-babl 0.1.14
+babl 0.1.16
Dynamic; any to any, pixel format conversion library.
installation (or a variation on this theme):
------------------------------------------------------------
- foo$ wget ftp://ftp.gtk.org/pub/babl/0.1/babl-0.1.14.tar.bz2
- foo$ tar jxf babl-0.1.14.tar.gz
- foo$ cd babl-0.1.14
- foo/babl-0.1.14$ ./configure && make && sudo make install
+ foo$ wget ftp://ftp.gtk.org/pub/babl/0.1/babl-0.1.16.tar.bz2
+ foo$ tar jxf babl-0.1.16.tar.gz
+ foo$ cd babl-0.1.16
+ foo/babl-0.1.16$ ./configure && make && sudo make install
------------------------------------------------------------
ECHO_T = @ECHO_T@
EGREP = @EGREP@
EXEEXT = @EXEEXT@
+F16C_EXTRA_CFLAGS = @F16C_EXTRA_CFLAGS@
FGREP = @FGREP@
GREP = @GREP@
INSTALL = @INSTALL@
SHELL = @SHELL@
SHREXT = @SHREXT@
SSE2_EXTRA_CFLAGS = @SSE2_EXTRA_CFLAGS@
+SSE4_1_EXTRA_CFLAGS = @SSE4_1_EXTRA_CFLAGS@
SSE_EXTRA_CFLAGS = @SSE_EXTRA_CFLAGS@
STRIP = @STRIP@
THREAD_LIB = @THREAD_LIB@
-->
+2016-02-12 babl-0.1.16 </dt><dd>
+Improvements to half float reference, SIMD, and fast-paths, and CIE float fast
+paths; cleanups of fast path extensions.
+ </dd><dt>
2015-11-19 babl-0.1.14 </dt><dd>
sRGB precision tuning, stability fixes, locale fix for setting error tolerance.
</dd><dt>
-Babl-0.1.14
+Babl-0.1.16
Contents
release is done a babl release is most often put out just prior to the
GEGL release.
+2016-02-12 babl-0.1.16
+ Improvements to half float reference, SIMD, and fast-paths, and CIE
+ float fast paths; cleanups of fast path extensions.
2015-11-19 babl-0.1.14
sRGB precision tuning, stability fixes, locale fix for setting
error tolerance.
saturation
value
+CMYK
+
+components
+
+ cyan
+ magenta
+ yellow
+ key
+
+CMY
+
+components
+
+ cyan
+ magenta
+ yellow
+
CIE Lab
components
CIE H(ab)
A
-CMYK
-
-components
-
- cyan
- magenta
- yellow
- key
-
-CMY
-
-components
-
- cyan
- magenta
- yellow
-
HSLA
components
u8 Y
-B'aG'aR'aA u8
-
-bytes/pixel
- 4
-model
- R'aG'aB'aA
-components
-
- u8 B'a
- u8 G'a
- u8 R'a
- u8 A
-
cairo-RGB24
bytes/pixel
float saturation
float value
+B'aG'aR'aA u8
+
+bytes/pixel
+ 4
+model
+ R'aG'aB'aA
+components
+
+ u8 B'a
+ u8 G'a
+ u8 R'a
+ u8 A
+
+CMYK double
+
+bytes/pixel
+ 32
+model
+ CMYK
+components
+
+ double cyan
+ double magenta
+ double yellow
+ double key
+
+CMY double
+
+bytes/pixel
+ 24
+model
+ CMY
+components
+
+ double cyan
+ double magenta
+ double yellow
+
+CMYK float
+
+bytes/pixel
+ 16
+model
+ CMYK
+components
+
+ float cyan
+ float magenta
+ float yellow
+ float key
+
+CMY float
+
+bytes/pixel
+ 12
+model
+ CMY
+components
+
+ float cyan
+ float magenta
+ float yellow
+
+CMYK u8
+
+bytes/pixel
+ 4
+model
+ CMYK
+components
+
+ u8 cyan
+ u8 magenta
+ u8 yellow
+ u8 key
+
CIE Lab double
bytes/pixel
float CIE H(ab)
float A
-CMYK double
-
-bytes/pixel
- 32
-model
- CMYK
-components
-
- double cyan
- double magenta
- double yellow
- double key
-
-CMY double
-
-bytes/pixel
- 24
-model
- CMY
-components
-
- double cyan
- double magenta
- double yellow
-
-CMYK float
-
-bytes/pixel
- 16
-model
- CMYK
-components
-
- float cyan
- float magenta
- float yellow
- float key
-
-CMY float
-
-bytes/pixel
- 12
-model
- CMY
-components
-
- float cyan
- float magenta
- float yellow
-
-CMYK u8
-
-bytes/pixel
- 4
-model
- CMYK
-components
-
- u8 cyan
- u8 magenta
- u8 yellow
- u8 key
-
HSLA double
bytes/pixel
chain of conversions. A DHTML version is also available.
