-128), 76309);
- c->table_bU[i] = (uint8_t *)table_b + entry_size * div_round (cbu * (i-128), 76309);
- }
-
- av_free(c->yuvTable);
- c->yuvTable= table_start;
return 0;
}