summaryrefslogtreecommitdiff
path: root/libswscale
diff options
context:
space:
mode:
Diffstat (limited to 'libswscale')
-rw-r--r--libswscale/Makefile4
-rw-r--r--libswscale/bfin/internal_bfin.S8
-rw-r--r--libswscale/bfin/swscale_bfin.c8
-rw-r--r--libswscale/bfin/yuv2rgb_bfin.c11
-rw-r--r--libswscale/colorspace-test.c8
-rw-r--r--libswscale/mlib/yuv2rgb_mlib.c8
-rw-r--r--libswscale/options.c8
-rw-r--r--libswscale/ppc/swscale_altivec.c8
-rw-r--r--libswscale/ppc/yuv2rgb_altivec.c23
-rw-r--r--libswscale/ppc/yuv2rgb_altivec.h8
-rw-r--r--libswscale/ppc/yuv2yuv_altivec.c8
-rw-r--r--libswscale/rgb2rgb.c44
-rw-r--r--libswscale/rgb2rgb.h8
-rw-r--r--libswscale/rgb2rgb_template.c68
-rw-r--r--libswscale/sparc/yuv2rgb_vis.c8
-rw-r--r--libswscale/swscale-test.c13
-rw-r--r--libswscale/swscale.c395
-rw-r--r--libswscale/swscale.h13
-rw-r--r--libswscale/swscale_internal.h78
-rw-r--r--libswscale/swscale_unscaled.c413
-rw-r--r--libswscale/utils.c108
-rw-r--r--libswscale/x86/rgb2rgb.c11
-rw-r--r--libswscale/x86/rgb2rgb_template.c106
-rw-r--r--libswscale/x86/scale.asm2
-rw-r--r--libswscale/x86/swscale_mmx.c94
-rw-r--r--libswscale/x86/swscale_template.c200
-rw-r--r--libswscale/x86/yuv2rgb_mmx.c8
-rw-r--r--libswscale/x86/yuv2rgb_template.c33
-rw-r--r--libswscale/yuv2rgb.c14
29 files changed, 1044 insertions, 674 deletions
diff --git a/libswscale/Makefile b/libswscale/Makefile
index 5671b2e1b3..942545d282 100644
--- a/libswscale/Makefile
+++ b/libswscale/Makefile
@@ -1,3 +1,5 @@
+include $(SUBDIR)../config.mak
+
NAME = swscale
FFLIBS = avutil
@@ -19,6 +21,8 @@ OBJS-$(HAVE_MMX) += x86/rgb2rgb.o \
OBJS-$(HAVE_VIS) += sparc/yuv2rgb_vis.o
OBJS-$(HAVE_YASM) += x86/scale.o
+$(SUBDIR)x86/swscale_mmx.o: CFLAGS += $(NOREDZONE_FLAGS)
+
TESTPROGS = colorspace swscale
DIRS = bfin mlib ppc sparc x86
diff --git a/libswscale/bfin/internal_bfin.S b/libswscale/bfin/internal_bfin.S
index 9f985e7824..cb8d71253c 100644
--- a/libswscale/bfin/internal_bfin.S
+++ b/libswscale/bfin/internal_bfin.S
@@ -5,20 +5,20 @@
* Blackfin video color space converter operations
* convert I420 YV12 to RGB in various formats
*
- * 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
*/
diff --git a/libswscale/bfin/swscale_bfin.c b/libswscale/bfin/swscale_bfin.c
index 0c5f004930..870636ea05 100644
--- a/libswscale/bfin/swscale_bfin.c
+++ b/libswscale/bfin/swscale_bfin.c
@@ -3,20 +3,20 @@
*
* Blackfin software video scaler operations
*
- * 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
*/
diff --git a/libswscale/bfin/yuv2rgb_bfin.c b/libswscale/bfin/yuv2rgb_bfin.c
index 68af522642..7a7dc7f0e6 100644
--- a/libswscale/bfin/yuv2rgb_bfin.c
+++ b/libswscale/bfin/yuv2rgb_bfin.c
@@ -4,20 +4,20 @@
* Blackfin video color space converter operations
* convert I420 YV12 to RGB in various formats
*
- * 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
*/
@@ -28,6 +28,7 @@
#include <assert.h>
#include "config.h"
#include <unistd.h>
+#include "libavutil/pixdesc.h"
#include "libswscale/rgb2rgb.h"
#include "libswscale/swscale.h"
#include "libswscale/swscale_internal.h"
@@ -197,7 +198,7 @@ SwsFunc ff_yuv2rgb_get_func_ptr_bfin(SwsContext *c)
}
av_log(c, AV_LOG_INFO, "BlackFin accelerated color space converter %s\n",
- sws_format_name (c->dstFormat));
+ av_get_pix_fmt_name(c->dstFormat));
return f;
}
diff --git a/libswscale/colorspace-test.c b/libswscale/colorspace-test.c
index 07c1cbd803..34095d8532 100644
--- a/libswscale/colorspace-test.c
+++ b/libswscale/colorspace-test.c
@@ -1,20 +1,20 @@
/*
* Copyright (C) 2002 Michael Niedermayer <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
*/
diff --git a/libswscale/mlib/yuv2rgb_mlib.c b/libswscale/mlib/yuv2rgb_mlib.c
index 526c97505c..e9f11494ee 100644
--- a/libswscale/mlib/yuv2rgb_mlib.c
+++ b/libswscale/mlib/yuv2rgb_mlib.c
@@ -3,20 +3,20 @@
*
* Copyright (C) 2003 Michael Niedermayer <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
*/
diff --git a/libswscale/options.c b/libswscale/options.c
index 9ba6e5e714..65095d5fd1 100644
--- a/libswscale/options.c
+++ b/libswscale/options.c
@@ -1,20 +1,20 @@
/*
* Copyright (C) 2001-2003 Michael Niedermayer <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
*/
diff --git a/libswscale/ppc/swscale_altivec.c b/libswscale/ppc/swscale_altivec.c
index 87059d9430..f2a8600925 100644
--- a/libswscale/ppc/swscale_altivec.c
+++ b/libswscale/ppc/swscale_altivec.c
@@ -4,20 +4,20 @@
* Copyright (C) 2004 Romain Dolbeau <romain@dolbeau.org>
* based on the equivalent C code in swscale.c
*
- * 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
*/
diff --git a/libswscale/ppc/yuv2rgb_altivec.c b/libswscale/ppc/yuv2rgb_altivec.c
index 73c02e9494..8e84c26382 100644
--- a/libswscale/ppc/yuv2rgb_altivec.c
+++ b/libswscale/ppc/yuv2rgb_altivec.c
@@ -3,20 +3,20 @@
*
* copyright (C) 2004 Marc Hoffman <marc.hoffman@analog.com>
*
- * 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
*/
@@ -95,6 +95,7 @@ adjustment.
#include "libswscale/swscale.h"
#include "libswscale/swscale_internal.h"
#include "libavutil/cpu.h"
+#include "libavutil/pixdesc.h"
#include "yuv2rgb_altivec.h"
#undef PROFILE_THE_BEAST
@@ -298,7 +299,7 @@ static int altivec_##name (SwsContext *c, \
vector signed short R1,G1,B1; \
vector unsigned char R,G,B; \
\
- vector unsigned char *y1ivP, *y2ivP, *uivP, *vivP; \
+ const vector unsigned char *y1ivP, *y2ivP, *uivP, *vivP; \
vector unsigned char align_perm; \
\
vector signed short \
@@ -335,10 +336,10 @@ static int altivec_##name (SwsContext *c, \
\
for (j=0;j<w/16;j++) { \
\
- y1ivP = (vector unsigned char *)y1i; \
- y2ivP = (vector unsigned char *)y2i; \
- uivP = (vector unsigned char *)ui; \
- vivP = (vector unsigned char *)vi; \
+ y1ivP = (const vector unsigned char *)y1i; \
+ y2ivP = (const vector unsigned char *)y2i; \
+ uivP = (const vector unsigned char *)ui; \
+ vivP = (const vector unsigned char *)vi; \
\
align_perm = vec_lvsl (0, y1i); \
y0 = (vector unsigned char) \
@@ -720,7 +721,7 @@ ff_yuv2packedX_altivec(SwsContext *c, const int16_t *lumFilter,
static int printed_error_message;
if (!printed_error_message) {
av_log(c, AV_LOG_ERROR, "altivec_yuv2packedX doesn't support %s output\n",
- sws_format_name(c->dstFormat));
+ av_get_pix_fmt_name(c->dstFormat));
printed_error_message=1;
}
return;
@@ -795,7 +796,7 @@ ff_yuv2packedX_altivec(SwsContext *c, const int16_t *lumFilter,
default:
/* Unreachable, I think. */
av_log(c, AV_LOG_ERROR, "altivec_yuv2packedX doesn't support %s output\n",
- sws_format_name(c->dstFormat));
+ av_get_pix_fmt_name(c->dstFormat));
return;
}
diff --git a/libswscale/ppc/yuv2rgb_altivec.h b/libswscale/ppc/yuv2rgb_altivec.h
index b809fe13fe..163eba6eb7 100644
--- a/libswscale/ppc/yuv2rgb_altivec.h
+++ b/libswscale/ppc/yuv2rgb_altivec.h
@@ -4,20 +4,20 @@
* Copyright (C) 2004 Romain Dolbeau <romain@dolbeau.org>
* based on the equivalent C code in swscale.c
*
- * 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
*/
diff --git a/libswscale/ppc/yuv2yuv_altivec.c b/libswscale/ppc/yuv2yuv_altivec.c
index 4cd02ffe1d..82c265afd2 100644
--- a/libswscale/ppc/yuv2yuv_altivec.c
+++ b/libswscale/ppc/yuv2yuv_altivec.c
@@ -4,20 +4,20 @@
* Copyright (C) 2004 Romain Dolbeau <romain@dolbeau.org>
* based on the equivalent C code in swscale.c
*
- * 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
*/
diff --git a/libswscale/rgb2rgb.c b/libswscale/rgb2rgb.c
index 51cb600e6b..adb3005753 100644
--- a/libswscale/rgb2rgb.c
+++ b/libswscale/rgb2rgb.c
@@ -6,20 +6,20 @@
* 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>
@@ -171,13 +171,13 @@ void rgb16tobgr32(const uint8_t *src, uint8_t *dst, int src_size)
bgr = *s++;
#if HAVE_BIGENDIAN
*d++ = 255;
- *d++ = (bgr&0x1F)<<3;
- *d++ = (bgr&0x7E0)>>3;
- *d++ = (bgr&0xF800)>>8;
+ *d++ = ((bgr&0x1F)<<3) | ((bgr&0x1F)>>2);
+ *d++ = ((bgr&0x7E0)>>3) | ((bgr&0x7E0)>>9);
+ *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
#else
- *d++ = (bgr&0xF800)>>8;
- *d++ = (bgr&0x7E0)>>3;
- *d++ = (bgr&0x1F)<<3;
+ *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
+ *d++ = ((bgr&0x7E0)>>3) | ((bgr&0x7E0)>>9);
+ *d++ = ((bgr&0x1F)<<3) | ((bgr&0x1F)>>2);
*d++ = 255;
#endif
}
@@ -192,9 +192,9 @@ void rgb16to24(const uint8_t *src, uint8_t *dst, int src_size)
while (s < end) {
register uint16_t bgr;
bgr = *s++;
- *d++ = (bgr&0xF800)>>8;
- *d++ = (bgr&0x7E0)>>3;
- *d++ = (bgr&0x1F)<<3;
+ *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
+ *d++ = ((bgr&0x7E0)>>3) | ((bgr&0x7E0)>>9);
+ *d++ = ((bgr&0x1F)<<3) | ((bgr&0x1F)>>2);
}
}
@@ -231,13 +231,13 @@ void rgb15tobgr32(const uint8_t *src, uint8_t *dst, int src_size)
bgr = *s++;
#if HAVE_BIGENDIAN
*d++ = 255;
- *d++ = (bgr&0x1F)<<3;
- *d++ = (bgr&0x3E0)>>2;
- *d++ = (bgr&0x7C00)>>7;
+ *d++ = ((bgr&0x1F)<<3) | ((bgr&0x1F)>>2);
+ *d++ = ((bgr&0x3E0)>>2) | ((bgr&0x3E0)>>7);
+ *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
#else
- *d++ = (bgr&0x7C00)>>7;
- *d++ = (bgr&0x3E0)>>2;
- *d++ = (bgr&0x1F)<<3;
+ *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
+ *d++ = ((bgr&0x3E0)>>2) | ((bgr&0x3E0)>>7);
+ *d++ = ((bgr&0x1F)<<3) | ((bgr&0x1F)>>2);
*d++ = 255;
#endif
}
@@ -252,9 +252,9 @@ void rgb15to24(const uint8_t *src, uint8_t *dst, int src_size)
while (s < end) {
register uint16_t bgr;
bgr = *s++;
- *d++ = (bgr&0x7C00)>>7;
- *d++ = (bgr&0x3E0)>>2;
- *d++ = (bgr&0x1F)<<3;
+ *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
+ *d++ = ((bgr&0x3E0)>>2) | ((bgr&0x3E0)>>7);
+ *d++ = ((bgr&0x1F)<<3) | ((bgr&0x1F)>>2);
}
}
diff --git a/libswscale/rgb2rgb.h b/libswscale/rgb2rgb.h
index 9d051de4f6..e3edac88d4 100644
--- a/libswscale/rgb2rgb.h
+++ b/libswscale/rgb2rgb.h
@@ -6,20 +6,20 @@
* Written by Nick Kurshev.
* 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
*/
diff --git a/libswscale/rgb2rgb_template.c b/libswscale/rgb2rgb_template.c
index c02015e5b3..6363bc19aa 100644
--- a/libswscale/rgb2rgb_template.c
+++ b/libswscale/rgb2rgb_template.c
@@ -7,20 +7,20 @@
* palette & YUV & runtime CPU stuff by Michael (michaelni@gmx.at)
* lot of big-endian byte order fixes by Alex Beregszaszi
*
- * 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
*/
@@ -227,27 +227,6 @@ static inline void rgb24to15_c(const uint8_t *src, uint8_t *dst, int src_size)
}
}
-/*
- I use less accurate approximation here by simply left-shifting the input
- value and filling the low order bits with zeroes. This method improves PNG
- compression but this scheme cannot reproduce white exactly, since it does
- not generate an all-ones maximum value; the net effect is to darken the
- image slightly.
-
- The better method should be "left bit replication":
-
- 4 3 2 1 0
- ---------
- 1 1 0 1 1
-
- 7 6 5 4 3 2 1 0
- ----------------
- 1 1 0 1 1 1 1 0
- |=======| |===|
- | leftmost bits repeated to fill open bits
- |
- original bits
-*/
static inline void rgb15tobgr24_c(const uint8_t *src, uint8_t *dst, int src_size)
{
const uint16_t *end;
@@ -257,9 +236,9 @@ static inline void rgb15tobgr24_c(const uint8_t *src, uint8_t *dst, int src_size
while (s < end) {
register uint16_t bgr;
bgr = *s++;
- *d++ = (bgr&0x1F)<<3;
- *d++ = (bgr&0x3E0)>>2;
- *d++ = (bgr&0x7C00)>>7;
+ *d++ = ((bgr&0x1F)<<3) | ((bgr&0x1F)>>2);
+ *d++ = ((bgr&0x3E0)>>2) | ((bgr&0x3E0)>>7);
+ *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
}
}
@@ -272,9 +251,9 @@ static inline void rgb16tobgr24_c(const uint8_t *src, uint8_t *dst, int src_size
while (s < end) {
register uint16_t bgr;
bgr = *s++;
- *d++ = (bgr&0x1F)<<3;
- *d++ = (bgr&0x7E0)>>3;
- *d++ = (bgr&0xF800)>>8;
+ *d++ = ((bgr&0x1F)<<3) | ((bgr&0x1F)>>2);
+ *d++ = ((bgr&0x7E0)>>3) | ((bgr&0x7E0)>>9);
+ *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
}
}
@@ -289,13 +268,13 @@ static inline void rgb15to32_c(const uint8_t *src, uint8_t *dst, int src_size)
bgr = *s++;
#if HAVE_BIGENDIAN
*d++ = 255;
- *d++ = (bgr&0x7C00)>>7;
- *d++ = (bgr&0x3E0)>>2;
- *d++ = (bgr&0x1F)<<3;
+ *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
+ *d++ = ((bgr&0x3E0)>>2) | ((bgr&0x3E0)>>7);
+ *d++ = ((bgr&0x1F)<<3) | ((bgr&0x1F)>>2);
#else
- *d++ = (bgr&0x1F)<<3;
- *d++ = (bgr&0x3E0)>>2;
- *d++ = (bgr&0x7C00)>>7;
+ *d++ = ((bgr&0x1F)<<3) | ((bgr&0x1F)>>2);
+ *d++ = ((bgr&0x3E0)>>2) | ((bgr&0x3E0)>>7);
+ *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
*d++ = 255;
#endif
}
@@ -312,13 +291,13 @@ static inline void rgb16to32_c(const uint8_t *src, uint8_t *dst, int src_size)
bgr = *s++;
#if HAVE_BIGENDIAN
*d++ = 255;
- *d++ = (bgr&0xF800)>>8;
- *d++ = (bgr&0x7E0)>>3;
- *d++ = (bgr&0x1F)<<3;
+ *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
+ *d++ = ((bgr&0x7E0)>>3) | ((bgr&0x7E0)>>9);
+ *d++ = ((bgr&0x1F)<<3) | ((bgr&0x1F)>>2);
#else
- *d++ = (bgr&0x1F)<<3;
- *d++ = (bgr&0x7E0)>>3;
- *d++ = (bgr&0xF800)>>8;
+ *d++ = ((bgr&0x1F)<<3) | ((bgr&0x1F)>>2);
+ *d++ = ((bgr&0x7E0)>>3) | ((bgr&0x7E0)>>9);
+ *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
*d++ = 255;
#endif
}
@@ -658,6 +637,9 @@ void rgb24toyv12_c(const uint8_t *src, uint8_t *ydst, uint8_t *udst,
ydst += lumStride;
src += srcStride;
+ if(y+1 == height)
+ break;
+
for (i=0; i<chromWidth; i++) {
unsigned int b = src[6*i+0];
unsigned int g = src[6*i+1];
diff --git a/libswscale/sparc/yuv2rgb_vis.c b/libswscale/sparc/yuv2rgb_vis.c
index 2111ea8f64..cc98f04053 100644
--- a/libswscale/sparc/yuv2rgb_vis.c
+++ b/libswscale/sparc/yuv2rgb_vis.c
@@ -2,20 +2,20 @@
* VIS optimized software YUV to RGB converter
* Copyright (c) 2007 Denes Balatoni <dbalatoni@programozo.hu>
*
- * 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
*/
diff --git a/libswscale/swscale-test.c b/libswscale/swscale-test.c
index b5cf1d202e..5dd2e34870 100644
--- a/libswscale/swscale-test.c
+++ b/libswscale/swscale-test.c
@@ -1,20 +1,20 @@
/*
- * Copyright (C) 2003 Michael Niedermayer <michaelni@gmx.at>
+ * Copyright (C) 2003-2011 Michael Niedermayer <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
*/
@@ -104,6 +104,7 @@ static int doTest(uint8_t *ref[4], int refStride[4], int w, int h,
av_image_fill_linesizes(srcStride, srcFormat, srcW);
for (p = 0; p < 4; p++) {
+ srcStride[p] = FFALIGN(srcStride[p], 16);
if (srcStride[p])
src[p] = av_mallocz(srcStride[p]*srcH+16);
if (srcStride[p] && !src[p]) {
@@ -139,6 +140,7 @@ static int doTest(uint8_t *ref[4], int refStride[4], int w, int h,
* allocated with av_malloc). */
/* An extra 16 bytes is being allocated because some scalers may write
* out of bounds. */
+ dstStride[i] = FFALIGN(dstStride[i], 16);
if (dstStride[i])
dst[i]= av_mallocz(dstStride[i]*dstH+16);
if (dstStride[i] && !dst[i]) {
@@ -178,6 +180,7 @@ static int doTest(uint8_t *ref[4], int refStride[4], int w, int h,
ssdA = r->ssdA;
} else {
for (i=0; i<4; i++) {
+ refStride[i] = FFALIGN(refStride[i], 16);
if (refStride[i])
out[i]= av_mallocz(refStride[i]*h);
if (refStride[i] && !out[i]) {
diff --git a/libswscale/swscale.c b/libswscale/swscale.c
index 975a0bdf1f..71033d3df8 100644
--- a/libswscale/swscale.c
+++ b/libswscale/swscale.c
@@ -1,20 +1,20 @@
/*
- * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>
+ * Copyright (C) 2001-2011 Michael Niedermayer <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
*/
@@ -60,6 +60,7 @@ untested special converters
#include "swscale.h"
#include "swscale_internal.h"
#include "rgb2rgb.h"
+#include "libavutil/avassert.h"
#include "libavutil/intreadwrite.h"
#include "libavutil/cpu.h"
#include "libavutil/avutil.h"
@@ -67,7 +68,6 @@ untested special converters
#include "libavutil/bswap.h"
#include "libavutil/pixdesc.h"
-#define DITHER1XBPP
#define RGB2YUV_SHIFT 15
#define BY ( (int)(0.114*219/255*(1<<RGB2YUV_SHIFT)+0.5))
@@ -195,6 +195,101 @@ DECLARE_ALIGNED(8, const uint8_t, dither_8x8_128)[8][8] = {
DECLARE_ALIGNED(8, const uint8_t, ff_sws_pb_64)[8] =
{ 64, 64, 64, 64, 64, 64, 64, 64 };
+DECLARE_ALIGNED(8, const uint8_t, dithers)[8][8][8]={
+{
+ { 0, 1, 0, 1, 0, 1, 0, 1,},
+ { 1, 0, 1, 0, 1, 0, 1, 0,},
+ { 0, 1, 0, 1, 0, 1, 0, 1,},
+ { 1, 0, 1, 0, 1, 0, 1, 0,},
+ { 0, 1, 0, 1, 0, 1, 0, 1,},
+ { 1, 0, 1, 0, 1, 0, 1, 0,},
+ { 0, 1, 0, 1, 0, 1, 0, 1,},
+ { 1, 0, 1, 0, 1, 0, 1, 0,},
+},{
+ { 1, 2, 1, 2, 1, 2, 1, 2,},
+ { 3, 0, 3, 0, 3, 0, 3, 0,},
+ { 1, 2, 1, 2, 1, 2, 1, 2,},
+ { 3, 0, 3, 0, 3, 0, 3, 0,},
+ { 1, 2, 1, 2, 1, 2, 1, 2,},