- ▂·▃▂·▁ ·· · ·· ▁▂▂▃▃▂ ▂▃▂ ▃ ▃ ▁▁▁▂▂ ▂▁▂▁ ▂ ▂ ▃·· ···· ▁▂ ·· ·· ·· ▃ ── 0 RGBA double
+ ▂·▃▂·▁▂▁ ▃▂ ·· ·▂ ▂▃ ▃ ·· ▁▂▂▃▃▂▃▂▃▂▃ ▃▃ ▃ ▁▁▁▂▂▂▂▁▂▁▂▃▂▂▃▂▃·· ·· ····▂▁▂ ·· ·· ▃ ── 0 RGBA double
── 1 RGB double
── 2 RaGaBaA double
── 3 R'G'B' double
-▂ ▃▂▁▁▂ ▃▃ ▃ ▃▃ ·▁▁▂▂▁ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▂▃▂ ▃ ▃ ▂▃▃ ▃▃▃▃ ▂▃ ▃▃ ▃▃ ▃▃ ▃ ── 4 R'G'B'A double
-· ▃▃▁▂ ▁▁ ▁ ▁▁ ▂▃ ▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▂▃▂ ▃ ▃ ▃▁▁ ▁▁▁▁ ▂▃ ▁▁ ▁▁ ▁▁ ── 5 R'aG'aB'aA double
-â\96\81 â\96\81 ··· â\96\82â\96\82 â\96\82 â\96\82â\96\82 ·â\96\81â\96\81â\96\81···â\96\83â\96\83â\96\82 â\96\83 â\96\83 â\96\81â\96\81·â\96\81â\96\82â\96\81â\96\81â\96\81â\96\82â\96\81 â\96\82 â\96\82 ·â\96\82â\96\82 â\96\82â\96\82â\96\83â\96\82 â\96\81â\96\82 â\96\82â\96\82 â\96\82â\96\82 â\96\82▂ ▁ ── 6 R'G'B' u8
-â\96\81 â\96\82â\96\82 ·â\96\81 â\96\82â\96\82 â\96\83 â\96\82â\96\82 ··â\96\81â\96\82â\96\82â\96\81â\96\81â\96\82â\96\83â\96\82 â\96\83 â\96\83 â\96\81â\96\81â\96\81â\96\82â\96\82â\96\81â\96\82â\96\81â\96\82â\96\81 â\96\82 â\96\82 ·â\96\82â\96\82 â\96\82â\96\82â\96\82â\96\82 â\96\81â\96\82 â\96\83â\96\83 â\96\82â\96\83 ▂▂ ▃ ── 7 R'G'B'A u8
-· ▁▁▂ · ▁▁ ▁ ▁▁ ·▁▁▂▂▁ ▁▂▁ ▂▃▂ ···▁▁ ▁·▁· ▁ ▁ ·▁▁ ▁▁▁▁ ·▁ ▁▁ ▁▁ ▁▁ ▃ ── 8 RGBA float
-â\96\81 â\96\82â\96\82â\96\81â\96\82 â\96\82â\96\82 â\96\82 â\96\82â\96\82 â\96\81â\96\82â\96\82â\96\83â\96\82· â\96\82â\96\83â\96\82 â\96\83 â\96\83 â\96\81â\96\81·â\96\81â\96\81 ·â\96\81â\96\82â\96\81 â\96\82 â\96\82 ·â\96\82â\96\82 â\96\83â\96\82â\96\82â\96\82 â\96\81â\96\82 â\96\82â\96\82 â\96\82â\96\82 â\96\82â\96\82 ▃ ── 9 RGB float