+ { 3, 0, 3, 0, 3, 0, 3, 0,},
+ { 1, 2, 1, 2, 1, 2, 1, 2,},
+ { 3, 0, 3, 0, 3, 0, 3, 0,},
+},{
+ { 2, 4, 3, 5, 2, 4, 3, 5,},
+ { 6, 0, 7, 1, 6, 0, 7, 1,},
+ { 3, 5, 2, 4, 3, 5, 2, 4,},
+ { 7, 1, 6, 0, 7, 1, 6, 0,},
+ { 2, 4, 3, 5, 2, 4, 3, 5,},
+ { 6, 0, 7, 1, 6, 0, 7, 1,},
+ { 3, 5, 2, 4, 3, 5, 2, 4,},
+ { 7, 1, 6, 0, 7, 1, 6, 0,},
+},{
+ { 4, 8, 7, 11, 4, 8, 7, 11,},
+ { 12, 0, 15, 3, 12, 0, 15, 3,},
+ { 6, 10, 5, 9, 6, 10, 5, 9,},
+ { 14, 2, 13, 1, 14, 2, 13, 1,},
+ { 4, 8, 7, 11, 4, 8, 7, 11,},
+ { 12, 0, 15, 3, 12, 0, 15, 3,},
+ { 6, 10, 5, 9, 6, 10, 5, 9,},
+ { 14, 2, 13, 1, 14, 2, 13, 1,},
+},{
+ { 9, 17, 15, 23, 8, 16, 14, 22,},
+ { 25, 1, 31, 7, 24, 0, 30, 6,},
+ { 13, 21, 11, 19, 12, 20, 10, 18,},
+ { 29, 5, 27, 3, 28, 4, 26, 2,},
+ { 8, 16, 14, 22, 9, 17, 15, 23,},
+ { 24, 0, 30, 6, 25, 1, 31, 7,},
+ { 12, 20, 10, 18, 13, 21, 11, 19,},
+ { 28, 4, 26, 2, 29, 5, 27, 3,},
+},{
+ { 18, 34, 30, 46, 17, 33, 29, 45,},
+ { 50, 2, 62, 14, 49, 1, 61, 13,},
+ { 26, 42, 22, 38, 25, 41, 21, 37,},
+ { 58, 10, 54, 6, 57, 9, 53, 5,},
+ { 16, 32, 28, 44, 19, 35, 31, 47,},
+ { 48, 0, 60, 12, 51, 3, 63, 15,},
+ { 24, 40, 20, 36, 27, 43, 23, 39,},
+ { 56, 8, 52, 4, 59, 11, 55, 7,},
+},{
+ { 18, 34, 30, 46, 17, 33, 29, 45,},
+ { 50, 2, 62, 14, 49, 1, 61, 13,},
+ { 26, 42, 22, 38, 25, 41, 21, 37,},
+ { 58, 10, 54, 6, 57, 9, 53, 5,},
+ { 16, 32, 28, 44, 19, 35, 31, 47,},
+ { 48, 0, 60, 12, 51, 3, 63, 15,},
+ { 24, 40, 20, 36, 27, 43, 23, 39,},
+ { 56, 8, 52, 4, 59, 11, 55, 7,},
+},{
+ { 36, 68, 60, 92, 34, 66, 58, 90,},
+ { 100, 4,124, 28, 98, 2,122, 26,},
+ { 52, 84, 44, 76, 50, 82, 42, 74,},
+ { 116, 20,108, 12,114, 18,106, 10,},
+ { 32, 64, 56, 88, 38, 70, 62, 94,},
+ { 96, 0,120, 24,102, 6,126, 30,},
+ { 48, 80, 40, 72, 54, 86, 46, 78,},
+ { 112, 16,104, 8,118, 22,110, 14,},
+}};
+
+static const uint8_t flat64[8]={64,64,64,64,64,64,64,64};
+
+const uint16_t dither_scale[15][16]={
+{ 2, 3, 3, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5,},
+{ 2, 3, 7, 7, 13, 13, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25,},
+{ 3, 3, 4, 15, 15, 29, 57, 57, 57, 113, 113, 113, 113, 113, 113, 113,},
+{ 3, 4, 4, 5, 31, 31, 61, 121, 241, 241, 241, 241, 481, 481, 481, 481,},
+{ 3, 4, 5, 5, 6, 63, 63, 125, 249, 497, 993, 993, 993, 993, 993, 1985,},
+{ 3, 5, 6, 6, 6, 7, 127, 127, 253, 505, 1009, 2017, 4033, 4033, 4033, 4033,},
+{ 3, 5, 6, 7, 7, 7, 8, 255, 255, 509, 1017, 2033, 4065, 8129,16257,16257,},
+{ 3, 5, 6, 8, 8, 8, 8, 9, 511, 511, 1021, 2041, 4081, 8161,16321,32641,},
+{ 3, 5, 7, 8, 9, 9, 9, 9, 10, 1023, 1023, 2045, 4089, 8177,16353,32705,},
+{ 3, 5, 7, 8, 10, 10, 10, 10, 10, 11, 2047, 2047, 4093, 8185,16369,32737,},
+{ 3, 5, 7, 8, 10, 11, 11, 11, 11, 11, 12, 4095, 4095, 8189,16377,32753,},
+{ 3, 5, 7, 9, 10, 12, 12, 12, 12, 12, 12, 13, 8191, 8191,16381,32761,},
+{ 3, 5, 7, 9, 10, 12, 13, 13, 13, 13, 13, 13, 14,16383,16383,32765,},
+{ 3, 5, 7, 9, 10, 12, 14, 14, 14, 14, 14, 14, 14, 15,32767,32767,},
+{ 3, 5, 7, 9, 11, 12, 14, 15, 15, 15, 15, 15, 15, 15, 16,65535,},
+};
+
#define output_pixel(pos, val, bias, signedness) \
if (big_endian) { \
AV_WB16(pos, bias + av_clip_ ## signedness ## 16(val >> shift)); \
@@ -207,7 +302,8 @@ yuv2plane1_16_c_template(const int32_t *src, uint16_t *dest, int dstW,
int big_endian, int output_bits)
{
int i;
- int shift = 19 - output_bits;
+ int shift = 3;
+ av_assert0(output_bits == 16);
for (i = 0; i < dstW; i++) {
int val = src[i] + (1 << (shift - 1));
@@ -221,10 +317,11 @@ yuv2planeX_16_c_template(const int16_t *filter, int filterSize,
int big_endian, int output_bits)
{
int i;
- int shift = 15 + 16 - output_bits;
+ int shift = 15;
+ av_assert0(output_bits == 16);
for (i = 0; i < dstW; i++) {
- int val = 1 << (30-output_bits);
+ int val = 1 << (shift - 1);
int j;
/* range of val is [0,0x7FFFFFFF], so 31 bits, but with lanczos/spline
@@ -271,7 +368,7 @@ yuv2planeX_10_c_template(const int16_t *filter, int filterSize,
int shift = 11 + 16 - output_bits;
for (i = 0; i < dstW; i++) {
- int val = 1 << (26-output_bits);
+ int val = 1 << (shift - 1);
int j;
for (j = 0; j < filterSize; j++)
@@ -433,8 +530,8 @@ yuv2gray16_1_c_template(SwsContext *c, const int32_t *buf0,
int i;
for (i = 0; i < (dstW >> 1); i++) {
- int Y1 = buf0[i * 2 ] << 1;
- int Y2 = buf0[i * 2 + 1] << 1;
+ int Y1 = (buf0[i * 2 ]+4)>>3;
+ int Y2 = (buf0[i * 2 + 1]+4)>>3;
output_pixel(&dest[i * 2 + 0], Y1);
output_pixel(&dest[i * 2 + 1], Y2);
@@ -882,8 +979,8 @@ yuv2rgb48_1_c_template(SwsContext *c, const int32_t *buf0,
for (i = 0; i < (dstW >> 1); i++) {
int Y1 = (buf0[i * 2] ) >> 2;
int Y2 = (buf0[i * 2 + 1]) >> 2;
- int U = (ubuf0[i] + ubuf1[i] + (-128 << 11)) >> 3;
- int V = (vbuf0[i] + vbuf1[i] + (-128 << 11)) >> 3;
+ int U = (ubuf0[i] + ubuf1[i] + (-128 << 12)) >> 3;
+ int V = (vbuf0[i] + vbuf1[i] + (-128 << 12)) >> 3;
int R, G, B;
Y1 -= c->yuv2rgb_y_offset;
@@ -954,6 +1051,7 @@ yuv2rgb_write(uint8_t *_dest, int i, int Y1, int Y2,
#define r_b ((target == PIX_FMT_RGB24) ? r : b)
#define b_r ((target == PIX_FMT_RGB24) ? b : r)
+
dest[i * 6 + 0] = r_b[Y1];
dest[i * 6 + 1] = g[Y1];
dest[i * 6 + 2] = b_r[Y1];
@@ -1244,9 +1342,9 @@ yuv2rgb_full_X_c_template(SwsContext *c, const int16_t *lumFilter,
for (i = 0; i < dstW; i++) {
int j;
- int Y = 0;
- int U = -128 << 19;
- int V = -128 << 19;
+ int Y = 1<<9;
+ int U = (1<<9)-(128 << 19);
+ int V = (1<<9)-(128 << 19);
int av_unused A;
int R, G, B;
@@ -1261,7 +1359,7 @@ yuv2rgb_full_X_c_template(SwsContext *c, const int16_t *lumFilter,
U >>= 10;
V >>= 10;
if (hasAlpha) {
- A = 1 << 21;
+ A = 1 << 18;
for (j = 0; j < lumFilterSize; j++) {
A += alpSrc[j][i] * lumFilter[j];
}
@@ -1304,7 +1402,6 @@ yuv2rgb_full_X_c_template(SwsContext *c, const int16_t *lumFilter,
dest[1] = B >> 22;
dest[2] = G >> 22;
dest[3] = R >> 22;
- dest += 4;
break;
case PIX_FMT_BGR24:
dest[0] = B >> 22;
@@ -1412,7 +1509,7 @@ rgb48ToUV_half_c_template(uint16_t *dstU, uint16_t *dstV,
#undef input_pixel
#define rgb48funcs(pattern, BE_LE, origin) \
-static void pattern ## 48 ## BE_LE ## ToY_c(uint8_t *_dst, const uint8_t *_src, \
+static void pattern ## 48 ## BE_LE ## ToY_c(uint8_t *_dst, const uint8_t *_src, const uint8_t *unused0, const uint8_t *unused1,\
int width, uint32_t *unused) \
{ \
const uint16_t *src = (const uint16_t *) _src; \
@@ -1421,7 +1518,7 @@ static void pattern ## 48 ## BE_LE ## ToY_c(uint8_t *_dst, const uint8_t *_src,
} \
\
static void pattern ## 48 ## BE_LE ## ToUV_c(uint8_t *_dstU, uint8_t *_dstV, \
- 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) \
{ \
const uint16_t *src1 = (const uint16_t *) _src1, \
@@ -1431,7 +1528,7 @@ static void pattern ## 48 ## BE_LE ## ToUV_c(uint8_t *_dstU, uint8_t *_dstV, \
} \
\
static void pattern ## 48 ## BE_LE ## ToUV_half_c(uint8_t *_dstU, uint8_t *_dstV, \
- 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) \
{ \
const uint16_t *src1 = (const uint16_t *) _src1, \
@@ -1450,14 +1547,14 @@ rgb48funcs(bgr, BE, PIX_FMT_BGR48BE);
(isBE(origin) ? AV_RB16(&src[(i)*2]) : AV_RL16(&src[(i)*2])))
static av_always_inline void
-rgb16_32ToY_c_template(uint8_t *dst, const uint8_t *src,
+rgb16_32ToY_c_template(int16_t *dst, const uint8_t *src,
int width, enum PixelFormat origin,
int shr, int shg, int shb, int shp,
int maskr, int maskg, int maskb,
int rsh, int gsh, int bsh, int S)
{
const int ry = RY << rsh, gy = GY << gsh, by = BY << bsh,
- rnd = 33 << (S - 1);
+ rnd = (32<<((S)-1)) + (1<<(S-7));
int i;
for (i = 0; i < width; i++) {
@@ -1466,12 +1563,12 @@ rgb16_32ToY_c_template(uint8_t *dst, const uint8_t *src,
int g = (px & maskg) >> shg;
int r = (px & maskr) >> shr;
- dst[i] = (ry * r + gy * g + by * b + rnd) >> S;
+ dst[i] = (ry * r + gy * g + by * b + rnd) >> ((S)-6);
}
}
static av_always_inline void
-rgb16_32ToUV_c_template(uint8_t *dstU, uint8_t *dstV,
+rgb16_32ToUV_c_template(int16_t *dstU, int16_t *dstV,
const uint8_t *src, int width,
enum PixelFormat origin,
int shr, int shg, int shb, int shp,
@@ -1480,7 +1577,7 @@ rgb16_32ToUV_c_template(uint8_t *dstU, uint8_t *dstV,
{
const int ru = RU << rsh, gu = GU << gsh, bu = BU << bsh,
rv = RV << rsh, gv = GV << gsh, bv = BV << bsh,
- rnd = 257 << (S - 1);
+ rnd = (256<<((S)-1)) + (1<<(S-7));
int i;
for (i = 0; i < width; i++) {
@@ -1489,13 +1586,13 @@ rgb16_32ToUV_c_template(uint8_t *dstU, uint8_t *dstV,
int g = (px & maskg) >> shg;
int r = (px & maskr) >> shr;
- dstU[i] = (ru * r + gu * g + bu * b + rnd) >> S;
- dstV[i] = (rv * r + gv * g + bv * b + rnd) >> S;
+ dstU[i] = (ru * r + gu * g + bu * b + rnd) >> ((S)-6);
+ dstV[i] = (rv * r + gv * g + bv * b + rnd) >> ((S)-6);
}
}
static av_always_inline void
-rgb16_32ToUV_half_c_template(uint8_t *dstU, uint8_t *dstV,
+rgb16_32ToUV_half_c_template(int16_t *dstU, int16_t *dstV,
const uint8_t *src, int width,
enum PixelFormat origin,
int shr, int shg, int shb, int shp,
@@ -1504,7 +1601,7 @@ rgb16_32ToUV_half_c_template(uint8_t *dstU, uint8_t *dstV,
{
const int ru = RU << rsh, gu = GU << gsh, bu = BU << bsh,
rv = RV << rsh, gv = GV << gsh, bv = BV << bsh,
- rnd = 257 << S, maskgx = ~(maskr | maskb);
+ rnd = (256U<<(S)) + (1<<(S-6)), maskgx = ~(maskr | maskb);
int i;
maskr |= maskr << 1; maskb |= maskb << 1; maskg |= maskg << 1;
@@ -1523,8 +1620,8 @@ rgb16_32ToUV_half_c_template(uint8_t *dstU, uint8_t *dstV,
}
r = (rb & maskr) >> shr;
- dstU[i] = (ru * r + gu * g + bu * b + rnd) >> (S + 1);
- dstV[i] = (rv * r + gv * g + bv * b + rnd) >> (S + 1);
+ dstU[i] = (ru * r + gu * g + bu * b + (unsigned)rnd) >> ((S)-6+1);
+ dstV[i] = (rv * r + gv * g + bv * b + (unsigned)rnd) >> ((S)-6+1);
}
}
@@ -1532,26 +1629,29 @@ rgb16_32ToUV_half_c_template(uint8_t *dstU, uint8_t *dstV,
#define rgb16_32_wrapper(fmt, name, shr, shg, shb, shp, maskr, \
maskg, maskb, rsh, gsh, bsh, S) \
-static void name ## ToY_c(uint8_t *dst, const uint8_t *src, \
+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) \
{ \
- rgb16_32ToY_c_template(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); \
} \
\
static void name ## ToUV_c(uint8_t *dstU, uint8_t *dstV, \
- 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) \
{ \
- rgb16_32ToUV_c_template(dstU, dstV, src, width, fmt, shr, shg, shb, shp, \
+ rgb16_32ToUV_c_template((int16_t*)dstU, (int16_t*)dstV, src, width, fmt, \
+ shr, shg, shb, shp, \
maskr, maskg, maskb, rsh, gsh, bsh, S); \
} \
\
static void name ## ToUV_half_c(uint8_t *dstU, uint8_t *dstV, \
- 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) \
{ \
- rgb16_32ToUV_half_c_template(dstU, dstV, src, width, fmt, shr, shg, shb, shp, \
+ rgb16_32ToUV_half_c_template((int16_t*)dstU, (int16_t*)dstV, src, width, fmt, \
+ shr, shg, shb, shp, \
maskr, maskg, maskb, rsh, gsh, bsh, S); \
}
@@ -1568,71 +1668,104 @@ rgb16_32_wrapper(PIX_FMT_BGR555BE, bgr15be, 0, 0, 0, 0, 0x001F, 0x03E0, 0x7
rgb16_32_wrapper(PIX_FMT_RGB565BE, rgb16be, 0, 0, 0, 0, 0xF800, 0x07E0, 0x001F, 0, 5, 11, RGB2YUV_SHIFT+8);
rgb16_32_wrapper(PIX_FMT_RGB555BE, rgb15be, 0, 0, 0, 0, 0x7C00, 0x03E0, 0x001F, 0, 5, 10, RGB2YUV_SHIFT+7);
-static void abgrToA_c(uint8_t *dst, const uint8_t *src, int width, uint32_t *unused)
+static void gbr24pToUV_half_c(uint16_t *dstU, uint16_t *dstV,
+ const uint8_t *gsrc, const uint8_t *bsrc, const uint8_t *rsrc,
+ int width, enum PixelFormat origin)
+{
+ int i;
+ for (i = 0; i < width; i++) {
+ unsigned int g = gsrc[2*i] + gsrc[2*i+1];
+ unsigned int b = bsrc[2*i] + bsrc[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);
+ dstV[i] = (RV*r + GV*g + BV*b + (0x4001<<(RGB2YUV_SHIFT-6))) >> (RGB2YUV_SHIFT-6+1);
+ }
+}
+
+static void abgrToA_c(int16_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, int width, uint32_t *unused)
+{
+ int i;
+ for (i=0; i<width; i++) {
+ dst[i]= src[4*i]<<6;
+ }
+}
+
+static void rgbaToA_c(int16_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, int width, uint32_t *unused)
{
int i;
for (i=0; i<width; i++) {
- dst[i]= src[4*i];
+ dst[i]= src[4*i+3]<<6;
}
}
-static void rgbaToA_c(uint8_t *dst, const uint8_t *src, int width, uint32_t *unused)
+static void palToA_c(int16_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, int width, uint32_t *pal)
{
int i;
for (i=0; i<width; i++) {
- dst[i]= src[4*i+3];
+ int d= src[i];
+
+ dst[i]= (pal[d] >> 24)<<6;
}
}
-static void palToY_c(uint8_t *dst, const uint8_t *src, int width, uint32_t *pal)
+static void palToY_c(int16_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, long width, uint32_t *pal)
{
int i;
for (i=0; i<width; i++) {
int d= src[i];
- dst[i]= pal[d] & 0xFF;
+ dst[i]= (pal[d] & 0xFF)<<6;
}
}
-static void palToUV_c(uint8_t *dstU, uint8_t *dstV,
- const uint8_t *src1, const uint8_t *src2,
- int width, uint32_t *pal)
+static void palToUV_c(uint16_t *dstU, int16_t *dstV,
+ const uint8_t *unused0, const uint8_t *src1, const uint8_t *src2,
+ int width, uint32_t *pal)
{
int i;
assert(src1 == src2);
for (i=0; i<width; i++) {
int p= pal[src1[i]];
- dstU[i]= p>>8;
- dstV[i]= p>>16;
+ dstU[i]= (uint8_t)(p>> 8)<<6;
+ dstV[i]= (uint8_t)(p>>16)<<6;
}
}
-static void monowhite2Y_c(uint8_t *dst, const uint8_t *src,
- int width, uint32_t *unused)
+static void monowhite2Y_c(int16_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, int width, uint32_t *unused)
{
int i, j;
for (i=0; i<width/8; i++) {
int d= ~src[i];
for(j=0; j<8; j++)
- dst[8*i+j]= ((d>>(7-j))&1)*255;
+ dst[8*i+j]= ((d>>(7-j))&1)*16383;
+ }
+ if(width&7){
+ int d= ~src[i];
+ for(j=0; j<(width&7); j++)
+ dst[8*i+j]= ((d>>(7-j))&1)*16383;
}
}
-static void monoblack2Y_c(uint8_t *dst, const uint8_t *src,
- int width, uint32_t *unused)
+static void monoblack2Y_c(int16_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, int width, uint32_t *unused)
{
int i, j;
for (i=0; i<width/8; i++) {
int d= src[i];
for(j=0; j<8; j++)
- dst[8*i+j]= ((d>>(7-j))&1)*255;
+ dst[8*i+j]= ((d>>(7-j))&1)*16383;
+ }
+ if(width&7){
+ int d= src[i];
+ for(j=0; j<(width&7); j++)
+ dst[8*i+j]= ((d>>(7-j))&1)*16383;
}
}
//FIXME yuy2* can read up to 7 samples too much
-static void yuy2ToY_c(uint8_t *dst, const uint8_t *src, int width,
+static void yuy2ToY_c(uint8_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, int width,
uint32_t *unused)
{
int i;
@@ -1640,7 +1773,7 @@ static void yuy2ToY_c(uint8_t *dst, const uint8_t *src, int width,
dst[i]= src[2*i];
}
-static void yuy2ToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
+static void yuy2ToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *unused0, const uint8_t *src1,
const uint8_t *src2, int width, uint32_t *unused)
{
int i;
@@ -1651,7 +1784,7 @@ static void yuy2ToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
assert(src1 == src2);
}
-static void bswap16Y_c(uint8_t *_dst, const uint8_t *_src, int width, uint32_t *unused)
+static void bswap16Y_c(uint8_t *_dst, const uint8_t *_src, const uint8_t *unused1, const uint8_t *unused2, int width, uint32_t *unused)
{
int i;
const uint16_t *src = (const uint16_t *) _src;
@@ -1661,7 +1794,7 @@ static void bswap16Y_c(uint8_t *_dst, const uint8_t *_src, int width, uint32_t *
}
}
-static void bswap16UV_c(uint8_t *_dstU, uint8_t *_dstV, const uint8_t *_src1,
+static void bswap16UV_c(uint8_t *_dstU, uint8_t *_dstV, const uint8_t *unused0, const uint8_t *_src1,
const uint8_t *_src2, int width, uint32_t *unused)
{
int i;
@@ -1676,7 +1809,7 @@ static void bswap16UV_c(uint8_t *_dstU, uint8_t *_dstV, const uint8_t *_src1,
/* This is almost identical to the previous, end exists only because
* yuy2ToY/UV)(dst, src+1, ...) would have 100% unaligned accesses. */
-static void uyvyToY_c(uint8_t *dst, const uint8_t *src, int width,
+static void uyvyToY_c(uint8_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, int width,
uint32_t *unused)
{
int i;
@@ -1684,7 +1817,7 @@ static void uyvyToY_c(uint8_t *dst, const uint8_t *src, int width,
dst[i]= src[2*i+1];
}
-static void uyvyToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
+static void uyvyToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *unused0, const uint8_t *src1,
const uint8_t *src2, int width, uint32_t *unused)
{
int i;
@@ -1706,14 +1839,14 @@ static av_always_inline void nvXXtoUV_c(uint8_t *dst1, uint8_t *dst2,
}
static void nv12ToUV_c(uint8_t *dstU, uint8_t *dstV,
- 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)
{
nvXXtoUV_c(dstU, dstV, src1, width);
}
static void nv21ToUV_c(uint8_t *dstU, uint8_t *dstV,
- 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)
{
nvXXtoUV_c(dstV, dstU, src1, width);
@@ -1721,7 +1854,7 @@ static void nv21ToUV_c(uint8_t *dstU, uint8_t *dstV,
#define input_pixel(pos) (isBE(origin) ? AV_RB16(pos) : AV_RL16(pos))
-static void bgr24ToY_c(uint8_t *dst, const uint8_t *src,
+static void bgr24ToY_c(int16_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2,
int width, uint32_t *unused)
{
int i;
@@ -1730,11 +1863,11 @@ static void bgr24ToY_c(uint8_t *dst, const uint8_t *src,
int g= src[i*3+1];
int r= src[i*3+2];
- dst[i]= ((RY*r + GY*g + BY*b + (33<<(RGB2YUV_SHIFT-1)))>>RGB2YUV_SHIFT);
+ 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 *src1,
+static void bgr24ToUV_c(int16_t *dstU, int16_t *dstV, const uint8_t *unused0, const uint8_t *src1,
const uint8_t *src2, int width, uint32_t *unused)
{
int i;
@@ -1743,13 +1876,13 @@ static void bgr24ToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
int g= src1[3*i + 1];
int r= src1[3*i + 2];
- dstU[i]= (RU*r + GU*g + BU*b + (257<<(RGB2YUV_SHIFT-1)))>>RGB2YUV_SHIFT;
- dstV[i]= (RV*r + GV*g + BV*b + (257<<(RGB2YUV_SHIFT-1)))>>RGB2YUV_SHIFT;
+ 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);
}
assert(src1 == src2);
}
-static void bgr24ToUV_half_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
+static void bgr24ToUV_half_c(int16_t *dstU, int16_t *dstV, const uint8_t *unused0, const uint8_t *src1,
const uint8_t *src2, int width, uint32_t *unused)
{
int i;
@@ -1758,13 +1891,13 @@ static void bgr24ToUV_half_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
int g= src1[6*i + 1] + src1[6*i + 4];
int r= src1[6*i + 2] + src1[6*i + 5];
- dstU[i]= (RU*r + GU*g + BU*b + (257<<RGB2YUV_SHIFT))>>(RGB2YUV_SHIFT+1);
- dstV[i]= (RV*r + GV*g + BV*b + (257<<RGB2YUV_SHIFT))>>(RGB2YUV_SHIFT+1);
+ 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);
}
assert(src1 == src2);
}
-static void rgb24ToY_c(uint8_t *dst, const uint8_t *src, int width,
+static void rgb24ToY_c(int16_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, int width,
uint32_t *unused)
{
int i;
@@ -1773,11 +1906,11 @@ static void rgb24ToY_c(uint8_t *dst, const uint8_t *src, int width,
int g= src[i*3+1];
int b= src[i*3+2];
- dst[i]= ((RY*r + GY*g + BY*b + (33<<(RGB2YUV_SHIFT-1)))>>RGB2YUV_SHIFT);
+ 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 *src1,
+static void rgb24ToUV_c(int16_t *dstU, int16_t *dstV, const uint8_t *unused0, const uint8_t *src1,
const uint8_t *src2, int width, uint32_t *unused)
{
int i;
@@ -1787,13 +1920,13 @@ static void rgb24ToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
int g= src1[3*i + 1];
int b= src1[3*i + 2];
- dstU[i]= (RU*r + GU*g + BU*b + (257<<(RGB2YUV_SHIFT-1)))>>RGB2YUV_SHIFT;
- dstV[i]= (RV*r + GV*g + BV*b + (257<<(RGB2YUV_SHIFT-1)))>>RGB2YUV_SHIFT;
+ 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);
}
}
-static void rgb24ToUV_half_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
- const uint8_t *src2, int width, uint32_t *unused)
+static void rgb24ToUV_half_c(int16_t *dstU, int16_t *dstV, const uint8_t *unused0, const uint8_t *src1,
+ const uint8_t *src2, int width, uint32_t *unused)
{
int i;
assert(src1==src2);
@@ -1802,12 +1935,12 @@ static void rgb24ToUV_half_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
int g= src1[6*i + 1] + src1[6*i + 4];
int b= src1[6*i + 2] + src1[6*i + 5];
- dstU[i]= (RU*r + GU*g + BU*b + (257<<RGB2YUV_SHIFT))>>(RGB2YUV_SHIFT+1);
- dstV[i]= (RV*r + GV*g + BV*b + (257<<RGB2YUV_SHIFT))>>(RGB2YUV_SHIFT+1);
+ 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);
}
}
-static void planar_rgb_to_y(uint8_t *dst, const uint8_t *src[4], int width)
+static void planar_rgb_to_y(uint16_t *dst, const uint8_t *src[4], int width)
{
int i;
for (i = 0; i < width; i++) {
@@ -1815,7 +1948,7 @@ static void planar_rgb_to_y(uint8_t *dst, const uint8_t *src[4], int width)
int b = src[1][i];
int r = src[2][i];
- dst[i] = ((RY * r + GY * g + BY * b + (33 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT);
+ dst[i] = (RY*r + GY*g + BY*b + (0x801<<(RGB2YUV_SHIFT-7))) >> (RGB2YUV_SHIFT-6);
}
}
@@ -1847,7 +1980,7 @@ static void planar_rgb16be_to_y(uint8_t *_dst, 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)
+static void planar_rgb_to_uv(uint16_t *dstU, uint16_t *dstV, const uint8_t *src[4], int width)
{
int i;
for (i = 0; i < width; i++) {
@@ -1855,8 +1988,8 @@ static void planar_rgb_to_uv(uint8_t *dstU, uint8_t *dstV, const uint8_t *src[4]
int b = src[1][i];
int r = src[2][i];
- dstU[i] = (RU * r + GU * g + BU * b + (257 << RGB2YUV_SHIFT)) >> (RGB2YUV_SHIFT + 1);
- dstV[i] = (RV * r + GV * g + BV * b + (257 << RGB2YUV_SHIFT)) >> (RGB2YUV_SHIFT + 1);
+ 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);
}
}
@@ -1902,6 +2035,9 @@ static void hScale16To19_c(SwsContext *c, int16_t *_dst, int dstW, const uint8_t
int bits = av_pix_fmt_descriptors[c->srcFormat].comp[0].depth_minus1;
int sh = bits - 4;
+ if((isAnyRGB(c->srcFormat) || c->srcFormat==PIX_FMT_PAL8) && av_pix_fmt_descriptors[c->srcFormat].comp[0].depth_minus1<15)
+ sh= 9;
+
for (i = 0; i < dstW; i++) {
int j;
int srcPos = filterPos[i];
@@ -1923,6 +2059,9 @@ static void hScale16To15_c(SwsContext *c, int16_t *dst, int dstW, const uint8_t
const uint16_t *src = (const uint16_t *) _src;
int sh = av_pix_fmt_descriptors[c->srcFormat].comp[0].depth_minus1;
+ if(sh<15)
+ sh= isAnyRGB(c->srcFormat) || c->srcFormat==PIX_FMT_PAL8 ? 13 : av_pix_fmt_descriptors[c->srcFormat].comp[0].depth_minus1;
+
for (i = 0; i < dstW; i++) {
int j;
int srcPos = filterPos[i];
@@ -2037,7 +2176,7 @@ static void lumRangeFromJpeg16_c(int16_t *_dst, int width)
int i;
int32_t *dst = (int32_t *) _dst;
for (i = 0; i < width; i++)
- dst[i] = (dst[i]*14071 + (33561947<<4))>>14;
+ dst[i] = (dst[i]*(14071/4) + (33561947<<4)/4)>>12;
}
static void hyscale_fast_c(SwsContext *c, int16_t *dst, int dstWidth,
@@ -2051,6 +2190,8 @@ static void hyscale_fast_c(SwsContext *c, int16_t *dst, int dstWidth,
dst[i]= (src[xx]<<7) + (src[xx+1] - src[xx])*xalpha;
xpos+=xInc;
}
+ for (i=dstWidth-1; (i*xInc)>>16 >=srcW-1; i--)
+ dst[i] = src[srcW-1]*128;
}
// *** horizontal scale Y line to temp buffer
@@ -2061,12 +2202,12 @@ static av_always_inline void hyscale(SwsContext *c, int16_t *dst, int dstWidth,
uint8_t *formatConvBuffer,
uint32_t *pal, int isAlpha)
{
- void (*toYV12)(uint8_t *, const uint8_t *, int, uint32_t *) = isAlpha ? c->alpToYV12 : c->lumToYV12;
+ void (*toYV12)(uint8_t *, const uint8_t *, const uint8_t *, const uint8_t *, int, uint32_t *) = isAlpha ? c->alpToYV12 : c->lumToYV12;
void (*convertRange)(int16_t *, int) = isAlpha ? NULL : c->lumConvertRange;
const uint8_t *src = src_in[isAlpha ? 3 : 0];
if (toYV12) {
- toYV12(formatConvBuffer, src, srcW, pal);
+ toYV12(formatConvBuffer, src, src_in[1], src_in[2], srcW, pal);
src= formatConvBuffer;
} else if (c->readLumPlanar && !isAlpha) {
c->readLumPlanar(formatConvBuffer, src_in, srcW);
@@ -2096,6 +2237,10 @@ static void hcscale_fast_c(SwsContext *c, int16_t *dst1, int16_t *dst2,
dst2[i]=(src2[xx]*(xalpha^127)+src2[xx+1]*xalpha);
xpos+=xInc;
}
+ for (i=dstWidth-1; (i*xInc)>>16 >=srcW-1; i--) {
+ dst1[i] = src1[srcW-1]*128;
+ dst2[i] = src2[srcW-1]*128;
+ }
}
static av_always_inline void hcscale(SwsContext *c, int16_t *dst1, int16_t *dst2, int dstWidth,
@@ -2106,12 +2251,12 @@ static av_always_inline void hcscale(SwsContext *c, int16_t *dst1, int16_t *dst2
{
const uint8_t *src1 = src_in[1], *src2 = src_in[2];
if (c->chrToYV12) {
- uint8_t *buf2 = formatConvBuffer + FFALIGN(srcW * FFALIGN(c->srcBpc, 8) >> 3, 16);
- c->chrToYV12(formatConvBuffer, buf2, src1, src2, srcW, pal);
+ uint8_t *buf2 = formatConvBuffer + FFALIGN(srcW*2+78, 16);
+ c->chrToYV12(formatConvBuffer, buf2, src_in[0], src1, src2, srcW, pal);
src1= formatConvBuffer;
src2= buf2;
} else if (c->readChrPlanar) {
- uint8_t *buf2 = formatConvBuffer + FFALIGN(srcW * FFALIGN(c->srcBpc, 8) >> 3, 16);
+ uint8_t *buf2 = formatConvBuffer + FFALIGN(srcW*2+78, 16);
c->readChrPlanar(formatConvBuffer, buf2, src_in, srcW);
src1= formatConvBuffer;
src2= buf2;
@@ -2220,7 +2365,10 @@ find_c_packed_planar_out_funcs(SwsContext *c,
*yuv2packedX = yuv2bgr24_full_X_c;
break;
}
+ if(!*yuv2packedX)
+ goto YUV_PACKED;
} else {
+ YUV_PACKED:
switch (dstFormat) {
case PIX_FMT_GRAY16BE:
*yuv2packed1 = yuv2gray16BE_1_c;
@@ -2391,8 +2539,6 @@ static int swScale(SwsContext *c, const uint8_t* src[],
int16_t *vChrFilterPos= c->vChrFilterPos;
int16_t *hLumFilterPos= c->hLumFilterPos;
int16_t *hChrFilterPos= c->hChrFilterPos;
- int16_t *vLumFilter= c->vLumFilter;
- int16_t *vChrFilter= c->vChrFilter;
int16_t *hLumFilter= c->hLumFilter;
int16_t *hChrFilter= c->hChrFilter;
int32_t *lumMmxFilter= c->lumMmxFilter;
@@ -2413,13 +2559,14 @@ static int swScale(SwsContext *c, const uint8_t* src[],
const int chrSrcSliceH= -((-srcSliceH) >> c->chrSrcVSubSample);
int lastDstY;
uint32_t *pal=c->pal_yuv;
+ int should_dither= isNBPS(c->srcFormat) || is16BPS(c->srcFormat);
+
yuv2planar1_fn yuv2plane1 = c->yuv2plane1;
yuv2planarX_fn yuv2planeX = c->yuv2planeX;
yuv2interleavedX_fn yuv2nv12cX = c->yuv2nv12cX;
yuv2packed1_fn yuv2packed1 = c->yuv2packed1;
yuv2packed2_fn yuv2packed2 = c->yuv2packed2;
yuv2packedX_fn yuv2packedX = c->yuv2packedX;
- int should_dither = is9_OR_10BPS(c->srcFormat) || is16BPS(c->srcFormat);
/* vars which will change and which we need to store back in the context */
int dstY= c->dstY;
@@ -2449,7 +2596,7 @@ static int swScale(SwsContext *c, const uint8_t* src[],
DEBUG_BUFFERS("vLumFilterSize: %d vLumBufSize: %d vChrFilterSize: %d vChrBufSize: %d\n",
vLumFilterSize, vLumBufSize, vChrFilterSize, vChrBufSize);
- if (dstStride[0]%8 !=0 || dstStride[1]%8 !=0 || dstStride[2]%8 !=0 || dstStride[3]%8 != 0) {
+ if (dstStride[0]%16 !=0 || dstStride[1]%16 !=0 || dstStride[2]%16 !=0 || dstStride[3]%16 != 0) {
static int warnedAlready=0; //FIXME move this into the context perhaps
if (flags & SWS_PRINT_INFO && !warnedAlready) {
av_log(c, AV_LOG_WARNING, "Warning: dstStride is not aligned!\n"
@@ -2458,6 +2605,18 @@ static int swScale(SwsContext *c, const uint8_t* src[],
}
}
+ if ((int)dst[0]%16 || (int)dst[1]%16 || (int)dst[2]%16 || (int)src[0]%16 || (int)src[1]%16 || (int)src[2]%16
+ || dstStride[0]%16 || dstStride[1]%16 || dstStride[2]%16 || dstStride[3]%16
+ || srcStride[0]%16 || srcStride[1]%16 || srcStride[2]%16 || srcStride[3]%16
+ ) {
+ static int warnedAlready=0;
+ int cpu_flags = av_get_cpu_flags();
+ if (HAVE_MMX2 && (cpu_flags & AV_CPU_FLAG_SSE2) && !warnedAlready){
+ av_log(c, AV_LOG_WARNING, "Warning: data is not aligned! This can lead to a speedloss\n");
+ warnedAlready=1;
+ }
+ }
+
/* Note the user might start scaling the picture in the middle so this
will not get executed. This is not really intended but works
currently, so people might do it. */
@@ -2482,6 +2641,7 @@ static int swScale(SwsContext *c, const uint8_t* src[],
dst[2] + dstStride[2] * chrDstY,
(CONFIG_SWSCALE_ALPHA && alpPixBuf) ? dst[3] + dstStride[3] * dstY : NULL,
};
+ int use_mmx_vfilter= c->use_mmx_vfilter;
const int firstLumSrcY= vLumFilterPos[dstY]; //First line needed as input
const int firstLumSrcY2= vLumFilterPos[FFMIN(dstY | ((1<<c->chrDstVSubSample) - 1), dstH-1)];
@@ -2577,6 +2737,7 @@ static int swScale(SwsContext *c, const uint8_t* src[],
// hmm looks like we can't use MMX here without overwriting this array's tail
find_c_packed_planar_out_funcs(c, &yuv2plane1, &yuv2planeX, &yuv2nv12cX,
&yuv2packed1, &yuv2packed2, &yuv2packedX);
+ use_mmx_vfilter= 0;
}
{
@@ -2584,35 +2745,58 @@ static int swScale(SwsContext *c, const uint8_t* src[],
const int16_t **chrUSrcPtr= (const int16_t **) chrUPixBuf + chrBufIndex + firstChrSrcY - lastInChrBuf + vChrBufSize;
const int16_t **chrVSrcPtr= (const int16_t **) chrVPixBuf + chrBufIndex + firstChrSrcY - lastInChrBuf + vChrBufSize;
const int16_t **alpSrcPtr= (CONFIG_SWSCALE_ALPHA && alpPixBuf) ? (const int16_t **) alpPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize : NULL;
+ int16_t *vLumFilter= c->vLumFilter;
+ int16_t *vChrFilter= c->vChrFilter;
+
if (isPlanarYUV(dstFormat) || dstFormat==PIX_FMT_GRAY8) { //YV12 like
const int chrSkipMask= (1<<c->chrDstVSubSample)-1;
+ vLumFilter += dstY * vLumFilterSize;
+ vChrFilter += chrDstY * vChrFilterSize;
+
+ av_assert0(use_mmx_vfilter != (
+ yuv2planeX == yuv2planeX_10BE_c
+ || yuv2planeX == yuv2planeX_10LE_c
+ || yuv2planeX == yuv2planeX_9BE_c
+ || yuv2planeX == yuv2planeX_9LE_c
+ || yuv2planeX == yuv2planeX_16BE_c
+ || yuv2planeX == yuv2planeX_16LE_c
+ || yuv2planeX == yuv2planeX_8_c) || !ARCH_X86);
+
+ if(use_mmx_vfilter){
+ vLumFilter= c->lumMmxFilter;
+ vChrFilter= c->chrMmxFilter;
+ }
+
if (vLumFilterSize == 1) {
yuv2plane1(lumSrcPtr[0], dest[0], dstW, c->lumDither8, 0);
} else {
- yuv2planeX(vLumFilter + dstY * vLumFilterSize, vLumFilterSize,
+ yuv2planeX(vLumFilter, vLumFilterSize,
lumSrcPtr, dest[0], dstW, c->lumDither8, 0);
}
if (!((dstY&chrSkipMask) || isGray(dstFormat))) {
if (yuv2nv12cX) {
- yuv2nv12cX(c, vChrFilter + chrDstY * vChrFilterSize, vChrFilterSize, chrUSrcPtr, chrVSrcPtr, dest[1], chrDstW);
+ yuv2nv12cX(c, vChrFilter, vChrFilterSize, chrUSrcPtr, chrVSrcPtr, dest[1], chrDstW);
} else if (vChrFilterSize == 1) {
yuv2plane1(chrUSrcPtr[0], dest[1], chrDstW, c->chrDither8, 0);
yuv2plane1(chrVSrcPtr[0], dest[2], chrDstW, c->chrDither8, 3);
} else {
- yuv2planeX(vChrFilter + chrDstY * vChrFilterSize, vChrFilterSize,
+ yuv2planeX(vChrFilter, vChrFilterSize,
chrUSrcPtr, dest[1], chrDstW, c->chrDither8, 0);
- yuv2planeX(vChrFilter + chrDstY * vChrFilterSize, vChrFilterSize,
- chrVSrcPtr, dest[2], chrDstW, c->chrDither8, 3);
+ yuv2planeX(vChrFilter, vChrFilterSize,
+ chrVSrcPtr, dest[2], chrDstW, c->chrDither8, use_mmx_vfilter ? (c->uv_offx2 >> 1) : 3);
}
}
if (CONFIG_SWSCALE_ALPHA && alpPixBuf){
+ if(use_mmx_vfilter){
+ vLumFilter= c->alpMmxFilter;
+ }
if (vLumFilterSize == 1) {
yuv2plane1(alpSrcPtr[0], dest[3], dstW, c->lumDither8, 0);
} else {
- yuv2planeX(vLumFilter + dstY * vLumFilterSize, vLumFilterSize,
+ yuv2planeX(vLumFilter, vLumFilterSize,
alpSrcPtr, dest[3], dstW, c->lumDither8, 0);
}
}
@@ -2695,8 +2879,8 @@ static av_cold void sws_init_swScale_c(SwsContext *c)
case PIX_FMT_YUV422P9LE:
case PIX_FMT_YUV420P9LE:
case PIX_FMT_YUV422P10LE:
- case PIX_FMT_YUV444P10LE:
case PIX_FMT_YUV420P10LE:
+ case PIX_FMT_YUV444P10LE:
case PIX_FMT_YUV420P16LE:
case PIX_FMT_YUV422P16LE:
case PIX_FMT_YUV444P16LE: c->chrToYV12 = bswap16UV_c; break;
@@ -2732,6 +2916,7 @@ static av_cold void sws_init_swScale_c(SwsContext *c)
case PIX_FMT_RGB565BE: c->chrToYV12 = rgb16beToUV_half_c; break;
case PIX_FMT_RGB555LE: c->chrToYV12 = rgb15leToUV_half_c; break;
case PIX_FMT_RGB555BE: c->chrToYV12 = rgb15beToUV_half_c; break;
+ case PIX_FMT_GBR24P : c->chrToYV12 = gbr24pToUV_half_c; break;
}
} else {
switch(srcFormat) {
@@ -2770,9 +2955,9 @@ static av_cold void sws_init_swScale_c(SwsContext *c)
case PIX_FMT_YUV444P9LE:
case PIX_FMT_YUV422P9LE:
case PIX_FMT_YUV420P9LE:
- case PIX_FMT_YUV444P10LE:
case PIX_FMT_YUV422P10LE:
case PIX_FMT_YUV420P10LE:
+ case PIX_FMT_YUV444P10LE:
case PIX_FMT_YUV420P16LE:
case PIX_FMT_YUV422P16LE:
case PIX_FMT_YUV444P16LE:
@@ -2825,9 +3010,11 @@ static av_cold void sws_init_swScale_c(SwsContext *c)
case PIX_FMT_ABGR:
case PIX_FMT_ARGB: c->alpToYV12 = abgrToA_c; break;
case PIX_FMT_Y400A: c->alpToYV12 = uyvyToY_c; break;
+ case PIX_FMT_PAL8 : c->alpToYV12 = palToA_c; break;
}
}
+
if (c->srcBpc == 8) {
if (c->dstBpc <= 10) {
c->hyScale = c->hcScale = hScale8To15_c;
diff --git a/libswscale/swscale.h b/libswscale/swscale.h
index 91378a40c4..845121a0c0 100644
--- a/libswscale/swscale.h
+++ b/libswscale/swscale.h
@@ -1,20 +1,20 @@
/*
- * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>
+ * Copyright (C) 2001-2011 Michael Niedermayer <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
*/
@@ -55,6 +55,9 @@
#ifndef FF_API_SWS_CPU_CAPS
#define FF_API_SWS_CPU_CAPS (LIBSWSCALE_VERSION_MAJOR < 3)
#endif
+#ifndef FF_API_SWS_FORMAT_NAME
+#define FF_API_SWS_FORMAT_NAME (LIBSWSCALE_VERSION_MAJOR < 3)
+#endif
/**
* Returns the LIBSWSCALE_VERSION_INT constant.