- ──10 RGB half
- ──11 RGBA half
+▂ ▃▂▁▁▂▃▂ ▂▁ ▃▃ ▃▃ ▃ ▃▃ ·▁▁▂▂▁▂▃ ▃ ▃ ▃ ▂▂▂▃▃▃▃▂▃▂▃ ▃▃ ▃▂▃▃ ▃▃ ▃▃▃▃▃▂▃ ▃▃ ▃▃ ▃ ── 4 R'G'B'A double
+· ▃▃▁▂▃▂ ▃ ▁▁ ▁▃ ▃ ▁▁ ▂▃ ▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃▃▃▂▃▃▃ ▃▃ ▂▂▁▁ ▁▁ ▁▁▁▁▃▂▃ ▁▁ ▁▁ ── 5 R'aG'aB'aA double
+â\96\81 â\96\82 ···â\96\81â\96\81 â\96\81â\96\82 â\96\82â\96\82 â\96\82â\96\82 â\96\82â\96\83 â\96\83 â\96\82â\96\82 â\96\81â\96\81â\96\81â\96\81···â\96\82â\96\83â\96\82â\96\83 â\96\83â\96\83 â\96\83 â\96\81â\96\81·â\96\81â\96\82â\96\81â\96\81â\96\81â\96\82â\96\81â\96\82â\96\83â\96\82â\96\82â\96\83â\96\82·â\96\82â\96\83 â\96\83â\96\82 â\96\82â\96\82â\96\83â\96\82â\96\83â\96\81â\96\82 â\96\83â\96\82 â\96\83▂ ▁ ── 6 R'G'B' u8
+â\96\81 â\96\81â\96\82 ·â\96\81â\96\82â\96\81 â\96\82â\96\81 â\96\82â\96\82 â\96\82â\96\82 â\96\82â\96\83 â\96\83 â\96\82â\96\82 ··â\96\81â\96\82â\96\82â\96\81â\96\81â\96\82â\96\83â\96\82â\96\83 â\96\83â\96\83 â\96\83 â\96\81â\96\81â\96\81â\96\82â\96\82â\96\82â\96\82â\96\81â\96\82â\96\81â\96\82â\96\83â\96\82â\96\82 â\96\82·â\96\82â\96\82 â\96\82â\96\82 â\96\82â\96\82â\96\82â\96\82â\96\82â\96\81â\96\82 â\96\82â\96\82 ▂▂ ▃ ── 7 R'G'B'A u8
+· ▁▁▂ ·▁· ▂▁ ▁▁ ▁▁ ▁▂ ▂ ▁▁ ·▁▁▂▂▁▂▁▂▁▂▃▂▂▃▂ ···▁▁▁▁·▁·▁▂▁▁▂▁▂▁▁ ▁▁ ▁▁▁▁▁·▁ ▁▁ ▁▁ ▃ ── 8 RGBA float