diff --git a/libswscale/swscale_internal.h b/libswscale/swscale_internal.h
index 7bfdc7eb6b..44ff166e55 100644
--- a/libswscale/swscale_internal.h
+++ b/libswscale/swscale_internal.h
@@ -1,20 +1,20 @@
/*
- * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>
+ * Copyright (C) 2001-2011 Michael Niedermayer <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
*/
@@ -38,6 +38,8 @@
#define MAX_FILTER_SIZE 256
+#define DITHER1XBPP
+
#if HAVE_BIGENDIAN
#define ALT32_CORR (-1)
#else
@@ -347,8 +349,8 @@ typedef struct SwsContext {
#define V_TEMP "11*8+4*4*256*2+32"
#define Y_TEMP "11*8+4*4*256*2+40"
#define ALP_MMX_FILTER_OFFSET "11*8+4*4*256*2+48"
-#define UV_OFF_PX "11*8+4*4*256*3+48"
-#define UV_OFF_BYTE "11*8+4*4*256*3+56"
+#define UV_OFF "11*8+4*4*256*3+48"
+#define UV_OFFx2 "11*8+4*4*256*3+56"
#define DITHER16 "11*8+4*4*256*3+64"
#define DITHER32 "11*8+4*4*256*3+80"
@@ -376,8 +378,8 @@ typedef struct SwsContext {
// alignment of these values is not necessary, but merely here
// to maintain the same offset across x8632 and x86-64. Once we
// use proper offset macros in the asm, they can be removed.
- DECLARE_ALIGNED(8, ptrdiff_t, uv_off_px); ///< offset (in pixels) between u and v planes
- DECLARE_ALIGNED(8, ptrdiff_t, uv_off_byte); ///< offset (in bytes) between u and v planes
+ DECLARE_ALIGNED(8, ptrdiff_t, uv_off); ///< offset (in pixels) between u and v planes
+ DECLARE_ALIGNED(8, ptrdiff_t, uv_offx2); ///< offset (in bytes) between u and v planes
DECLARE_ALIGNED(8, uint16_t, dither16)[8];
DECLARE_ALIGNED(8, uint32_t, dither32)[8];
@@ -411,6 +413,7 @@ typedef struct SwsContext {
#if HAVE_VIS
DECLARE_ALIGNED(8, uint64_t, sparc_coeffs)[10];
#endif
+ int use_mmx_vfilter;
/* function pointers for swScale() */
yuv2planar1_fn yuv2plane1;
@@ -420,12 +423,12 @@ typedef struct SwsContext {
yuv2packed2_fn yuv2packed2;
yuv2packedX_fn yuv2packedX;
- void (*lumToYV12)(uint8_t *dst, const uint8_t *src,
+ void (*lumToYV12)(uint8_t *dst, const uint8_t *src, const uint8_t *src2, const uint8_t *src3,
int width, uint32_t *pal); ///< Unscaled conversion of luma plane to YV12 for horizontal scaler.
- void (*alpToYV12)(uint8_t *dst, const uint8_t *src,
+ void (*alpToYV12)(uint8_t *dst, const uint8_t *src, const uint8_t *src2, const uint8_t *src3,
int width, uint32_t *pal); ///< Unscaled conversion of alpha plane to YV12 for horizontal scaler.
void (*chrToYV12)(uint8_t *dstU, uint8_t *dstV,
- const uint8_t *src1, const uint8_t *src2,
+ const uint8_t *src1, const uint8_t *src2, const uint8_t *src3,
int width, uint32_t *pal); ///< Unscaled conversion of chroma planes to YV12 for horizontal scaler.
/**
@@ -529,7 +532,13 @@ SwsFunc ff_yuv2rgb_init_altivec(SwsContext *c);
SwsFunc ff_yuv2rgb_get_func_ptr_bfin(SwsContext *c);
void ff_bfin_get_unscaled_swscale(SwsContext *c);
+#if FF_API_SWS_FORMAT_NAME
+/**
+ * @deprecated Use av_get_pix_fmt_name() instead.
+ */
+attribute_deprecated
const char *sws_format_name(enum PixelFormat format);
+#endif
#define is16BPS(x) \
(av_pix_fmt_descriptors[x].comp[0].depth_minus1 == 15)
@@ -538,6 +547,8 @@ const char *sws_format_name(enum PixelFormat format);
(av_pix_fmt_descriptors[x].comp[0].depth_minus1 == 8 || \
av_pix_fmt_descriptors[x].comp[0].depth_minus1 == 9)
+#define isNBPS(x) is9_OR_10BPS(x)
+
#define isBE(x) \
(av_pix_fmt_descriptors[x].flags & PIX_FMT_BE)
@@ -551,7 +562,6 @@ const char *sws_format_name(enum PixelFormat format);
#define isRGB(x) \
(av_pix_fmt_descriptors[x].flags & PIX_FMT_RGB)
-
#if 0 // FIXME
#define isGray(x) \
(!(av_pix_fmt_descriptors[x].flags & PIX_FMT_PAL) && \
@@ -559,7 +569,7 @@ const char *sws_format_name(enum PixelFormat format);
#else
#define isGray(x) ( \
(x)==PIX_FMT_GRAY8 \
- || (x)==PIX_FMT_Y400A \
+ || (x)==PIX_FMT_GRAY8A \
|| (x)==PIX_FMT_GRAY16BE \
|| (x)==PIX_FMT_GRAY16LE \
)
@@ -568,6 +578,8 @@ const char *sws_format_name(enum PixelFormat format);
#define isRGBinInt(x) ( \
(x)==PIX_FMT_RGB48BE \
|| (x)==PIX_FMT_RGB48LE \
+ || (x)==PIX_FMT_RGBA64BE \
+ || (x)==PIX_FMT_RGBA64LE \
|| (x)==PIX_FMT_RGB32 \
|| (x)==PIX_FMT_RGB32_1 \
|| (x)==PIX_FMT_RGB24 \
@@ -586,6 +598,8 @@ const char *sws_format_name(enum PixelFormat format);
#define isBGRinInt(x) ( \
(x)==PIX_FMT_BGR48BE \
|| (x)==PIX_FMT_BGR48LE \
+ || (x)==PIX_FMT_BGRA64BE \
+ || (x)==PIX_FMT_BGRA64LE \
|| (x)==PIX_FMT_BGR32 \
|| (x)==PIX_FMT_BGR32_1 \
|| (x)==PIX_FMT_BGR24 \
@@ -601,18 +615,51 @@ const char *sws_format_name(enum PixelFormat format);
|| (x)==PIX_FMT_MONOBLACK \
|| (x)==PIX_FMT_MONOWHITE \
)
+
+#define isRGBinBytes(x) ( \
+ (x)==PIX_FMT_RGB48BE \
+ || (x)==PIX_FMT_RGB48LE \
+ || (x)==PIX_FMT_RGBA64BE \
+ || (x)==PIX_FMT_RGBA64LE \
+ || (x)==PIX_FMT_RGBA \
+ || (x)==PIX_FMT_ARGB \
+ || (x)==PIX_FMT_RGB24 \
+ )
+#define isBGRinBytes(x) ( \
+ (x)==PIX_FMT_BGR48BE \
+ || (x)==PIX_FMT_BGR48LE \
+ || (x)==PIX_FMT_BGRA64BE \
+ || (x)==PIX_FMT_BGRA64LE \
+ || (x)==PIX_FMT_BGRA \
+ || (x)==PIX_FMT_ABGR \
+ || (x)==PIX_FMT_BGR24 \
+ )
+
#define isAnyRGB(x) ( \
isRGBinInt(x) \
|| isBGRinInt(x) \
+ || (x)==PIX_FMT_GBR24P \
)
+
#define isALPHA(x) \
(av_pix_fmt_descriptors[x].nb_components == 2 || \
av_pix_fmt_descriptors[x].nb_components == 4)
+#if 1
+#define isPacked(x) ( \
+ (x)==PIX_FMT_PAL8 \
+ || (x)==PIX_FMT_YUYV422 \
+ || (x)==PIX_FMT_UYVY422 \
+ || (x)==PIX_FMT_Y400A \
+ || isRGBinInt(x) \
+ || isBGRinInt(x) \
+ )
+#else
#define isPacked(x) \
(av_pix_fmt_descriptors[x].nb_components >= 2 && \
!(av_pix_fmt_descriptors[x].flags & PIX_FMT_PLANAR))
+#endif
#define isPlanar(x) \
(av_pix_fmt_descriptors[x].nb_components >= 2 && \
(av_pix_fmt_descriptors[x].flags & PIX_FMT_PLANAR))
@@ -621,6 +668,9 @@ const char *sws_format_name(enum PixelFormat format);
extern const uint64_t ff_dither4[2];
extern const uint64_t ff_dither8[2];
+extern const uint8_t dithers[8][8][8];
+extern const uint16_t dither_scale[15][16];
+
extern const AVClass sws_context_class;
diff --git a/libswscale/swscale_unscaled.c b/libswscale/swscale_unscaled.c
index 86bc5acda6..80ae5fc287 100644
--- a/libswscale/swscale_unscaled.c
+++ b/libswscale/swscale_unscaled.c
@@ -1,20 +1,20 @@
/*
- * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>
+ * Copyright (C) 2001-2011 Michael Niedermayer <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
*/
@@ -34,48 +34,6 @@
#include "libavutil/bswap.h"
#include "libavutil/pixdesc.h"
-DECLARE_ALIGNED(8, const uint8_t, dither_8x8_1)[8][8] = {
- { 0, 1, 0, 1, 0, 1, 0, 1,},
- { 1, 0, 1, 0, 1, 0, 1, 0,},
- { 0, 1, 0, 1, 0, 1, 0, 1,},
- { 1, 0, 1, 0, 1, 0, 1, 0,},
- { 0, 1, 0, 1, 0, 1, 0, 1,},
- { 1, 0, 1, 0, 1, 0, 1, 0,},
- { 0, 1, 0, 1, 0, 1, 0, 1,},
- { 1, 0, 1, 0, 1, 0, 1, 0,},
-};
-DECLARE_ALIGNED(8, const uint8_t, dither_8x8_3)[8][8] = {
- { 1, 2, 1, 2, 1, 2, 1, 2,},
- { 3, 0, 3, 0, 3, 0, 3, 0,},
- { 1, 2, 1, 2, 1, 2, 1, 2,},
- { 3, 0, 3, 0, 3, 0, 3, 0,},
- { 1, 2, 1, 2, 1, 2, 1, 2,},
- { 3, 0, 3, 0, 3, 0, 3, 0,},
- { 1, 2, 1, 2, 1, 2, 1, 2,},
- { 3, 0, 3, 0, 3, 0, 3, 0,},
-};
-DECLARE_ALIGNED(8, const uint8_t, dither_8x8_64)[8][8] = {
- { 18, 34, 30, 46, 17, 33, 29, 45,},
- { 50, 2, 62, 14, 49, 1, 61, 13,},
- { 26, 42, 22, 38, 25, 41, 21, 37,},
- { 58, 10, 54, 6, 57, 9, 53, 5,},
- { 16, 32, 28, 44, 19, 35, 31, 47,},
- { 48, 0, 60, 12, 51, 3, 63, 15,},
- { 24, 40, 20, 36, 27, 43, 23, 39,},
- { 56, 8, 52, 4, 59, 11, 55, 7,},
-};
-extern const uint8_t dither_8x8_128[8][8];
-DECLARE_ALIGNED(8, const uint8_t, dither_8x8_256)[8][8] = {
- { 72, 136, 120, 184, 68, 132, 116, 180,},
- { 200, 8, 248, 56, 196, 4, 244, 52,},
- { 104, 168, 88, 152, 100, 164, 84, 148,},
- { 232, 40, 216, 24, 228, 36, 212, 20,},
- { 64, 128, 102, 176, 76, 140, 124, 188,},
- { 192, 0, 240, 48, 204, 12, 252, 60,},
- { 96, 160, 80, 144, 108, 172, 92, 156,},
- { 224, 32, 208, 16, 236, 44, 220, 28,},
-};
-
#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))
@@ -263,7 +221,7 @@ static int palToRgbWrapper(SwsContext *c, const uint8_t* src[], int srcStride[],
uint8_t *dstPtr= dst[0] + dstStride[0]*srcSliceY;
const uint8_t *srcPtr= src[0];
- if (srcFormat == PIX_FMT_Y400A) {
+ if (srcFormat == PIX_FMT_GRAY8A) {
switch (dstFormat) {
case PIX_FMT_RGB32 : conv = gray8aToPacked32; break;
case PIX_FMT_BGR32 : conv = gray8aToPacked32; break;
@@ -285,7 +243,7 @@ static int palToRgbWrapper(SwsContext *c, const uint8_t* src[], int srcStride[],
if (!conv)
av_log(c, AV_LOG_ERROR, "internal error %s -> %s converter\n",
- sws_format_name(srcFormat), sws_format_name(dstFormat));
+ av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));
else {
for (i=0; i<srcSliceH; i++) {
conv(srcPtr, dstPtr, c->srcW, (uint8_t *) c->pal_rgb);
@@ -297,6 +255,92 @@ static int palToRgbWrapper(SwsContext *c, const uint8_t* src[], int srcStride[],
return srcSliceH;
}
+static void gbr24ptopacked24(const uint8_t* src[], int srcStride[], uint8_t* dst, int dstStride, int srcSliceH, int width)
+{
+ int x, h, i;
+ for (h = 0; h < srcSliceH; h++) {
+ uint8_t *dest = dst + dstStride * h;
+ for (x = 0; x < width; x++) {
+ *dest++ = src[0][x];
+ *dest++ = src[1][x];
+ *dest++ = src[2][x];
+ }
+
+ for (i = 0; i < 3; i++)
+ src[i] += srcStride[i];
+ }
+}
+
+static void gbr24ptopacked32(const uint8_t* src[], int srcStride[], uint8_t* dst, int dstStride, int srcSliceH, int alpha_first, int width)
+{
+ int x, h, i;
+ for (h = 0; h < srcSliceH; h++) {
+ uint8_t *dest = dst + dstStride * h;
+
+ if (alpha_first) {
+ for (x = 0; x < width; x++) {
+ *dest++ = 0xff;
+ *dest++ = src[0][x];
+ *dest++ = src[1][x];
+ *dest++ = src[2][x];
+ }
+ } else {
+ for (x = 0; x < width; x++) {
+ *dest++ = src[0][x];
+ *dest++ = src[1][x];
+ *dest++ = src[2][x];
+ *dest++ = 0xff;
+ }
+ }
+
+ for (i = 0; i < 3; i++)
+ src[i] += srcStride[i];
+ }
+}
+
+static int planarRgbToRgbWrapper(SwsContext *c, const uint8_t* src[], int srcStride[], int srcSliceY,
+ int srcSliceH, uint8_t* dst[], int dstStride[])
+{
+ int alpha_first = 0;
+ if (c->srcFormat != PIX_FMT_GBR24P) {
+ av_log(c, AV_LOG_ERROR, "unsupported planar RGB conversion %s -> %s\n",
+ av_get_pix_fmt_name(c->srcFormat), av_get_pix_fmt_name(c->dstFormat));
+ return srcSliceH;
+ }
+
+ switch (c->dstFormat) {
+ case PIX_FMT_BGR24:
+ gbr24ptopacked24((const uint8_t* []) {src[1], src[0], src[2]}, (int []) {srcStride[1], srcStride[0], srcStride[2]},
+ dst[0] + srcSliceY * dstStride[0], dstStride[0], srcSliceH, c->srcW);
+ break;
+
+ case PIX_FMT_RGB24:
+ gbr24ptopacked24((const uint8_t* []) {src[2], src[0], src[1]}, (int []) {srcStride[2], srcStride[0], srcStride[1]},
+ dst[0] + srcSliceY * dstStride[0], dstStride[0], srcSliceH, c->srcW);
+ break;
+
+ case PIX_FMT_ARGB:
+ alpha_first = 1;
+ case PIX_FMT_RGBA:
+ gbr24ptopacked32((const uint8_t* []) {src[2], src[0], src[1]}, (int []) {srcStride[2], srcStride[0], srcStride[1]},
+ dst[0] + srcSliceY * dstStride[0], dstStride[0], srcSliceH, alpha_first, c->srcW);
+ break;
+
+ case PIX_FMT_ABGR:
+ alpha_first = 1;
+ case PIX_FMT_BGRA:
+ gbr24ptopacked32((const uint8_t* []) {src[1], src[0], src[2]}, (int []) {srcStride[1], srcStride[0], srcStride[2]},
+ dst[0] + srcSliceY * dstStride[0], dstStride[0], srcSliceH, alpha_first, c->srcW);
+ break;
+
+ default:
+ av_log(c, AV_LOG_ERROR, "unsupported planar RGB conversion %s -> %s\n",
+ av_get_pix_fmt_name(c->srcFormat), av_get_pix_fmt_name(c->dstFormat));
+ }
+
+ return srcSliceH;
+}
+
#define isRGBA32(x) ( \
(x) == PIX_FMT_ARGB \
|| (x) == PIX_FMT_RGBA \
@@ -372,7 +416,7 @@ static int rgbToRgbWrapper(SwsContext *c, const uint8_t* src[], int srcStride[],
if (!conv) {
av_log(c, AV_LOG_ERROR, "internal error %s -> %s converter\n",
- sws_format_name(srcFormat), sws_format_name(dstFormat));
+ av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));
} else {
const uint8_t *srcPtr= src[0];
uint8_t *dstPtr= dst[0];
@@ -454,27 +498,28 @@ static int packedCopyWrapper(SwsContext *c, const uint8_t* src[], int srcStride[
return srcSliceH;
}
-#define clip9(x) av_clip_uintp2(x, 9)
-#define clip10(x) av_clip_uintp2(x, 10)
-#define DITHER_COPY(dst, dstStride, wfunc, src, srcStride, rfunc, dithers, shift, clip) \
- for (i = 0; i < height; i++) { \
- const uint8_t *dither = dithers[i & 7]; \
- for (j = 0; j < length - 7; j += 8) { \
- wfunc(&dst[j + 0], clip((rfunc(&src[j + 0]) + dither[0]) >> shift)); \
- wfunc(&dst[j + 1], clip((rfunc(&src[j + 1]) + dither[1]) >> shift)); \
- wfunc(&dst[j + 2], clip((rfunc(&src[j + 2]) + dither[2]) >> shift)); \
- wfunc(&dst[j + 3], clip((rfunc(&src[j + 3]) + dither[3]) >> shift)); \
- wfunc(&dst[j + 4], clip((rfunc(&src[j + 4]) + dither[4]) >> shift)); \
- wfunc(&dst[j + 5], clip((rfunc(&src[j + 5]) + dither[5]) >> shift)); \
- wfunc(&dst[j + 6], clip((rfunc(&src[j + 6]) + dither[6]) >> shift)); \
- wfunc(&dst[j + 7], clip((rfunc(&src[j + 7]) + dither[7]) >> shift)); \
- } \
- for (; j < length; j++) \