+â\96\81 â\96\82â\96\83â\96\81â\96\82 ·â\96\81 â\96\81â\96\82 â\96\82â\96\82 â\96\82â\96\82 â\96\82â\96\83 â\96\83 â\96\82â\96\83 â\96\81â\96\82â\96\81â\96\82â\96\82·â\96\81â\96\82â\96\83â\96\82â\96\83 â\96\83â\96\83 â\96\83 â\96\81â\96\81·â\96\81â\96\81··â\96\81â\96\82â\96\81â\96\82â\96\83â\96\82â\96\82â\96\83â\96\82â\96\82â\96\83â\96\82 â\96\82â\96\82 â\96\83â\96\83â\96\82â\96\82·â\96\81â\96\82 â\96\82â\96\82 â\96\82â\96\83 ▃ ── 9 RGB float
+▂ ▃▃▂▃▁ ▂ ▂▃ ▃▃ ▃▃ ▃ ▃▃ ▂▃▂▃▃▁▂▃ ▃ ▃ ▃ ▂▂▁▂▃▁▁▂▃▂▃ ▃▃ ▃▃▃▃ ▃▃ ▃▃▃▃▁▃▃ ▃▃ ▃▃ ▃ ──10 RGB half
+▁ ▂▂▃▂·▁ ▃▂ ▂▂ ▂▂ ▂▃ ▃ ▂▂ ▁▂▂▃▃▂▃▂▃▃▃ ▃▃ ▃ ▁▁▁▂▂▂▂▁▂▁▂▃▂▂▃▂▃▂▂ ▂▂ ▂▂▂▂▂▁▂ ▂▂ ▂▂ ▃ ──11 RGBA half
──12 RaGaBaA half
- ──13 R'G'B' half
- ──14 R'G'B'A half
+▃ ▂ ▁▂▂▁▂▃ ▂ ▁▂▁▂▂·▁ ▃▃▂▃▃▂▂▃ ▃ ▃▂ ▂▃ ▃ ──13 R'G'B' half
+▂ ▁▃▂▁▁▂▃▂ ▃▃ ▃▃ ▃ ▃▃ ·▁▁▂▂▁▂▃ ▃ ▃ ▃ ▂▂▂▃▃▃▃▂▃▂▃ ▃▃ ▃▂▃▃ ▃▃ ▃▃▃▃▃▂▃ ▃▃ ▃▃ ▃ ──14 R'G'B'A half
──15 R'aG'aB'aA half
──16 RGB u15
──17 RGBA u15
──25 R'G'B' u32
──26 R'G'B'A u32
──27 R'aG'aB'aA u32
-· â\96\83â\96\81â\96\83â\96\83â\96\81â\96\82 â\96\81 â\96\81 â\96\81â\96\81 â\96\82â\96\83â\96\83â\96\83â\96\83â\96\83 â\96\83 â\96\83 â\96\83 â\96\83 â\96\82â\96\82â\96\82â\96\83â\96\83 â\96\83â\96\82â\96\83 â\96\83 â\96\83â\96\81â\96\81 â\96\81â\96\81â\96\81â\96\81 â\96\82â\96\83 â\96\81â\96\81 ▁▁ ▁▁ ──28 Y double
-· â\96\83â\96\81â\96\83â\96\83â\96\81â\96\83 â\96\81 â\96\81â\96\81 â\96\82â\96\83â\96\83â\96\83â\96\83â\96\83 â\96\83 â\96\83 â\96\83 â\96\83 â\96\82â\96\82â\96\82â\96\83â\96\83 â\96\83 â\96\83â\96\82 â\96\83 â\96\83â\96\81â\96\81 â\96\81â\96\81â\96\81â\96\81 â\96\82â\96\83 â\96\81â\96\81 ▁▁ ▁▁ ──29 YA double