- wfunc(&dst[j], (rfunc(&src[j]) + dither[j & 7]) >> shift); \
- dst += dstStride; \
- src += srcStride; \
+#define DITHER_COPY(dst, dstStride, src, srcStride, bswap, dbswap)\
+ uint16_t scale= dither_scale[dst_depth-1][src_depth-1];\
+ int shift= src_depth-dst_depth + dither_scale[src_depth-2][dst_depth-1];\
+ for (i = 0; i < height; i++) {\
+ const uint8_t *dither= dithers[src_depth-9][i&7];\
+ for (j = 0; j < length-7; j+=8){\
+ dst[j+0] = dbswap((bswap(src[j+0]) + dither[0])*scale>>shift);\
+ dst[j+1] = dbswap((bswap(src[j+1]) + dither[1])*scale>>shift);\
+ dst[j+2] = dbswap((bswap(src[j+2]) + dither[2])*scale>>shift);\
+ dst[j+3] = dbswap((bswap(src[j+3]) + dither[3])*scale>>shift);\
+ dst[j+4] = dbswap((bswap(src[j+4]) + dither[4])*scale>>shift);\
+ dst[j+5] = dbswap((bswap(src[j+5]) + dither[5])*scale>>shift);\
+ dst[j+6] = dbswap((bswap(src[j+6]) + dither[6])*scale>>shift);\
+ dst[j+7] = dbswap((bswap(src[j+7]) + dither[7])*scale>>shift);\
+ }\
+ for (; j < length; j++)\
+ dst[j] = dbswap((bswap(src[j]) + dither[j&7])*scale>>shift);\
+ dst += dstStride;\
+ src += srcStride;\
}
+
static int planarCopyWrapper(SwsContext *c, const uint8_t* src[], int srcStride[], int srcSliceY,
int srcSliceH, uint8_t* dst[], int dstStride[])
{
@@ -485,6 +530,7 @@ static int planarCopyWrapper(SwsContext *c, const uint8_t* src[], int srcStride[
int height= (plane==0 || plane==3) ? srcSliceH: -((-srcSliceH)>>c->chrDstVSubSample);
const uint8_t *srcPtr= src[plane];
uint8_t *dstPtr= dst[plane] + dstStride[plane]*y;
+ int shiftonly= plane==1 || plane==2 || (!c->srcRange && plane==0);
if (!dst[plane]) continue;
// ignore palette for GRAY8
@@ -494,151 +540,84 @@ static int planarCopyWrapper(SwsContext *c, const uint8_t* src[], int srcStride[
length*=2;
fillPlane(dst[plane], dstStride[plane], length, height, y, (plane==3) ? 255 : 128);
} else {
- if(is9_OR_10BPS(c->srcFormat)) {
+ if(isNBPS(c->srcFormat) || isNBPS(c->dstFormat)
+ || (is16BPS(c->srcFormat) != is16BPS(c->dstFormat))
+ ) {
const int src_depth = av_pix_fmt_descriptors[c->srcFormat].comp[plane].depth_minus1+1;
const int dst_depth = av_pix_fmt_descriptors[c->dstFormat].comp[plane].depth_minus1+1;
const uint16_t *srcPtr2 = (const uint16_t*)srcPtr;
+ uint16_t *dstPtr2 = (uint16_t*)dstPtr;
- if (is16BPS(c->dstFormat)) {
- uint16_t *dstPtr2 = (uint16_t*)dstPtr;
-#define COPY9_OR_10TO16(rfunc, wfunc) \
- for (i = 0; i < height; i++) { \
- for (j = 0; j < length; j++) { \
- int srcpx = rfunc(&srcPtr2[j]); \
- wfunc(&dstPtr2[j], (srcpx<<(16-src_depth)) | (srcpx>>(2*src_depth-16))); \
- } \
- dstPtr2 += dstStride[plane]/2; \
- srcPtr2 += srcStride[plane]/2; \
+ if (dst_depth == 8) {
+ if(isBE(c->srcFormat) == HAVE_BIGENDIAN){
+ DITHER_COPY(dstPtr, dstStride[plane], srcPtr2, srcStride[plane]/2, , )
+ } else {
+ DITHER_COPY(dstPtr, dstStride[plane], srcPtr2, srcStride[plane]/2, av_bswap16, )
}
- if (isBE(c->dstFormat)) {
- if (isBE(c->srcFormat)) {
- COPY9_OR_10TO16(AV_RB16, AV_WB16);
- } else {
- COPY9_OR_10TO16(AV_RL16, AV_WB16);
+ } else if (src_depth == 8) {
+ for (i = 0; i < height; i++) {
+ #define COPY816(w)\
+ if(shiftonly){\
+ for (j = 0; j < length; j++)\
+ w(&dstPtr2[j], srcPtr[j]<<(dst_depth-8));\
+ }else{\
+ for (j = 0; j < length; j++)\
+ w(&dstPtr2[j], (srcPtr[j]<<(dst_depth-8)) |\
+ (srcPtr[j]>>(2*8-dst_depth)));\
}
- } else {
- if (isBE(c->srcFormat)) {
- COPY9_OR_10TO16(AV_RB16, AV_WL16);
+ if(isBE(c->dstFormat)){
+ COPY816(AV_WB16)
} else {
- COPY9_OR_10TO16(AV_RL16, AV_WL16);
+ COPY816(AV_WL16)
}
+ dstPtr2 += dstStride[plane]/2;
+ srcPtr += srcStride[plane];
}
- } else if (is9_OR_10BPS(c->dstFormat)) {
- uint16_t *dstPtr2 = (uint16_t*)dstPtr;
-#define COPY9_OR_10TO9_OR_10(loop) \
- for (i = 0; i < height; i++) { \
- for (j = 0; j < length; j++) { \
- loop; \
- } \
- dstPtr2 += dstStride[plane]/2; \
- srcPtr2 += srcStride[plane]/2; \
- }
-#define COPY9_OR_10TO9_OR_10_2(rfunc, wfunc) \
- if (dst_depth > src_depth) { \
- COPY9_OR_10TO9_OR_10(int srcpx = rfunc(&srcPtr2[j]); \
- wfunc(&dstPtr2[j], (srcpx << 1) | (srcpx >> 9))); \
- } else if (dst_depth < src_depth) { \
- DITHER_COPY(dstPtr2, dstStride[plane]/2, wfunc, \
- srcPtr2, srcStride[plane]/2, rfunc, \
- dither_8x8_1, 1, clip9); \
- } else { \
- COPY9_OR_10TO9_OR_10(wfunc(&dstPtr2[j], rfunc(&srcPtr2[j]))); \
- }
- if (isBE(c->dstFormat)) {
- if (isBE(c->srcFormat)) {
- COPY9_OR_10TO9_OR_10_2(AV_RB16, AV_WB16);
- } else {
- COPY9_OR_10TO9_OR_10_2(AV_RL16, AV_WB16);
- }
- } else {
- if (isBE(c->srcFormat)) {
- COPY9_OR_10TO9_OR_10_2(AV_RB16, AV_WL16);
+ } else if (src_depth <= dst_depth) {
+ for (i = 0; i < height; i++) {
+#define COPY_UP(r,w) \
+ if(shiftonly){\
+ for (j = 0; j < length; j++){ \
+ unsigned int v= r(&srcPtr2[j]);\
+ w(&dstPtr2[j], v<<(dst_depth-src_depth));\
+ }\
+ }else{\
+ for (j = 0; j < length; j++){ \
+ unsigned int v= r(&srcPtr2[j]);\
+ w(&dstPtr2[j], (v<<(dst_depth-src_depth)) | \
+ (v>>(2*src_depth-dst_depth)));\
+ }\
+ }
+ if(isBE(c->srcFormat)){
+ if(isBE(c->dstFormat)){
+ COPY_UP(AV_RB16, AV_WB16)
+ } else {
+ COPY_UP(AV_RB16, AV_WL16)
+ }
} else {
- COPY9_OR_10TO9_OR_10_2(AV_RL16, AV_WL16);
+ if(isBE(c->dstFormat)){
+ COPY_UP(AV_RL16, AV_WB16)
+ } else {
+ COPY_UP(AV_RL16, AV_WL16)
+ }
}
+ dstPtr2 += dstStride[plane]/2;
+ srcPtr2 += srcStride[plane]/2;
}
} else {
-#define W8(a, b) { *(a) = (b); }
-#define COPY9_OR_10TO8(rfunc) \
- if (src_depth == 9) { \
- DITHER_COPY(dstPtr, dstStride[plane], W8, \
- srcPtr2, srcStride[plane]/2, rfunc, \
- dither_8x8_1, 1, av_clip_uint8); \
- } else { \
- DITHER_COPY(dstPtr, dstStride[plane], W8, \
- srcPtr2, srcStride[plane]/2, rfunc, \
- dither_8x8_3, 2, av_clip_uint8); \
- }
- if (isBE(c->srcFormat)) {
- COPY9_OR_10TO8(AV_RB16);
- } else {
- COPY9_OR_10TO8(AV_RL16);
- }
- }
- } else if(is9_OR_10BPS(c->dstFormat)) {
- const int dst_depth = av_pix_fmt_descriptors[c->dstFormat].comp[plane].depth_minus1+1;
- uint16_t *dstPtr2 = (uint16_t*)dstPtr;
-
- if (is16BPS(c->srcFormat)) {
- const uint16_t *srcPtr2 = (const uint16_t*)srcPtr;
-#define COPY16TO9_OR_10(rfunc, wfunc) \
- if (dst_depth == 9) { \
- DITHER_COPY(dstPtr2, dstStride[plane]/2, wfunc, \
- srcPtr2, srcStride[plane]/2, rfunc, \
- dither_8x8_128, 7, clip9); \
- } else { \
- DITHER_COPY(dstPtr2, dstStride[plane]/2, wfunc, \
- srcPtr2, srcStride[plane]/2, rfunc, \
- dither_8x8_64, 6, clip10); \
- }
- if (isBE(c->dstFormat)) {
- if (isBE(c->srcFormat)) {
- COPY16TO9_OR_10(AV_RB16, AV_WB16);
+ if(isBE(c->srcFormat) == HAVE_BIGENDIAN){
+ if(isBE(c->dstFormat) == HAVE_BIGENDIAN){
+ DITHER_COPY(dstPtr2, dstStride[plane]/2, srcPtr2, srcStride[plane]/2, , )
} else {
- COPY16TO9_OR_10(AV_RL16, AV_WB16);
+ DITHER_COPY(dstPtr2, dstStride[plane]/2, srcPtr2, srcStride[plane]/2, , av_bswap16)
}
- } else {
- if (isBE(c->srcFormat)) {
- COPY16TO9_OR_10(AV_RB16, AV_WL16);
+ }else{
+ if(isBE(c->dstFormat) == HAVE_BIGENDIAN){
+ DITHER_COPY(dstPtr2, dstStride[plane]/2, srcPtr2, srcStride[plane]/2, av_bswap16, )
} else {
- COPY16TO9_OR_10(AV_RL16, AV_WL16);
+ DITHER_COPY(dstPtr2, dstStride[plane]/2, srcPtr2, srcStride[plane]/2, av_bswap16, av_bswap16)
}
}
- } else /* 8bit */ {
-#define COPY8TO9_OR_10(wfunc) \
- for (i = 0; i < height; i++) { \
- for (j = 0; j < length; j++) { \
- const int srcpx = srcPtr[j]; \
- wfunc(&dstPtr2[j], (srcpx<<(dst_depth-8)) | (srcpx >> (16-dst_depth))); \
- } \
- dstPtr2 += dstStride[plane]/2; \
- srcPtr += srcStride[plane]; \
- }
- if (isBE(c->dstFormat)) {
- COPY8TO9_OR_10(AV_WB16);
- } else {
- COPY8TO9_OR_10(AV_WL16);
- }
- }
- } else if(is16BPS(c->srcFormat) && !is16BPS(c->dstFormat)) {
- const uint16_t *srcPtr2 = (const uint16_t*)srcPtr;
-#define COPY16TO8(rfunc) \
- DITHER_COPY(dstPtr, dstStride[plane], W8, \
- srcPtr2, srcStride[plane]/2, rfunc, \
- dither_8x8_256, 8, av_clip_uint8);
- if (isBE(c->srcFormat)) {
- COPY16TO8(AV_RB16);
- } else {
- COPY16TO8(AV_RL16);
- }
- } else if(!is16BPS(c->srcFormat) && is16BPS(c->dstFormat)) {
- for (i=0; i<height; i++) {
- for (j=0; j<length; j++) {
- dstPtr[ j<<1 ] = srcPtr[j];
- dstPtr[(j<<1)+1] = srcPtr[j];
- }
- srcPtr+= srcStride[plane];
- dstPtr+= dstStride[plane];
}
} else if(is16BPS(c->srcFormat) && is16BPS(c->dstFormat)
&& isBE(c->srcFormat) != isBE(c->dstFormat)) {
@@ -715,13 +694,18 @@ void ff_get_unscaled_swscale(SwsContext *c)
&& (!needsDither || (c->flags&(SWS_FAST_BILINEAR|SWS_POINT))))
c->swScale= rgbToRgbWrapper;
- if ((usePal(srcFormat) && (
- dstFormat == PIX_FMT_RGB32 ||
- dstFormat == PIX_FMT_RGB32_1 ||
- dstFormat == PIX_FMT_RGB24 ||
- dstFormat == PIX_FMT_BGR32 ||
- dstFormat == PIX_FMT_BGR32_1 ||
- dstFormat == PIX_FMT_BGR24)))
+#define isByteRGB(f) (\
+ f == PIX_FMT_RGB32 ||\
+ f == PIX_FMT_RGB32_1 ||\
+ f == PIX_FMT_RGB24 ||\
+ f == PIX_FMT_BGR32 ||\
+ f == PIX_FMT_BGR32_1 ||\
+ f == PIX_FMT_BGR24)
+
+ if (isAnyRGB(srcFormat) && isPlanar(srcFormat) && isByteRGB(dstFormat))
+ c->swScale= planarRgbToRgbWrapper;
+
+ if (usePal(srcFormat) && isByteRGB(dstFormat))
c->swScale= palToRgbWrapper;
if (srcFormat == PIX_FMT_YUV422P) {
@@ -787,7 +771,7 @@ static void reset_ptr(const uint8_t* src[], int format)
}
}
-static int check_image_pointers(uint8_t *data[4], enum PixelFormat pix_fmt,
+static int check_image_pointers(const uint8_t * const data[4], enum PixelFormat pix_fmt,
const int linesizes[4])
{
const AVPixFmtDescriptor *desc = &av_pix_fmt_descriptors[pix_fmt];
@@ -822,7 +806,7 @@ int attribute_align_arg sws_scale(struct SwsContext *c, const uint8_t* const src
av_log(c, AV_LOG_ERROR, "bad src image pointers\n");
return 0;
}
- if (!check_image_pointers(dst, c->dstFormat, dstStride)) {
+ if (!check_image_pointers((const uint8_t* const*)dst, c->dstFormat, dstStride)) {
av_log(c, AV_LOG_ERROR, "bad dst image pointers\n");
return 0;
}
@@ -837,9 +821,10 @@ int attribute_align_arg sws_scale(struct SwsContext *c, const uint8_t* const src
if (usePal(c->srcFormat)) {
for (i=0; i<256; i++) {
- int p, r, g, b,y,u,v;
+ int p, r, g, b, y, u, v, a = 0xff;
if(c->srcFormat == PIX_FMT_PAL8) {
p=((const uint32_t*)(srcSlice[1]))[i];
+ a= (p>>24)&0xFF;
r= (p>>16)&0xFF;
g= (p>> 8)&0xFF;
b= p &0xFF;
@@ -855,7 +840,7 @@ int attribute_align_arg sws_scale(struct SwsContext *c, const uint8_t* const src
r= (i>>3 )*255;
g= ((i>>1)&3)*85;
b= (i&1 )*255;
- } else if(c->srcFormat == PIX_FMT_GRAY8 || c->srcFormat == PIX_FMT_Y400A) {
+ } else if(c->srcFormat == PIX_FMT_GRAY8 || c->srcFormat == PIX_FMT_GRAY8A) {
r = g = b = i;
} else {
assert(c->srcFormat == PIX_FMT_BGR4_BYTE);
@@ -866,33 +851,33 @@ int attribute_align_arg sws_scale(struct SwsContext *c, const uint8_t* const src
y= av_clip_uint8((RY*r + GY*g + BY*b + ( 33<<(RGB2YUV_SHIFT-1)))>>RGB2YUV_SHIFT);
u= av_clip_uint8((RU*r + GU*g + BU*b + (257<<(RGB2YUV_SHIFT-1)))>>RGB2YUV_SHIFT);
v= av_clip_uint8((RV*r + GV*g + BV*b + (257<<(RGB2YUV_SHIFT-1)))>>RGB2YUV_SHIFT);
- c->pal_yuv[i]= y + (u<<8) + (v<<16);
+ c->pal_yuv[i]= y + (u<<8) + (v<<16) + (a<<24);
switch(c->dstFormat) {
case PIX_FMT_BGR32:
#if !HAVE_BIGENDIAN
case PIX_FMT_RGB24:
#endif
- c->pal_rgb[i]= r + (g<<8) + (b<<16);
+ c->pal_rgb[i]= r + (g<<8) + (b<<16) + (a<<24);
break;
case PIX_FMT_BGR32_1:
#if HAVE_BIGENDIAN
case PIX_FMT_BGR24:
#endif
- c->pal_rgb[i]= (r + (g<<8) + (b<<16)) << 8;
+ c->pal_rgb[i]= a + (r<<8) + (g<<16) + (b<<24);
break;
case PIX_FMT_RGB32_1:
#if HAVE_BIGENDIAN
case PIX_FMT_RGB24:
#endif
- c->pal_rgb[i]= (b + (g<<8) + (r<<16)) << 8;
+ c->pal_rgb[i]= a + (b<<8) + (g<<16) + (r<<24);
break;
case PIX_FMT_RGB32:
#if !HAVE_BIGENDIAN
case PIX_FMT_BGR24:
#endif
default:
- c->pal_rgb[i]= b + (g<<8) + (r<<16);
+ c->pal_rgb[i]= b + (g<<8) + (r<<16) + (a<<24);
}
}
}
@@ -904,7 +889,7 @@ int attribute_align_arg sws_scale(struct SwsContext *c, const uint8_t* const src
int dstStride2[4]= {dstStride[0], dstStride[1], dstStride[2], dstStride[3]};
reset_ptr(src2, c->srcFormat);
- reset_ptr((const uint8_t**)dst2, c->dstFormat);
+ reset_ptr((void*)dst2, c->dstFormat);
/* reset slice direction at end of frame */
if (srcSliceY + srcSliceH == c->srcH)
@@ -927,7 +912,7 @@ int attribute_align_arg sws_scale(struct SwsContext *c, const uint8_t* const src
dst2[3] += ( c->dstH -1)*dstStride[3];
reset_ptr(src2, c->srcFormat);
- reset_ptr((const uint8_t**)dst2, c->dstFormat);
+ reset_ptr((void*)dst2, c->dstFormat);
/* reset slice direction at end of frame */
if (!srcSliceY)
diff --git a/libswscale/utils.c b/libswscale/utils.c
index b644ed9610..3240b96030 100644
--- a/libswscale/utils.c
+++ b/libswscale/utils.c
@@ -1,24 +1,25 @@
/*
* Copyright (C) 2001-2003 Michael Niedermayer <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
*/
#define _SVID_SOURCE //needed for MAP_ANONYMOUS
+#define _DARWIN_C_SOURCE // needed for MAP_ANON
#include <inttypes.h>
#include <string.h>
#include <math.h>
@@ -46,6 +47,7 @@
#include "libavutil/mathematics.h"
#include "libavutil/opt.h"