+· â\96\83â\96\81â\96\82â\96\83â\96\81â\96\82â\96\83â\96\82 â\96\83 â\96\81 â\96\81â\96\83 â\96\81â\96\81 â\96\82â\96\83â\96\83â\96\83â\96\83â\96\83 â\96\83 â\96\83 â\96\83 â\96\83 â\96\82â\96\82â\96\82â\96\83â\96\83â\96\83â\96\83â\96\82â\96\83 â\96\83 â\96\83 â\96\82â\96\81â\96\81 â\96\81â\96\81 â\96\81â\96\81â\96\81â\96\81â\96\83â\96\82â\96\83 ▁▁ ▁▁ ──28 Y double
+· â\96\83â\96\81â\96\82â\96\83â\96\81â\96\82â\96\83â\96\82 â\96\83 â\96\81 â\96\83 â\96\81â\96\81 â\96\82â\96\83â\96\83â\96\83â\96\83â\96\83 â\96\83 â\96\83 â\96\83 â\96\83 â\96\82â\96\82â\96\82â\96\83â\96\83â\96\83â\96\83 â\96\83â\96\83 â\96\83 â\96\83â\96\82â\96\81â\96\81 â\96\81â\96\81 â\96\81â\96\81â\96\81â\96\81â\96\83â\96\82â\96\83 ▁▁ ▁▁ ──29 YA double
──30 YaA double
──31 Y' double
──32 Y'A double
-· ▃▁▂▃▁▂ ▁▁ ▁▁ ▂▃▃▃▃▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▂▃▂ ▃ ▃ ▃▁▁ ▁▁▁▁ ▂▃ ▁▁ ▁▁ ▁▁ ──33 Y'aA double
- ──34 YA half
+· ▃▁▂▃▁▂▃▂ ▃ ▁▁ ▃ ▃ ▁▁ ▂▃▃▃▃▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃▃▃▂▃▂▃ ▃▃ ▃▂▁▁ ▁▁ ▁▁▁▁▃▂▃ ▁▁ ▁▁ ──33 Y'aA double
+▂ ▃▃▃▃▁▂▃▂ ▃ ▃▃ ▂▂ ▃ ▃▃ ▂▃▃▃▃▃ ▁▂▂▂▃▃▂▃▃ ▂▂▂▃▃▃▃·▁▁▁▂▂▁▂▂▂▃▃ ▃▃ ▃▃▃▃▃▂▁ ▃▃ ▃▃ ──34 YA half
──35 YaA half
- ──36 Y half
- ──37 Y'A half
+▃ ▃▃▂▁▂▃ ▃ ▂ ▃ ▂ ▃ ▃ ▃▂▃▂▂▁▃▃▂▃▃▂ ▃▃▂▃▃▂▂▁▁·▂▂▁▂▂▁▂ ▂▃▂ ▃ ──36 Y half
+▃ ▂ ▃▂▂▃ ▃ ▃▂ ▂ ▂ ▁▂▂▃▃▂▃·▁▁▁▃▂▁▃▂ ▃▃▃ ▃▁▂▂▂▃▃▂▃▃▃ ▃▂ ──37 Y'A half
──38 Y'aA half
- ──39 Y' half
+ ▃ ▂▃▃▂▃ ▂▃ ▃ ▂▂ ▂▃▂▃▃▁▂▁▁·▂▂▁▂▂▁ ▃ ▃▃▂▂▁▃▃▂▃▃▂▃ ▃ ▃ ▃ ──39 Y' half
──40 YA u15
──41 YaA u15
──42 Y u15
──49 Y'A u32
──50 Y'aA u32
──51 Y' u32