#include "libavutil/pixdesc.h"
+#include "libavutil/avassert.h"
unsigned swscale_version(void)
{
@@ -54,13 +56,13 @@ unsigned swscale_version(void)
const char *swscale_configuration(void)
{
- return LIBAV_CONFIGURATION;
+ return FFMPEG_CONFIGURATION;
}
const char *swscale_license(void)
{
#define LICENSE_PREFIX "libswscale license: "
- return LICENSE_PREFIX LIBAV_LICENSE + sizeof(LICENSE_PREFIX) - 1;
+ return LICENSE_PREFIX FFMPEG_LICENSE + sizeof(LICENSE_PREFIX) - 1;
}
#define RET 0xC3 //near return opcode for x86
@@ -99,6 +101,10 @@ const static FormatEntry format_entries[PIX_FMT_NB] = {
[PIX_FMT_RGBA] = { 1 , 1 },
[PIX_FMT_ABGR] = { 1 , 1 },
[PIX_FMT_BGRA] = { 1 , 1 },
+ [PIX_FMT_0RGB] = { 1 , 1 },
+ [PIX_FMT_RGB0] = { 1 , 1 },
+ [PIX_FMT_0BGR] = { 1 , 1 },
+ [PIX_FMT_BGR0] = { 1 , 1 },
[PIX_FMT_GRAY16BE] = { 1 , 1 },
[PIX_FMT_GRAY16LE] = { 1 , 1 },
[PIX_FMT_YUV440P] = { 1 , 1 },
@@ -106,6 +112,8 @@ const static FormatEntry format_entries[PIX_FMT_NB] = {
[PIX_FMT_YUVA420P] = { 1 , 1 },
[PIX_FMT_RGB48BE] = { 1 , 1 },
[PIX_FMT_RGB48LE] = { 1 , 1 },
+ [PIX_FMT_RGBA64BE] = { 0 , 0 },
+ [PIX_FMT_RGBA64LE] = { 0 , 0 },
[PIX_FMT_RGB565BE] = { 1 , 1 },
[PIX_FMT_RGB565LE] = { 1 , 1 },
[PIX_FMT_RGB555BE] = { 1 , 1 },
@@ -127,6 +135,8 @@ const static FormatEntry format_entries[PIX_FMT_NB] = {
[PIX_FMT_Y400A] = { 1 , 0 },
[PIX_FMT_BGR48BE] = { 1 , 1 },
[PIX_FMT_BGR48LE] = { 1 , 1 },
+ [PIX_FMT_BGRA64BE] = { 0 , 0 },
+ [PIX_FMT_BGRA64LE] = { 0 , 0 },
[PIX_FMT_YUV420P9BE] = { 1 , 1 },
[PIX_FMT_YUV420P9LE] = { 1 , 1 },
[PIX_FMT_YUV420P10BE] = { 1 , 1 },
@@ -139,6 +149,7 @@ const static FormatEntry format_entries[PIX_FMT_NB] = {
[PIX_FMT_YUV444P9LE] = { 1 , 1 },
[PIX_FMT_YUV444P10BE] = { 1 , 1 },
[PIX_FMT_YUV444P10LE] = { 1 , 1 },
+ [PIX_FMT_GBR24P] = { 1 , 0 },
[PIX_FMT_GBRP] = { 1 , 0 },
[PIX_FMT_GBRP9LE] = { 1 , 0 },
[PIX_FMT_GBRP9BE] = { 1 , 0 },
@@ -162,13 +173,12 @@ int sws_isSupportedOutput(enum PixelFormat pix_fmt)
extern const int32_t ff_yuv2rgb_coeffs[8][4];
+#if FF_API_SWS_FORMAT_NAME
const char *sws_format_name(enum PixelFormat format)
{
- if ((unsigned)format < PIX_FMT_NB && av_pix_fmt_descriptors[format].name)
- return av_pix_fmt_descriptors[format].name;
- else
- return "Unknown format";
+ return av_get_pix_fmt_name(format);
}
+#endif
static double getSplineCoeff(double a, double b, double c, double d, double dist)
{
@@ -717,7 +727,7 @@ int sws_getColorspaceDetails(struct SwsContext *c, int **inv_table,
int *srcRange, int **table, int *dstRange,
int *brightness, int *contrast, int *saturation)
{
- if (isYUV(c->dstFormat) || isGray(c->dstFormat)) return -1;
+ if (!c || isYUV(c->dstFormat) || isGray(c->dstFormat)) return -1;
*inv_table = c->srcColorspaceTable;
*table = c->dstColorspaceTable;
@@ -737,6 +747,10 @@ static int handle_jpeg(enum PixelFormat *format)
case PIX_FMT_YUVJ422P: *format = PIX_FMT_YUV422P; return 1;
case PIX_FMT_YUVJ444P: *format = PIX_FMT_YUV444P; return 1;
case PIX_FMT_YUVJ440P: *format = PIX_FMT_YUV440P; return 1;
+ case PIX_FMT_0BGR : *format = PIX_FMT_ABGR ; return 0;
+ case PIX_FMT_BGR0 : *format = PIX_FMT_BGRA ; return 0;
+ case PIX_FMT_0RGB : *format = PIX_FMT_ARGB ; return 0;
+ case PIX_FMT_RGB0 : *format = PIX_FMT_RGBA ; return 0;
default: return 0;
}
}
@@ -753,7 +767,7 @@ SwsContext *sws_alloc_context(void)
int sws_init_context(SwsContext *c, SwsFilter *srcFilter, SwsFilter *dstFilter)
{
- int i;
+ int i, j;
int usesVFilter, usesHFilter;
int unscaled;
SwsFilter dummyFilter= {NULL, NULL, NULL, NULL};
@@ -761,7 +775,7 @@ int sws_init_context(SwsContext *c, SwsFilter *srcFilter, SwsFilter *dstFilter)
int srcH= c->srcH;
int dstW= c->dstW;
int dstH= c->dstH;
- int dst_stride = FFALIGN(dstW * sizeof(int16_t) + 16, 16), dst_stride_px = dst_stride >> 1;
+ int dst_stride = FFALIGN(dstW * sizeof(int16_t)+66, 16);
int flags, cpu_flags;
enum PixelFormat srcFormat= c->srcFormat;
enum PixelFormat dstFormat= c->dstFormat;
@@ -773,12 +787,21 @@ int sws_init_context(SwsContext *c, SwsFilter *srcFilter, SwsFilter *dstFilter)
unscaled = (srcW == dstW && srcH == dstH);
+ handle_jpeg(&srcFormat);
+ handle_jpeg(&dstFormat);
+
+ if(srcFormat!=c->srcFormat || dstFormat!=c->dstFormat){
+ av_log(c, AV_LOG_WARNING, "deprecated pixel format used, make sure you did set range correctly\n");
+ c->srcFormat= srcFormat;
+ c->dstFormat= dstFormat;
+ }
+
if (!sws_isSupportedInput(srcFormat)) {
- av_log(c, AV_LOG_ERROR, "%s is not supported as input pixel format\n", sws_format_name(srcFormat));
+ av_log(c, AV_LOG_ERROR, "%s is not supported as input pixel format\n", av_get_pix_fmt_name(srcFormat));
return AVERROR(EINVAL);
}
if (!sws_isSupportedOutput(dstFormat)) {
- av_log(c, AV_LOG_ERROR, "%s is not supported as output pixel format\n", sws_format_name(dstFormat));
+ av_log(c, AV_LOG_ERROR, "%s is not supported as output pixel format\n", av_get_pix_fmt_name(dstFormat));
return AVERROR(EINVAL);
}
@@ -826,19 +849,6 @@ int sws_init_context(SwsContext *c, SwsFilter *srcFilter, SwsFilter *dstFilter)
getSubSampleFactors(&c->chrDstHSubSample, &c->chrDstVSubSample, dstFormat);
// reuse chroma for 2 pixels RGB/BGR unless user wants full chroma interpolation
- if (flags & SWS_FULL_CHR_H_INT &&
- dstFormat != PIX_FMT_RGBA &&
- dstFormat != PIX_FMT_ARGB &&
- dstFormat != PIX_FMT_BGRA &&
- dstFormat != PIX_FMT_ABGR &&
- dstFormat != PIX_FMT_RGB24 &&
- dstFormat != PIX_FMT_BGR24) {
- av_log(c, AV_LOG_ERROR,
- "full chroma interpolation for destination format '%s' not yet implemented\n",
- sws_format_name(dstFormat));
- flags &= ~SWS_FULL_CHR_H_INT;
- c->flags = flags;
- }
if (isAnyRGB(dstFormat) && !(flags&SWS_FULL_CHR_H_INT)) c->chrDstHSubSample=1;
// drop some chroma lines if the user wants it
@@ -866,7 +876,7 @@ int sws_init_context(SwsContext *c, SwsFilter *srcFilter, SwsFilter *dstFilter)
if (c->swScale) {
if (flags&SWS_PRINT_INFO)
av_log(c, AV_LOG_INFO, "using unscaled %s -> %s special converter\n",
- sws_format_name(srcFormat), sws_format_name(dstFormat));
+ av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));
return 0;
}
}
@@ -877,18 +887,18 @@ int sws_init_context(SwsContext *c, SwsFilter *srcFilter, SwsFilter *dstFilter)
c->dstBpc = 1 + av_pix_fmt_descriptors[dstFormat].comp[0].depth_minus1;
if (c->dstBpc < 8)
c->dstBpc = 8;
+ if (isAnyRGB(srcFormat) || srcFormat == PIX_FMT_PAL8)
+ c->srcBpc = 16;
if (c->dstBpc == 16)
dst_stride <<= 1;
- FF_ALLOC_OR_GOTO(c, c->formatConvBuffer,
- (FFALIGN(srcW, 16) * 2 * FFALIGN(c->srcBpc, 8) >> 3) + 16,
- fail);
+ FF_ALLOC_OR_GOTO(c, c->formatConvBuffer, FFALIGN(srcW*2+78, 16) * 2, fail);
if (HAVE_MMX2 && cpu_flags & AV_CPU_FLAG_MMX2 && c->srcBpc == 8 && c->dstBpc <= 10) {
c->canMMX2BeUsed= (dstW >=srcW && (dstW&31)==0 && (srcW&15)==0) ? 1 : 0;
if (!c->canMMX2BeUsed && dstW >=srcW && (srcW&15)==0 && (flags&SWS_FAST_BILINEAR)) {
if (flags&SWS_PRINT_INFO)
av_log(c, AV_LOG_INFO, "output width is not a multiple of 32 -> no MMX2 scaler\n");
}
- if (usesHFilter) c->canMMX2BeUsed=0;
+ if (usesHFilter || isNBPS(c->srcFormat) || is16BPS(c->srcFormat) || isAnyRGB(c->srcFormat)) c->canMMX2BeUsed=0;
}
else
c->canMMX2BeUsed=0;
@@ -908,7 +918,7 @@ int sws_init_context(SwsContext *c, SwsFilter *srcFilter, SwsFilter *dstFilter)
c->chrXInc+= 20;
}
//we don't use the x86 asm scaler if MMX is available
- else if (HAVE_MMX && cpu_flags & AV_CPU_FLAG_MMX) {
+ else if (HAVE_MMX && cpu_flags & AV_CPU_FLAG_MMX && c->dstBpc <= 10) {
c->lumXInc = ((srcW-2)<<16)/(dstW-2) - 20;
c->chrXInc = ((c->chrSrcW-2)<<16)/(c->chrDstW-2) - 20;
}
@@ -933,15 +943,19 @@ int sws_init_context(SwsContext *c, SwsFilter *srcFilter, SwsFilter *dstFilter)
c->chrMmx2FilterCode = av_malloc(c->chrMmx2FilterCodeSize);
#endif
+#ifdef MAP_ANONYMOUS
+ if (c->lumMmx2FilterCode == MAP_FAILED || c->chrMmx2FilterCode == MAP_FAILED)
+#else
if (!c->lumMmx2FilterCode || !c->chrMmx2FilterCode)
+#endif
return AVERROR(ENOMEM);
FF_ALLOCZ_OR_GOTO(c, c->hLumFilter , (dstW /8+8)*sizeof(int16_t), fail);
FF_ALLOCZ_OR_GOTO(c, c->hChrFilter , (c->chrDstW /4+8)*sizeof(int16_t), fail);
FF_ALLOCZ_OR_GOTO(c, c->hLumFilterPos, (dstW /2/8+8)*sizeof(int32_t), fail);
FF_ALLOCZ_OR_GOTO(c, c->hChrFilterPos, (c->chrDstW/2/4+8)*sizeof(int32_t), fail);
- initMMX2HScaler( dstW, c->lumXInc, c->lumMmx2FilterCode, c->hLumFilter, c->hLumFilterPos, 8);
- initMMX2HScaler(c->chrDstW, c->chrXInc, c->chrMmx2FilterCode, c->hChrFilter, c->hChrFilterPos, 4);
+ initMMX2HScaler( dstW, c->lumXInc, c->lumMmx2FilterCode, c->hLumFilter, (uint32_t*)c->hLumFilterPos, 8);
+ initMMX2HScaler(c->chrDstW, c->chrXInc, c->chrMmx2FilterCode, c->hChrFilter, (uint32_t*)c->hChrFilterPos, 4);
#ifdef MAP_ANONYMOUS
mprotect(c->lumMmx2FilterCode, c->lumMmx2FilterCodeSize, PROT_EXEC | PROT_READ);
@@ -1010,7 +1024,7 @@ int sws_init_context(SwsContext *c, SwsFilter *srcFilter, SwsFilter *dstFilter)
c->vLumBufSize= c->vLumFilterSize;
c->vChrBufSize= c->vChrFilterSize;
for (i=0; i<dstH; i++) {
- int chrI= i*c->chrDstH / dstH;
+ int chrI= (int64_t)i*c->chrDstH / dstH;
int nextSlice= FFMAX(c->vLumFilterPos[i ] + c->vLumFilterSize - 1,
((c->vChrFilterPos[chrI] + c->vChrFilterSize - 1)<<c->chrSrcVSubSample));
@@ -1035,9 +1049,9 @@ int sws_init_context(SwsContext *c, SwsFilter *srcFilter, SwsFilter *dstFilter)
FF_ALLOCZ_OR_GOTO(c, c->lumPixBuf[i+c->vLumBufSize], dst_stride+16, fail);
c->lumPixBuf[i] = c->lumPixBuf[i+c->vLumBufSize];
}
- // 64 / (c->dstBpc & ~7) is the same as 16 / sizeof(scaling_intermediate)
- c->uv_off_px = dst_stride_px + 64 / (c->dstBpc &~ 7);
- c->uv_off_byte = dst_stride + 16;
+ // 64 / c->scalingBpp is the same as 16 / sizeof(scaling_intermediate)
+ c->uv_off = (dst_stride>>1) + 64 / (c->dstBpc &~ 7);
+ c->uv_offx2 = dst_stride + 16;
for (i=0; i<c->vChrBufSize; i++) {
FF_ALLOC_OR_GOTO(c, c->chrUPixBuf[i+c->vChrBufSize], dst_stride*2+32, fail);
c->chrUPixBuf[i] = c->chrUPixBuf[i+c->vChrBufSize];
@@ -1051,7 +1065,13 @@ int sws_init_context(SwsContext *c, SwsFilter *srcFilter, SwsFilter *dstFilter)
//try to avoid drawing green stuff between the right end and the stride end
for (i=0; i<c->vChrBufSize; i++)
- memset(c->chrUPixBuf[i], 64, dst_stride*2+1);
+ if(av_pix_fmt_descriptors[c->dstFormat].comp[0].depth_minus1 == 15){
+ av_assert0(c->dstBpc > 10);
+ for(j=0; j<dst_stride/2+1; j++)
+ ((int32_t*)(c->chrUPixBuf[i]))[j] = 1<<18;
+ } else
+ for(j=0; j<dst_stride+1; j++)
+ ((int16_t*)(c->chrUPixBuf[i]))[j] = 1<<14;
assert(c->chrDstH <= dstH);
@@ -1070,7 +1090,7 @@ int sws_init_context(SwsContext *c, SwsFilter *srcFilter, SwsFilter *dstFilter)
else av_log(c, AV_LOG_INFO, "ehh flags invalid?! ");
av_log(c, AV_LOG_INFO, "from %s to %s%s ",
- sws_format_name(srcFormat),
+ av_get_pix_fmt_name(srcFormat),
#ifdef DITHER1XBPP
dstFormat == PIX_FMT_BGR555 || dstFormat == PIX_FMT_BGR565 ||
dstFormat == PIX_FMT_RGB444BE || dstFormat == PIX_FMT_RGB444LE ||
@@ -1078,7 +1098,7 @@ int sws_init_context(SwsContext *c, SwsFilter *srcFilter, SwsFilter *dstFilter)
#else
"",
#endif
- sws_format_name(dstFormat));
+ av_get_pix_fmt_name(dstFormat));
if (HAVE_MMX2 && cpu_flags & AV_CPU_FLAG_MMX2) av_log(c, AV_LOG_INFO, "using MMX2\n");
else if (HAVE_AMD3DNOW && cpu_flags & AV_CPU_FLAG_3DNOW) av_log(c, AV_LOG_INFO, "using 3DNOW\n");
@@ -1479,7 +1499,7 @@ void sws_freeContext(SwsContext *c)
#endif /* HAVE_MMX */
av_freep(&c->yuvTable);
- av_free(c->formatConvBuffer);
+ av_freep(&c->formatConvBuffer);
av_free(c);
}
diff --git a/libswscale/x86/rgb2rgb.c b/libswscale/x86/rgb2rgb.c
index 282618c301..9359f0b1f3 100644
--- a/libswscale/x86/rgb2rgb.c
+++ b/libswscale/x86/rgb2rgb.c
@@ -6,20 +6,20 @@
* 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
*/
@@ -68,6 +68,9 @@ 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;
+DECLARE_ASM_CONST(8, uint64_t, mul15_mid) = 0x4200420042004200ULL;
+DECLARE_ASM_CONST(8, uint64_t, mul15_hi) = 0x0210021002100210ULL;
+DECLARE_ASM_CONST(8, uint64_t, mul16_mid) = 0x2080208020802080ULL;
#define RGB2YUV_SHIFT 8
#define BY ((int)( 0.098*(1<<RGB2YUV_SHIFT)+0.5))
diff --git a/libswscale/x86/rgb2rgb_template.c b/libswscale/x86/rgb2rgb_template.c
index c255610193..3bca43c42e 100644
--- a/libswscale/x86/rgb2rgb_template.c
+++ b/libswscale/x86/rgb2rgb_template.c
@@ -7,20 +7,20 @@
* palette & YUV & runtime CPU stuff by Michael (michaelni@gmx.at)
* lot of big-endian byte order fixes by Alex Beregszaszi
*
- * 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
*/
@@ -801,27 +801,6 @@ static inline void RENAME(rgb24to15)(const uint8_t *src, uint8_t *dst, int src_s
}
}
-/*
- I use less accurate approximation here by simply left-shifting the input
- value and filling the low order bits with zeroes. This method improves PNG
- compression but this scheme cannot reproduce white exactly, since it does
- not generate an all-ones maximum value; the net effect is to darken the
- image slightly.
-
- The better method should be "left bit replication":
-
- 4 3 2 1 0
- ---------
- 1 1 0 1 1
-
- 7 6 5 4 3 2 1 0
- ----------------
- 1 1 0 1 1 1 1 0
- |=======| |===|
- | leftmost bits repeated to fill open bits
- |
- original bits
-*/
static inline void RENAME(rgb15tobgr24)(const uint8_t *src, uint8_t *dst, int src_size)
{
const uint16_t *end;
@@ -840,9 +819,10 @@ static inline void RENAME(rgb15tobgr24)(const uint8_t *src, uint8_t *dst, int sr
"pand %2, %%mm0 \n\t"
"pand %3, %%mm1 \n\t"
"pand %4, %%mm2 \n\t"
- "psllq $3, %%mm0 \n\t"
- "psrlq $2, %%mm1 \n\t"
- "psrlq $7, %%mm2 \n\t"
+ "psllq $5, %%mm0 \n\t"
+ "pmulhw %6, %%mm0 \n\t"
+ "pmulhw %6, %%mm1 \n\t"
+ "pmulhw %7, %%mm2 \n\t"
"movq %%mm0, %%mm3 \n\t"
"movq %%mm1, %%mm4 \n\t"
"movq %%mm2, %%mm5 \n\t"
@@ -870,9 +850,10 @@ static inline void RENAME(rgb15tobgr24)(const uint8_t *src, uint8_t *dst, int sr
"pand %2, %%mm0 \n\t"
"pand %3, %%mm1 \n\t"
"pand %4, %%mm2 \n\t"
- "psllq $3, %%mm0 \n\t"
- "psrlq $2, %%mm1 \n\t"
- "psrlq $7, %%mm2 \n\t"
+ "psllq $5, %%mm0 \n\t"
+ "pmulhw %6, %%mm0 \n\t"
+ "pmulhw %6, %%mm1 \n\t"
+ "pmulhw %7, %%mm2 \n\t"
"movq %%mm0, %%mm3 \n\t"
"movq %%mm1, %%mm4 \n\t"
"movq %%mm2, %%mm5 \n\t"
@@ -892,7 +873,7 @@ static inline void RENAME(rgb15tobgr24)(const uint8_t *src, uint8_t *dst, int sr
"por %%mm5, %%mm3 \n\t"
:"=m"(*d)
- :"m"(*s),"m"(mask15b),"m"(mask15g),"m"(mask15r), "m"(mmx_null)
+ :"m"(*s),"m"(mask15b),"m"(mask15g),"m"(mask15r),"m"(mmx_null),"m"(mul15_mid),"m"(mul15_hi)
:"memory");
/* borrowed 32 to 24 */
__asm__ volatile(
@@ -919,9 +900,9 @@ static inline void RENAME(rgb15tobgr24)(const uint8_t *src, uint8_t *dst, int sr
while (s < end) {
register uint16_t bgr;
bgr = *s++;
- *d++ = (bgr&0x1F)<<3;
- *d++ = (bgr&0x3E0)>>2;
- *d++ = (bgr&0x7C00)>>7;
+ *d++ = ((bgr&0x1F)<<3) | ((bgr&0x1F)>>2);
+ *d++ = ((bgr&0x3E0)>>2) | ((bgr&0x3E0)>>7);
+ *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
}
}
@@ -943,9 +924,11 @@ static inline void RENAME(rgb16tobgr24)(const uint8_t *src, uint8_t *dst, int sr
"pand %2, %%mm0 \n\t"
"pand %3, %%mm1 \n\t"
"pand %4, %%mm2 \n\t"
- "psllq $3, %%mm0 \n\t"
- "psrlq $3, %%mm1 \n\t"
- "psrlq $8, %%mm2 \n\t"
+ "psllq $5, %%mm0 \n\t"
+ "psrlq $1, %%mm2 \n\t"
+ "pmulhw %6, %%mm0 \n\t"
+ "pmulhw %8, %%mm1 \n\t"
+ "pmulhw %7, %%mm2 \n\t"
"movq %%mm0, %%mm3 \n\t"
"movq %%mm1, %%mm4 \n\t"
"movq %%mm2, %%mm5 \n\t"
@@ -973,9 +956,11 @@ static inline void RENAME(rgb16tobgr24)(const uint8_t *src, uint8_t *dst, int sr
"pand %2, %%mm0 \n\t"
"pand %3, %%mm1 \n\t"
"pand %4, %%mm2 \n\t"
- "psllq $3, %%mm0 \n\t"
- "psrlq $3, %%mm1 \n\t"
- "psrlq $8, %%mm2 \n\t"
+ "psllq $5, %%mm0 \n\t"
+ "psrlq $1, %%mm2 \n\t"
+ "pmulhw %6, %%mm0 \n\t"
+ "pmulhw %8, %%mm1 \n\t"
+ "pmulhw %7, %%mm2 \n\t"
"movq %%mm0, %%mm3 \n\t"
"movq %%mm1, %%mm4 \n\t"
"movq %%mm2, %%mm5 \n\t"
@@ -994,7 +979,7 @@ static inline void RENAME(rgb16tobgr24)(const uint8_t *src, uint8_t *dst, int sr
"por %%mm4, %%mm3 \n\t"
"por %%mm5, %%mm3 \n\t"
:"=m"(*d)
- :"m"(*s),"m"(mask16b),"m"(mask16g),"m"(mask16r),"m"(mmx_null)
+ :"m"(*s),"m"(mask16b),"m"(mask16g),"m"(mask16r),"m"(mmx_null),"m"(mul15_mid),"m"(mul15_hi),"m"(mul16_mid)
:"memory");
/* borrowed 32 to 24 */
__asm__ volatile(
@@ -1021,9 +1006,9 @@ static inline void RENAME(rgb16tobgr24)(const uint8_t *src, uint8_t *dst, int sr
while (s < end) {
register uint16_t bgr;
bgr = *s++;
- *d++ = (bgr&0x1F)<<3;
- *d++ = (bgr&0x7E0)>>3;
- *d++ = (bgr&0xF800)>>8;
+ *d++ = ((bgr&0x1F)<<3) | ((bgr&0x1F)>>2);
+ *d++ = ((bgr&0x7E0)>>3) | ((bgr&0x7E0)>>9);
+ *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
}
}
@@ -1066,12 +1051,13 @@ static inline void RENAME(rgb15to32)(const uint8_t *src, uint8_t *dst, int src_s
"pand %2, %%mm0 \n\t"
"pand %3, %%mm1 \n\t"
"pand %4, %%mm2 \n\t"
- "psllq $3, %%mm0 \n\t"
- "psrlq $2, %%mm1 \n\t"
- "psrlq $7, %%mm2 \n\t"
+ "psllq $5, %%mm0 \n\t"
+ "pmulhw %5, %%mm0 \n\t"
+ "pmulhw %5, %%mm1 \n\t"
+ "pmulhw %6, %%mm2 \n\t"
PACK_RGB32
:"=m"(*d)
- :"m"(*s),"m"(mask15b),"m"(mask15g),"m"(mask15r)
+ :"m"(*s),"m"(mask15b),"m"(mask15g),"m"(mask15r),"m"(mul15_mid),"m"(mul15_hi)
:"memory");
d += 16;
s += 4;
@@ -1081,9 +1067,9 @@ static inline void RENAME(rgb15to32)(const uint8_t *src, uint8_t *dst, int src_s
while (s < end) {
register uint16_t bgr;
bgr = *s++;
- *d++ = (bgr&0x1F)<<3;
- *d++ = (bgr&0x3E0)>>2;
- *d++ = (bgr&0x7C00)>>7;
+ *d++ = ((bgr&0x1F)<<3) | ((bgr&0x1F)>>2);
+ *d++ = ((bgr&0x3E0)>>2) | ((bgr&0x3E0)>>7);
+ *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
*d++ = 255;
}
}
@@ -1108,12 +1094,14 @@ static inline void RENAME(rgb16to32)(const uint8_t *src, uint8_t *dst, int src_s
"pand %2, %%mm0 \n\t"
"pand %3, %%mm1 \n\t"
"pand %4, %%mm2 \n\t"
- "psllq $3, %%mm0 \n\t"
- "psrlq $3, %%mm1 \n\t"
- "psrlq $8, %%mm2 \n\t"
+ "psllq $5, %%mm0 \n\t"
+ "psrlq $1, %%mm2 \n\t"
+ "pmulhw %5, %%mm0 \n\t"
+ "pmulhw %7, %%mm1 \n\t"
+ "pmulhw %6, %%mm2 \n\t"
PACK_RGB32
:"=m"(*d)
- :"m"(*s),"m"(mask16b),"m"(mask16g),"m"(mask16r)
+ :"m"(*s),"m"(mask16b),"m"(mask16g),"m"(mask16r),"m"(mul15_mid),"m"(mul15_hi),"m"(mul16_mid)
:"memory");
d += 16;
s += 4;
@@ -1123,9 +1111,9 @@ static inline void RENAME(rgb16to32)(const uint8_t *src, uint8_t *dst, int src_s
while (s < end) {
register uint16_t bgr;
bgr = *s++;
- *d++ = (bgr&0x1F)<<3;
- *d++ = (bgr&0x7E0)>>3;
- *d++ = (bgr&0xF800)>>8;
+ *d++ = ((bgr&0x1F)<<3) | ((bgr&0x1F)>>2);
+ *d++ = ((bgr&0x7E0)>>3) | ((bgr&0x7E0)>>9);
+ *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
*d++ = 255;
}
}
diff --git a/libswscale/x86/scale.asm b/libswscale/x86/scale.asm
index 3bdf71b542..e8a5e5cd7c 100644
--- a/libswscale/x86/scale.asm
+++ b/libswscale/x86/scale.asm
@@ -423,11 +423,13 @@ SCALE_FUNC %1, %2, X, X8, %3, 7, %4
SCALE_FUNCS 8, 15, %1, %2
SCALE_FUNCS 9, 15, %1, %3
SCALE_FUNCS 10, 15, %1, %3
+SCALE_FUNCS 14, 15, %1, %3
SCALE_FUNCS 16, 15, %1, %4
%endif ; !sse4
SCALE_FUNCS 8, 19, %1, %2
SCALE_FUNCS 9, 19, %1, %3
SCALE_FUNCS 10, 19, %1, %3
+SCALE_FUNCS 14, 19, %1, %3
SCALE_FUNCS 16, 19, %1, %4
%endmacro
diff --git a/libswscale/x86/swscale_mmx.c b/libswscale/x86/swscale_mmx.c
index b3f1deae60..360ec2a1c7 100644
--- a/libswscale/x86/swscale_mmx.c
+++ b/libswscale/x86/swscale_mmx.c
@@ -1,20 +1,20 @@
/*
- * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>
+ * Copyright (C) 2001-2011 Michael Niedermayer <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
*/
@@ -72,14 +72,14 @@ DECLARE_ASM_CONST(8, uint64_t, ff_bgr24toY1Coeff) = 0x0C88000040870C88ULL;
DECLARE_ASM_CONST(8, uint64_t, ff_bgr24toY2Coeff) = 0x20DE4087000020DEULL;
DECLARE_ASM_CONST(8, uint64_t, ff_rgb24toY1Coeff) = 0x20DE0000408720DEULL;
DECLARE_ASM_CONST(8, uint64_t, ff_rgb24toY2Coeff) = 0x0C88408700000C88ULL;
-DECLARE_ASM_CONST(8, uint64_t, ff_bgr24toYOffset) = 0x0008400000084000ULL;
+DECLARE_ASM_CONST(8, uint64_t, ff_bgr24toYOffset) = 0x0008010000080100ULL;
DECLARE_ASM_CONST(8, uint64_t, ff_bgr24toUV)[2][4] = {
{0x38380000DAC83838ULL, 0xECFFDAC80000ECFFULL, 0xF6E40000D0E3F6E4ULL, 0x3838D0E300003838ULL},
{0xECFF0000DAC8ECFFULL, 0x3838DAC800003838ULL, 0x38380000D0E33838ULL, 0xF6E4D0E30000F6E4ULL},
};
-DECLARE_ASM_CONST(8, uint64_t, ff_bgr24toUVOffset)= 0x0040400000404000ULL;
+DECLARE_ASM_CONST(8, uint64_t, ff_bgr24toUVOffset)= 0x0040010000400100ULL;
//MMX versions
#if HAVE_MMX
@@ -176,6 +176,67 @@ void updateMMXDitherTables(SwsContext *c, int dstY, int lumBufIndex, int chrBufI
}
}
+#if HAVE_MMX2
+static void yuv2yuvX_sse3(const int16_t *filter, int filterSize,
+ const int16_t **src, uint8_t *dest, int dstW,
+ const uint8_t *dither, int offset)
+{
+ if(((int)dest) & 15){
+ return yuv2yuvX_MMX2(filter, filterSize, src, dest, dstW, dither, offset);
+ }
+ if (offset) {
+ __asm__ volatile("movq (%0), %%xmm3\n\t"
+ "movdqa %%xmm3, %%xmm4\n\t"
+ "psrlq $24, %%xmm3\n\t"
+ "psllq $40, %%xmm4\n\t"
+ "por %%xmm4, %%xmm3\n\t"
+ :: "r"(dither)
+ );
+ } else {
+ __asm__ volatile("movq (%0), %%xmm3\n\t"
+ :: "r"(dither)
+ );
+ }
+ __asm__ volatile(
+ "pxor %%xmm0, %%xmm0\n\t"
+ "punpcklbw %%xmm0, %%xmm3\n\t"
+ "psraw $4, %%xmm3\n\t"
+ "movdqa %%xmm3, %%xmm4\n\t"
+ "movdqa %%xmm3, %%xmm7\n\t"
+ "movl %3, %%ecx\n\t"
+ "mov %0, %%"REG_d" \n\t"\
+ "mov (%%"REG_d"), %%"REG_S" \n\t"\
+ ".p2align 4 \n\t" /* FIXME Unroll? */\
+ "1: \n\t"\
+ "movddup 8(%%"REG_d"), %%xmm0 \n\t" /* filterCoeff */\
+ "movdqa (%%"REG_S", %%"REG_c", 2), %%xmm2 \n\t" /* srcData */\
+ "movdqa 16(%%"REG_S", %%"REG_c", 2), %%xmm5 \n\t" /* srcData */\
+ "add $16, %%"REG_d" \n\t"\
+ "mov (%%"REG_d"), %%"REG_S" \n\t"\
+ "test %%"REG_S", %%"REG_S" \n\t"\
+ "pmulhw %%xmm0, %%xmm2 \n\t"\
+ "pmulhw %%xmm0, %%xmm5 \n\t"\
+ "paddw %%xmm2, %%xmm3 \n\t"\
+ "paddw %%xmm5, %%xmm4 \n\t"\
+ " jnz 1b \n\t"\
+ "psraw $3, %%xmm3 \n\t"\
+ "psraw $3, %%xmm4 \n\t"\
+ "packuswb %%xmm4, %%xmm3 \n\t"
+ "movntdq %%xmm3, (%1, %%"REG_c")\n\t"
+ "add $16, %%"REG_c" \n\t"\
+ "cmp %2, %%"REG_c" \n\t"\
+ "movdqa %%xmm7, %%xmm3\n\t"
+ "movdqa %%xmm7, %%xmm4\n\t"
+ "mov %0, %%"REG_d" \n\t"\
+ "mov (%%"REG_d"), %%"REG_S" \n\t"\
+ "jb 1b \n\t"\
+ :: "g" (filter),
+ "r" (dest-offset), "g" ((x86_reg)(dstW+offset)), "m" (offset)
+ : "%"REG_d, "%"REG_S, "%"REG_c
+ );
+}
+#endif
+
#define SCALE_FUNC(filter_n, from_bpc, to_bpc, opt) \
extern void ff_hscale ## from_bpc ## to ## to_bpc ## _ ## filter_n ## _ ## opt( \
SwsContext *c, int16_t *data, \
@@ -187,10 +248,12 @@ extern void ff_hscale ## from_bpc ## to ## to_bpc ## _ ## filter_n ## _ ## opt(
SCALE_FUNC(filter_n, 8, 15, opt); \
SCALE_FUNC(filter_n, 9, 15, opt); \
SCALE_FUNC(filter_n, 10, 15, opt); \
+ SCALE_FUNC(filter_n, 14, 15, opt); \
SCALE_FUNC(filter_n, 16, 15, opt); \
SCALE_FUNC(filter_n, 8, 19, opt); \
SCALE_FUNC(filter_n, 9, 19, opt); \
SCALE_FUNC(filter_n, 10, 19, opt); \
+ SCALE_FUNC(filter_n, 14, 19, opt); \
SCALE_FUNC(filter_n, 16, 19, opt)
#define SCALE_FUNCS_MMX(opt) \
@@ -253,6 +316,10 @@ void ff_sws_init_swScale_mmx(SwsContext *c)
#if HAVE_MMX2
if (cpu_flags & AV_CPU_FLAG_MMX2)
sws_init_swScale_MMX2(c);
+ if (cpu_flags & AV_CPU_FLAG_SSE3){
+ if(c->use_mmx_vfilter && !(c->flags & SWS_ACCURATE_RND))
+ c->yuv2planeX = yuv2yuvX_sse3;
+ }
#endif
#if HAVE_YASM
@@ -266,7 +333,10 @@ void ff_sws_init_swScale_mmx(SwsContext *c)
} else if (c->srcBpc == 10) { \
hscalefn = c->dstBpc <= 10 ? ff_hscale10to15_ ## filtersize ## _ ## opt2 : \
ff_hscale10to19_ ## filtersize ## _ ## opt1; \
- } else /* c->srcBpc == 16 */ { \
+ } else if (c->srcBpc == 14 || ((c->srcFormat==PIX_FMT_PAL8||isAnyRGB(c->srcFormat)) && av_pix_fmt_descriptors[c->srcFormat].comp[0].depth_minus1<15)) { \
+ hscalefn = c->dstBpc <= 10 ? ff_hscale14to15_ ## filtersize ## _ ## opt2 : \
+ ff_hscale14to19_ ## filtersize ## _ ## opt1; \
+ } else { /* c->srcBpc == 16 */ \
hscalefn = c->dstBpc <= 10 ? ff_hscale16to15_ ## filtersize ## _ ## opt2 : \
ff_hscale16to19_ ## filtersize ## _ ## opt1; \
} \
@@ -279,10 +349,10 @@ void ff_sws_init_swScale_mmx(SwsContext *c)
}
#define ASSIGN_VSCALEX_FUNC(vscalefn, opt1, opt2, opt2chk, do_16_case) \
switch(c->dstBpc){ \
- case 16: do_16_case; break; \
- case 10: if (!isBE(c->dstFormat) && opt2chk) vscalefn = ff_yuv2planeX_10_ ## opt2; break; \
- case 9: if (!isBE(c->dstFormat) && opt2chk) vscalefn = ff_yuv2planeX_9_ ## opt2; break; \
- default: vscalefn = ff_yuv2planeX_8_ ## opt1; break; \
+ case 16: /*do_16_case;*/ break; \
+ case 10: if (!isBE(c->dstFormat) && opt2chk) /*vscalefn = ff_yuv2planeX_10_ ## opt2;*/ break; \
+ case 9: if (!isBE(c->dstFormat) && opt2chk) /*vscalefn = ff_yuv2planeX_9_ ## opt2;*/ break; \
+ default: /*vscalefn = ff_yuv2planeX_8_ ## opt1;*/ break; \
}
#define ASSIGN_VSCALE_FUNC(vscalefn, opt1, opt2, opt2chk) \
switch(c->dstBpc){ \
diff --git a/libswscale/x86/swscale_template.c b/libswscale/x86/swscale_template.c
index 5e7df5c4a0..c7de145b52 100644
--- a/libswscale/x86/swscale_template.c
+++ b/libswscale/x86/swscale_template.c
@@ -1,25 +1,26 @@
/*
- * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>
+ * Copyright (C) 2001-2011 Michael Niedermayer <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
*/
#undef REAL_MOVNTQ
#undef MOVNTQ
+#undef MOVNTQ2
#undef PREFETCH
#if COMPILE_TEMPLATE_MMX2
@@ -30,11 +31,84 @@
#if COMPILE_TEMPLATE_MMX2
#define REAL_MOVNTQ(a,b) "movntq " #a ", " #b " \n\t"
+#define MOVNTQ2 "movntq "
#else
#define REAL_MOVNTQ(a,b) "movq " #a ", " #b " \n\t"
+#define MOVNTQ2 "movq "
#endif
#define MOVNTQ(a,b) REAL_MOVNTQ(a,b)
+#if !COMPILE_TEMPLATE_MMX2
+static av_always_inline void
+dither_8to16(const uint8_t *srcDither, int rot)
+{
+ if (rot) {
+ __asm__ volatile("pxor %%mm0, %%mm0\n\t"
+ "movq (%0), %%mm3\n\t"
+ "movq %%mm3, %%mm4\n\t"
+ "psrlq $24, %%mm3\n\t"
+ "psllq $40, %%mm4\n\t"
+ "por %%mm4, %%mm3\n\t"
+ "movq %%mm3, %%mm4\n\t"
+ "punpcklbw %%mm0, %%mm3\n\t"
+ "punpckhbw %%mm0, %%mm4\n\t"
+ :: "r"(srcDither)
+ );
+ } else {
+ __asm__ volatile("pxor %%mm0, %%mm0\n\t"
+ "movq (%0), %%mm3\n\t"
+ "movq %%mm3, %%mm4\n\t"
+ "punpcklbw %%mm0, %%mm3\n\t"
+ "punpckhbw %%mm0, %%mm4\n\t"
+ :: "r"(srcDither)
+ );
+ }
+}
+#endif
+
+static void RENAME(yuv2yuvX)(const int16_t *filter, int filterSize,
+ const int16_t **src, uint8_t *dest, int dstW,
+ const uint8_t *dither, int offset)
+{
+ dither_8to16(dither, offset);
+ __asm__ volatile(\
+ "psraw $4, %%mm3\n\t"
+ "psraw $4, %%mm4\n\t"
+ "movq %%mm3, %%mm6\n\t"
+ "movq %%mm4, %%mm7\n\t"
+ "movl %3, %%ecx\n\t"
+ "mov %0, %%"REG_d" \n\t"\
+ "mov (%%"REG_d"), %%"REG_S" \n\t"\
+ ".p2align 4 \n\t" /* FIXME Unroll? */\
+ "1: \n\t"\
+ "movq 8(%%"REG_d"), %%mm0 \n\t" /* filterCoeff */\
+ "movq (%%"REG_S", %%"REG_c", 2), %%mm2 \n\t" /* srcData */\
+ "movq 8(%%"REG_S", %%"REG_c", 2), %%mm5 \n\t" /* srcData */\
+ "add $16, %%"REG_d" \n\t"\
+ "mov (%%"REG_d"), %%"REG_S" \n\t"\
+ "test %%"REG_S", %%"REG_S" \n\t"\
+ "pmulhw %%mm0, %%mm2 \n\t"\
+ "pmulhw %%mm0, %%mm5 \n\t"\
+ "paddw %%mm2, %%mm3 \n\t"\
+ "paddw %%mm5, %%mm4 \n\t"\
+ " jnz 1b \n\t"\
+ "psraw $3, %%mm3 \n\t"\
+ "psraw $3, %%mm4 \n\t"\
+ "packuswb %%mm4, %%mm3 \n\t"
+ MOVNTQ2 " %%mm3, (%1, %%"REG_c")\n\t"
+ "add $8, %%"REG_c" \n\t"\
+ "cmp %2, %%"REG_c" \n\t"\
+ "movq %%mm6, %%mm3\n\t"
+ "movq %%mm7, %%mm4\n\t"
+ "mov %0, %%"REG_d" \n\t"\
+ "mov (%%"REG_d"), %%"REG_S" \n\t"\
+ "jb 1b \n\t"\
+ :: "g" (filter),
+ "r" (dest-offset), "g" ((x86_reg)(dstW+offset)), "m" (offset)
+ : "%"REG_d, "%"REG_S, "%"REG_c
+ );
+}
+
#define YSCALEYUV2PACKEDX_UV \
__asm__ volatile(\
"xor %%"REG_a", %%"REG_a" \n\t"\
@@ -260,7 +334,7 @@ static void RENAME(yuv2rgb32_X_ar)(SwsContext *c, const int16_t *lumFilter,
{
x86_reg dummy=0;
x86_reg dstW_reg = dstW;
- x86_reg uv_off = c->uv_off_byte;
+ x86_reg uv_off = c->uv_offx2;
if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {
YSCALEYUV2PACKEDX_ACCURATE
@@ -293,7 +367,7 @@ static void RENAME(yuv2rgb32_X)(SwsContext *c, const int16_t *lumFilter,
{
x86_reg dummy=0;
x86_reg dstW_reg = dstW;
- x86_reg uv_off = c->uv_off_byte;
+ x86_reg uv_off = c->uv_offx2;
if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {
YSCALEYUV2PACKEDX
@@ -350,7 +424,7 @@ static void RENAME(yuv2rgb565_X_ar)(SwsContext *c, const int16_t *lumFilter,
{
x86_reg dummy=0;
x86_reg dstW_reg = dstW;
- x86_reg uv_off = c->uv_off_byte;
+ x86_reg uv_off = c->uv_offx2;
YSCALEYUV2PACKEDX_ACCURATE
YSCALEYUV2RGBX
@@ -374,7 +448,7 @@ static void RENAME(yuv2rgb565_X)(SwsContext *c, const int16_t *lumFilter,
{
x86_reg dummy=0;
x86_reg dstW_reg = dstW;
- x86_reg uv_off = c->uv_off_byte;
+ x86_reg uv_off = c->uv_offx2;
YSCALEYUV2PACKEDX
YSCALEYUV2RGBX
@@ -427,7 +501,7 @@ static void RENAME(yuv2rgb555_X_ar)(SwsContext *c, const int16_t *lumFilter,
{
x86_reg dummy=0;
x86_reg dstW_reg = dstW;
- x86_reg uv_off = c->uv_off_byte;
+ x86_reg uv_off = c->uv_offx2;
YSCALEYUV2PACKEDX_ACCURATE
YSCALEYUV2RGBX
@@ -451,7 +525,7 @@ static void RENAME(yuv2rgb555_X)(SwsContext *c, const int16_t *lumFilter,
{
x86_reg dummy=0;
x86_reg dstW_reg = dstW;
- x86_reg uv_off = c->uv_off_byte;
+ x86_reg uv_off = c->uv_offx2;
YSCALEYUV2PACKEDX
YSCALEYUV2RGBX
@@ -584,7 +658,7 @@ static void RENAME(yuv2bgr24_X_ar)(SwsContext *c, const int16_t *lumFilter,
{
x86_reg dummy=0;
x86_reg dstW_reg = dstW;
- x86_reg uv_off = c->uv_off_byte;
+ x86_reg uv_off = c->uv_offx2;
YSCALEYUV2PACKEDX_ACCURATE
YSCALEYUV2RGBX
@@ -608,7 +682,7 @@ static void RENAME(yuv2bgr24_X)(SwsContext *c, const int16_t *lumFilter,
{
x86_reg dummy=0;
x86_reg dstW_reg = dstW;
- x86_reg uv_off = c->uv_off_byte;
+ x86_reg uv_off = c->uv_offx2;
YSCALEYUV2PACKEDX
YSCALEYUV2RGBX
@@ -649,7 +723,7 @@ static void RENAME(yuv2yuyv422_X_ar)(SwsContext *c, const int16_t *lumFilter,
{
x86_reg dummy=0;
x86_reg dstW_reg = dstW;
- x86_reg uv_off = c->uv_off_byte;
+ x86_reg uv_off = c->uv_offx2;
YSCALEYUV2PACKEDX_ACCURATE
/* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */
@@ -670,7 +744,7 @@ static void RENAME(yuv2yuyv422_X)(SwsContext *c, const int16_t *lumFilter,
{
x86_reg dummy=0;
x86_reg dstW_reg = dstW;
- x86_reg uv_off = c->uv_off_byte;
+ x86_reg uv_off = c->uv_offx2;
YSCALEYUV2PACKEDX
/* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */
@@ -688,10 +762,10 @@ static void RENAME(yuv2yuyv422_X)(SwsContext *c, const int16_t *lumFilter,
"1: \n\t"\
"movq (%2, "#index"), %%mm2 \n\t" /* uvbuf0[eax]*/\
"movq (%3, "#index"), %%mm3 \n\t" /* uvbuf1[eax]*/\
- "add "UV_OFF_PX"("#c"), "#index" \n\t" \
+ "add "UV_OFFx2"("#c"), "#index" \n\t" \
"movq (%2, "#index"), %%mm5 \n\t" /* uvbuf0[eax+2048]*/\
"movq (%3, "#index"), %%mm4 \n\t" /* uvbuf1[eax+2048]*/\
- "sub "UV_OFF_PX"("#c"), "#index" \n\t" \
+ "sub "UV_OFFx2"("#c"), "#index" \n\t" \
"psubw %%mm3, %%mm2 \n\t" /* uvbuf0[eax] - uvbuf1[eax]*/\
"psubw %%mm4, %%mm5 \n\t" /* uvbuf0[eax+2048] - uvbuf1[eax+2048]*/\
"movq "CHR_MMX_FILTER_OFFSET"+8("#c"), %%mm0 \n\t"\
@@ -786,8 +860,8 @@ static void RENAME(yuv2rgb32_2)(SwsContext *c, const int16_t *buf[2],
: "%r8"
);
#else
- *(const uint16_t **)(&c->u_temp)=abuf0;
- *(const uint16_t **)(&c->v_temp)=abuf1;
+ c->u_temp=(intptr_t)abuf0;
+ c->v_temp=(intptr_t)abuf1;
__asm__ volatile(
"mov %%"REG_b", "ESP_OFFSET"(%5) \n\t"
"mov %4, %%"REG_b" \n\t"
@@ -919,10 +993,10 @@ static void RENAME(yuv2rgb565_2)(SwsContext *c, const int16_t *buf[2],
"1: \n\t"\
"movq (%2, "#index"), %%mm2 \n\t" /* uvbuf0[eax]*/\
"movq (%3, "#index"), %%mm3 \n\t" /* uvbuf1[eax]*/\
- "add "UV_OFF_PX"("#c"), "#index" \n\t" \
+ "add "UV_OFFx2"("#c"), "#index" \n\t" \
"movq (%2, "#index"), %%mm5 \n\t" /* uvbuf0[eax+2048]*/\
"movq (%3, "#index"), %%mm4 \n\t" /* uvbuf1[eax+2048]*/\
- "sub "UV_OFF_PX"("#c"), "#index" \n\t" \
+ "sub "UV_OFFx2"("#c"), "#index" \n\t" \
"psubw %%mm3, %%mm2 \n\t" /* uvbuf0[eax] - uvbuf1[eax]*/\
"psubw %%mm4, %%mm5 \n\t" /* uvbuf0[eax+2048] - uvbuf1[eax+2048]*/\
"movq "CHR_MMX_FILTER_OFFSET"+8("#c"), %%mm0 \n\t"\
@@ -974,9 +1048,9 @@ static void RENAME(yuv2yuyv422_2)(SwsContext *c, const int16_t *buf[2],
".p2align 4 \n\t"\
"1: \n\t"\
"movq (%2, "#index"), %%mm3 \n\t" /* uvbuf0[eax]*/\
- "add "UV_OFF_PX"("#c"), "#index" \n\t" \
+ "add "UV_OFFx2"("#c"), "#index" \n\t" \
"movq (%2, "#index"), %%mm4 \n\t" /* uvbuf0[eax+2048]*/\
- "sub "UV_OFF_PX"("#c"), "#index" \n\t" \
+ "sub "UV_OFFx2"("#c"), "#index" \n\t" \
"psraw $4, %%mm3 \n\t" /* uvbuf0[eax] - uvbuf1[eax] >>4*/\
"psraw $4, %%mm4 \n\t" /* uvbuf0[eax+2048] - uvbuf1[eax+2048] >>4*/\
"psubw "U_OFFSET"("#c"), %%mm3 \n\t" /* (U-128)8*/\
@@ -1027,10 +1101,10 @@ static void RENAME(yuv2yuyv422_2)(SwsContext *c, const int16_t *buf[2],
"1: \n\t"\
"movq (%2, "#index"), %%mm2 \n\t" /* uvbuf0[eax]*/\
"movq (%3, "#index"), %%mm3 \n\t" /* uvbuf1[eax]*/\
- "add "UV_OFF_PX"("#c"), "#index" \n\t" \
+ "add "UV_OFFx2"("#c"), "#index" \n\t" \
"movq (%2, "#index"), %%mm5 \n\t" /* uvbuf0[eax+2048]*/\
"movq (%3, "#index"), %%mm4 \n\t" /* uvbuf1[eax+2048]*/\
- "sub "UV_OFF_PX"("#c"), "#index" \n\t" \
+ "sub "UV_OFFx2"("#c"), "#index" \n\t" \
"paddw %%mm2, %%mm3 \n\t" /* uvbuf0[eax] + uvbuf1[eax]*/\
"paddw %%mm5, %%mm4 \n\t" /* uvbuf0[eax+2048] + uvbuf1[eax+2048]*/\
"psrlw $5, %%mm3 \n\t" /*FIXME might overflow*/\
@@ -1294,9 +1368,9 @@ static void RENAME(yuv2rgb565_1)(SwsContext *c, const int16_t *buf0,
".p2align 4 \n\t"\
"1: \n\t"\
"movq (%2, "#index"), %%mm3 \n\t" /* uvbuf0[eax]*/\
- "add "UV_OFF_PX"("#c"), "#index" \n\t" \
+ "add "UV_OFFx2"("#c"), "#index" \n\t" \
"movq (%2, "#index"), %%mm4 \n\t" /* uvbuf0[eax+2048]*/\
- "sub "UV_OFF_PX"("#c"), "#index" \n\t" \
+ "sub "UV_OFFx2"("#c"), "#index" \n\t" \
"psraw $7, %%mm3 \n\t" \
"psraw $7, %%mm4 \n\t" \
"movq (%0, "#index", 2), %%mm1 \n\t" /*buf0[eax]*/\
@@ -1312,10 +1386,10 @@ static void RENAME(yuv2rgb565_1)(SwsContext *c, const int16_t *buf0,
"1: \n\t"\
"movq (%2, "#index"), %%mm2 \n\t" /* uvbuf0[eax]*/\
"movq (%3, "#index"), %%mm3 \n\t" /* uvbuf1[eax]*/\
- "add "UV_OFF_PX"("#c"), "#index" \n\t" \
+ "add "UV_OFFx2"("#c"), "#index" \n\t" \
"movq (%2, "#index"), %%mm5 \n\t" /* uvbuf0[eax+2048]*/\
"movq (%3, "#index"), %%mm4 \n\t" /* uvbuf1[eax+2048]*/\
- "sub "UV_OFF_PX"("#c"), "#index" \n\t" \
+ "sub "UV_OFFx2"("#c"), "#index" \n\t" \
"paddw %%mm2, %%mm3 \n\t" /* uvbuf0[eax] + uvbuf1[eax]*/\
"paddw %%mm5, %%mm4 \n\t" /* uvbuf0[eax+2048] + uvbuf1[eax+2048]*/\
"psrlw $8, %%mm3 \n\t" \
@@ -1364,7 +1438,7 @@ static void RENAME(yuv2yuyv422_1)(SwsContext *c, const int16_t *buf0,
#if !COMPILE_TEMPLATE_MMX2
//FIXME yuy2* can read up to 7 samples too much
-static void RENAME(yuy2ToY)(uint8_t *dst, const uint8_t *src,
+static void RENAME(yuy2ToY)(uint8_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2,
int width, uint32_t *unused)
{
__asm__ volatile(
@@ -1385,7 +1459,7 @@ static void RENAME(yuy2ToY)(uint8_t *dst, const uint8_t *src,
}
static void RENAME(yuy2ToUV)(uint8_t *dstU, uint8_t *dstV,
- const uint8_t *src1, const uint8_t *src2,
+ const uint8_t *unused1, const uint8_t *src1, const uint8_t *src2,
int width, uint32_t *unused)
{
__asm__ volatile(
@@ -1414,7 +1488,7 @@ static void RENAME(yuy2ToUV)(uint8_t *dstU, uint8_t *dstV,
/* This is almost identical to the previous, end exists only because
* yuy2ToY/UV)(dst, src+1, ...) would have 100% unaligned accesses. */
-static void RENAME(uyvyToY)(uint8_t *dst, const uint8_t *src,
+static void RENAME(uyvyToY)(uint8_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2,
int width, uint32_t *unused)
{
__asm__ volatile(
@@ -1434,7 +1508,7 @@ static void RENAME(uyvyToY)(uint8_t *dst, const uint8_t *src,
}
static void RENAME(uyvyToUV)(uint8_t *dstU, uint8_t *dstV,
- const uint8_t *src1, const uint8_t *src2,
+ const uint8_t *unused1, const uint8_t *src1, const uint8_t *src2,
int width, uint32_t *unused)
{
__asm__ volatile(
@@ -1488,21 +1562,21 @@ static av_always_inline void RENAME(nvXXtoUV)(uint8_t *dst1, uint8_t *dst2,
}
static void RENAME(nv12ToUV)(uint8_t *dstU, uint8_t *dstV,
- const uint8_t *src1, const uint8_t *src2,
+ const uint8_t *unused1, const uint8_t *src1, const uint8_t *src2,
int width, uint32_t *unused)
{
RENAME(nvXXtoUV)(dstU, dstV, src1, width);
}
static void RENAME(nv21ToUV)(uint8_t *dstU, uint8_t *dstV,
- const uint8_t *src1, const uint8_t *src2,
+ const uint8_t *unused1, const uint8_t *src1, const uint8_t *src2,
int width, uint32_t *unused)
{
RENAME(nvXXtoUV)(dstV, dstU, src1, width);
}
#endif /* !COMPILE_TEMPLATE_MMX2 */
-static av_always_inline void RENAME(bgr24ToY_mmx)(uint8_t *dst, const uint8_t *src,
+static av_always_inline void RENAME(bgr24ToY_mmx)(int16_t *dst, const uint8_t *src,
int width, enum PixelFormat srcFormat)
{
@@ -1543,32 +1617,31 @@ static av_always_inline void RENAME(bgr24ToY_mmx)(uint8_t *dst, const uint8_t *s
"paddd %%mm3, %%mm2 \n\t"
"paddd %%mm4, %%mm0 \n\t"
"paddd %%mm4, %%mm2 \n\t"
- "psrad $15, %%mm0 \n\t"
- "psrad $15, %%mm2 \n\t"
+ "psrad $9, %%mm0 \n\t"
+ "psrad $9, %%mm2 \n\t"
"packssdw %%mm2, %%mm0 \n\t"
- "packuswb %%mm0, %%mm0 \n\t"
- "movd %%mm0, (%1, %%"REG_a") \n\t"
- "add $4, %%"REG_a" \n\t"
+ "movq %%mm0, (%1, %%"REG_a") \n\t"
+ "add $8, %%"REG_a" \n\t"
" js 1b \n\t"
: "+r" (src)
- : "r" (dst+width), "g" ((x86_reg)-width)
+ : "r" (dst+width), "g" ((x86_reg)-2*width)
: "%"REG_a
);
}
-static void RENAME(bgr24ToY)(uint8_t *dst, const uint8_t *src,
+static void RENAME(bgr24ToY)(int16_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2,
int width, uint32_t *unused)
{
RENAME(bgr24ToY_mmx)(dst, src, width, PIX_FMT_BGR24);
}
-static void RENAME(rgb24ToY)(uint8_t *dst, const uint8_t *src,
+static void RENAME(rgb24ToY)(int16_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2,
int width, uint32_t *unused)
{
RENAME(bgr24ToY_mmx)(dst, src, width, PIX_FMT_RGB24);
}
-static av_always_inline void RENAME(bgr24ToUV_mmx)(uint8_t *dstU, uint8_t *dstV,
+static av_always_inline void RENAME(bgr24ToUV_mmx)(int16_t *dstU, int16_t *dstV,
const uint8_t *src, int width,
enum PixelFormat srcFormat)
{
@@ -1610,34 +1683,32 @@ static av_always_inline void RENAME(bgr24ToUV_mmx)(uint8_t *dstU, uint8_t *dstV,
"paddd %%mm3, %%mm2 \n\t"
"paddd %%mm3, %%mm1 \n\t"
"paddd %%mm3, %%mm4 \n\t"
- "psrad $15, %%mm0 \n\t"
- "psrad $15, %%mm2 \n\t"
- "psrad $15, %%mm1 \n\t"
- "psrad $15, %%mm4 \n\t"
+ "psrad $9, %%mm0 \n\t"
+ "psrad $9, %%mm2 \n\t"
+ "psrad $9, %%mm1 \n\t"
+ "psrad $9, %%mm4 \n\t"
"packssdw %%mm1, %%mm0 \n\t"
"packssdw %%mm4, %%mm2 \n\t"
- "packuswb %%mm0, %%mm0 \n\t"
- "packuswb %%mm2, %%mm2 \n\t"
- "movd %%mm0, (%1, %%"REG_a") \n\t"
- "movd %%mm2, (%2, %%"REG_a") \n\t"
- "add $4, %%"REG_a" \n\t"
+ "movq %%mm0, (%1, %%"REG_a") \n\t"
+ "movq %%mm2, (%2, %%"REG_a") \n\t"
+ "add $8, %%"REG_a" \n\t"
" js 1b \n\t"
: "+r" (src)
- : "r" (dstU+width), "r" (dstV+width), "g" ((x86_reg)-width), "r"(ff_bgr24toUV[srcFormat == PIX_FMT_RGB24])
+ : "r" (dstU+width), "r" (dstV+width), "g" ((x86_reg)-2*width), "r"(ff_bgr24toUV[srcFormat == PIX_FMT_RGB24])
: "%"REG_a
);
}
-static void RENAME(bgr24ToUV)(uint8_t *dstU, uint8_t *dstV,
- const uint8_t *src1, const uint8_t *src2,
+static void RENAME(bgr24ToUV)(int16_t *dstU, int16_t *dstV,
+ const uint8_t *unused1, const uint8_t *src1, const uint8_t *src2,
int width, uint32_t *unused)
{
RENAME(bgr24ToUV_mmx)(dstU, dstV, src1, width, PIX_FMT_BGR24);
assert(src1 == src2);
}
-static void RENAME(rgb24ToUV)(uint8_t *dstU, uint8_t *dstV,
- const uint8_t *src1, const uint8_t *src2,
+static void RENAME(rgb24ToUV)(int16_t *dstU, int16_t *dstV,
+ const uint8_t *unused1, const uint8_t *src1, const uint8_t *src2,
int width, uint32_t *unused)
{
assert(src1==src2);
@@ -1784,10 +1855,9 @@ static av_cold void RENAME(sws_init_swScale)(SwsContext *c)
{
enum PixelFormat srcFormat = c->srcFormat,
dstFormat = c->dstFormat;
-
- if (!is16BPS(dstFormat) && !is9_OR_10BPS(dstFormat) &&
- dstFormat != PIX_FMT_NV12 && dstFormat != PIX_FMT_NV21) {
- if (!(c->flags & SWS_BITEXACT)) {
+ c->use_mmx_vfilter= 0;
+ if (!is16BPS(dstFormat) && !is9_OR_10BPS(dstFormat) && dstFormat != PIX_FMT_NV12
+ && dstFormat != PIX_FMT_NV21 && !(c->flags & SWS_BITEXACT)) {
if (c->flags & SWS_ACCURATE_RND) {
if (!(c->flags & SWS_FULL_CHR_H_INT)) {
switch (c->dstFormat) {
@@ -1800,6 +1870,9 @@ static av_cold void RENAME(sws_init_swScale)(SwsContext *c)
}
}
} else {
+ int should_dither= isNBPS(c->srcFormat) || is16BPS(c->srcFormat);
+ c->use_mmx_vfilter= 1;
+ c->yuv2planeX = RENAME(yuv2yuvX );
if (!(c->flags & SWS_FULL_CHR_H_INT)) {
switch (c->dstFormat) {
case PIX_FMT_RGB32: c->yuv2packedX = RENAME(yuv2rgb32_X); break;
@@ -1811,7 +1884,6 @@ static av_cold void RENAME(sws_init_swScale)(SwsContext *c)
}
}
}
- }
if (!(c->flags & SWS_FULL_CHR_H_INT)) {
switch (c->dstFormat) {
case PIX_FMT_RGB32:
diff --git a/libswscale/x86/yuv2rgb_mmx.c b/libswscale/x86/yuv2rgb_mmx.c
index 0eaea77485..df0e1a3726 100644
--- a/libswscale/x86/yuv2rgb_mmx.c
+++ b/libswscale/x86/yuv2rgb_mmx.c
@@ -7,20 +7,20 @@
* 1,4,8bpp support and context / deglobalize stuff
* by Michael Niedermayer (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
*/
diff --git a/libswscale/x86/yuv2rgb_template.c b/libswscale/x86/yuv2rgb_template.c
index 5d1fa5b309..926e3fb9c4 100644
--- a/libswscale/x86/yuv2rgb_template.c
+++ b/libswscale/x86/yuv2rgb_template.c
@@ -4,20 +4,20 @@
* Copyright (C) 2001-2007 Michael Niedermayer
* (c) 2010 Konstantin Shishkov
*
- * 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
*/
@@ -43,17 +43,14 @@
if (h_size * depth > FFABS(dstStride[0])) \
h_size -= 8; \
\
- if (c->srcFormat == PIX_FMT_YUV422P) { \
- srcStride[1] *= 2; \
- srcStride[2] *= 2; \
- } \
+ vshift = c->srcFormat != PIX_FMT_YUV422P; \
\
__asm__ volatile ("pxor %mm4, %mm4\n\t"); \
for (y = 0; y < srcSliceH; y++) { \
uint8_t *image = dst[0] + (y + srcSliceY) * dstStride[0]; \
const uint8_t *py = src[0] + y * 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 *pu = src[1] + (y >> vshift) * srcStride[1]; \
+ const uint8_t *pv = src[2] + (y >> vshift) * srcStride[2]; \
x86_reg index = -h_size / 2; \
#define YUV2RGB_INITIAL_LOAD \
@@ -188,7 +185,7 @@ static inline int RENAME(yuv420_rgb15)(SwsContext *c, const uint8_t *src[],
int srcSliceY, int srcSliceH,
uint8_t *dst[], int dstStride[])
{
- int y, h_size;
+ int y, h_size, vshift;
YUV2RGB_LOOP(2)
@@ -216,7 +213,7 @@ static inline int RENAME(yuv420_rgb16)(SwsContext *c, const uint8_t *src[],
int srcSliceY, int srcSliceH,
uint8_t *dst[], int dstStride[])
{
- int y, h_size;
+ int y, h_size, vshift;
YUV2RGB_LOOP(2)
@@ -306,7 +303,7 @@ static inline int RENAME(yuv420_rgb24)(SwsContext *c, const uint8_t *src[],
int srcSliceY, int srcSliceH,
uint8_t *dst[], int dstStride[])
{
- int y, h_size;
+ int y, h_size, vshift;
YUV2RGB_LOOP(3)
@@ -324,7 +321,7 @@ static inline int RENAME(yuv420_bgr24)(SwsContext *c, const uint8_t *src[],
int srcSliceY, int srcSliceH,
uint8_t *dst[], int dstStride[])
{
- int y, h_size;
+ int y, h_size, vshift;
YUV2RGB_LOOP(3)
@@ -368,7 +365,7 @@ static inline int RENAME(yuv420_rgb32)(SwsContext *c, const uint8_t *src[],
int srcSliceY, int srcSliceH,
uint8_t *dst[], int dstStride[])
{
- int y, h_size;
+ int y, h_size, vshift;
YUV2RGB_LOOP(4)
@@ -389,7 +386,7 @@ static inline int RENAME(yuva420_rgb32)(SwsContext *c, const uint8_t *src[],
int srcSliceY, int srcSliceH,
uint8_t *dst[], int dstStride[])
{
- int y, h_size;
+ int y, h_size, vshift;
YUV2RGB_LOOP(4)
@@ -411,7 +408,7 @@ static inline int RENAME(yuv420_bgr32)(SwsContext *c, const uint8_t *src[],
int srcSliceY, int srcSliceH,
uint8_t *dst[], int dstStride[])
{
- int y, h_size;
+ int y, h_size, vshift;
YUV2RGB_LOOP(4)
@@ -432,7 +429,7 @@ static inline int RENAME(yuva420_bgr32)(SwsContext *c, const uint8_t *src[],
int srcSliceY, int srcSliceH,
uint8_t *dst[], int dstStride[])
{
- int y, h_size;
+ int y, h_size, vshift;
YUV2RGB_LOOP(4)
diff --git a/libswscale/yuv2rgb.c b/libswscale/yuv2rgb.c
index 39c8b9c6fb..c2c9090a36 100644
--- a/libswscale/yuv2rgb.c
+++ b/libswscale/yuv2rgb.c
@@ -6,20 +6,20 @@
* 1,4,8bpp support and context / deglobalize stuff
* by Michael Niedermayer (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
*/
@@ -34,6 +34,7 @@
#include "swscale_internal.h"
#include "libavutil/cpu.h"
#include "libavutil/bswap.h"
+#include "libavutil/pixdesc.h"
extern const uint8_t dither_4x4_16[4][8];
extern const uint8_t dither_8x8_32[8][8];
@@ -521,7 +522,8 @@ SwsFunc ff_yuv2rgb_get_func_ptr(SwsContext *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));
+ av_log(c, AV_LOG_WARNING, "No accelerated colorspace conversion found from %s to %s.\n",
+ av_get_pix_fmt_name(c->srcFormat), av_get_pix_fmt_name(c->dstFormat));
switch (c->dstFormat) {
case PIX_FMT_BGR48BE:
@@ -613,7 +615,7 @@ av_cold int ff_yuv2rgb_c_init_tables(SwsContext *c, const int inv_table[4], int
uint8_t *y_table;
uint16_t *y_table16;
uint32_t *y_table32;
- int i, base, rbase, gbase, bbase, abase, needAlpha;
+ int i, base, rbase, gbase, bbase, av_uninit(abase), needAlpha;
const int yoffs = fullRange ? 384 : 326;
int64_t crv = inv_table[0];