-· ▃▁▃▃▁▂ ▁▁ ▁ ▁ ▂▃▃▃▃▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▂▃▂ ▃ ▃ ▃▁▁ ▁▁▁▁ ▂▃ ▁▁ ▁▁ ▁▁ ──52 Y'CbCr double
-· ▃▁▃▃▁▂ ▁▁ ▁ ▂▃▃▃▃▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▂▃▂ ▃ ▃ ▂▁▁ ▁▁▁▁ ▂▃ ▁▁ ▁▁ ▁▁ ──53 Y'CbCrA double
+· ▃▁▃▃▁▂▃▂ ▃ ▁▁ ▁▃ ▃ ▁ ▂▃▃▃▃▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃▃▃▃▃▂▃ ▃ ▃▂▁▁ ▁▁ ▁▁▁▁▃▂▃ ▁▁ ▁▁ ──52 Y'CbCr double
+· ▃▁▃▃▁▂▃▂ ▃ ▁▁ ▁▃ ▃ ▂▃▃▃▃▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃▃▃▂▃▂▃ ▃▃ ▂▂▁▁ ▁▁ ▁▁▁▁▃▂▃ ▁▁ ▁▁ ──53 Y'CbCrA double
──54 Y'CbCr u8
-▁ ·▂▁··▁ ▂▂ ▃ ▂▂ ··▁▁· ▂▃▂ ▃ ▃ ▁▁▁▂▂ ▂▁▂▂ ▂ ▂ ▁▂▂ ▂▂▂▂ ▁▂ ▂▂ ▂▂ ▂▂ ▂ ──55 R'G'B'A float
-▂ ▁▃▂▁▁▂ ▃▃ ▃ ▃▃ ·▁▁▁▁▃ ▃ ▃ ▃ ▂▂▂▃▃▂▃▂▃▃ ▃ ▃ ▂▃▃ ▃▃▃▃ ▂▃ ▃▃ ▃▃ ▃▃ ▂ ──56 R'G'B'A u16
-▂ ▁ ▂▁▁▂ ▃▃ ▃ ▃▃ · ··▁ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▂▃▂ ▃ ▃ ▂▃▃ ▃▃▃▃ ▂▃ ▃▃ ▃▃ ▃▃ ▁ ──57 R'aG'aB'aA float
-▃ ▂ ▃▂▂▃ ▃▃ ▃ ▃▃ ▁▂ ·▂▂▃ ▃ ▃ ▃ ▃▃▃▃▃▃▃▃▃▃ ▃ ▃ ▃▃▃ ▃▃▃▃ ▃▃ ▃▃ ▃▃ ▃▃ ▁ ──58 R'aG'aB'aA u16
-▃ ▂ ▂▃ ▁▂· ▂▂ ▃▃▃ ▃▃▃▃ ▃ ▃ ▃ · ──59 R'aG'aB'aA u8
-▂ ▃▃·▁▁· ▃▃ ▃ ▃▃ ▂▃▃▃ ▃ ▃ ▃ ▃ ▂▂▁▂▂ ▁▂▃▂ ▃ ▃ ▁▃▃ ▃▃▃▃ ▂▃ ▃▃ ▃▃ ▃▃ ▂ ──60 R'G'B' float
-▃ ·▁▂▁ ▃ ▃ ▁ ▃▃▃▃▃▃▂▃ ▃ ▃ ▁ ▃ ▂ ──61 R'G'B' u16
-▂ ▁▃▂▁▁▂ ▃▃ ▃ ▃▃ ·▁▁▂▂▁ ·· ▂ ·▁▁ ▂▂▂▃▃ ▃·▁▁ ▁ ▂ ▂▃▃ ▃▃▃▃ ▂▁ ▃▃ ▃▃ ▃▃ ▃ ──62 Y'A float
-▃ ▂ ▃▂▂▃ ▁▂·▁▁▂ ▁ ▁ ▁·▂ ▃▃▃ ▃▁▂▂ ▂ ▃ ▃ ▃▂ ▂ ──63 Y'aA float
-▃ ▂ ▁▂▂▁ ▁▂▁▂▂· · ▁ · ▃▃▂▃▃ ▂▁▁· ▂ ▁ ▂ ▃▂ ▃ ──64 Y' float
-▃ ▂ ▃▃▂▃ ▃ ▁▂▂▃▃▂▂·▂ ▃ ·▃▃ ▃▃▃ ▃▃▃▁▂▂ ▂ ▃ ▃ ▃▂ ▃ ──65 Y'A u16
- ▃ ▃ ▂▃▁▂▂▃ ▁·▂ ▂▁▃ ▃ ▃ ▃ ──66 Y'aA u16
- ▃ ▃▃▃▂ ▂▃▂▃▃▁▃▁▁· ▃▃· ▃ ▃▂▂▁ ▃ ▂ ▃ ▃ ▂ ──67 Y' u16
-▁ ▂▃▃▂·▂ ▂▂ ▂ ▂▂ ▁▃▂▃▃▂▂·▁▁·▃ ▂▂ ▁▁▂▂▂▁▂·▁▁ ▁ ▂ ▃▂▂ ▂▃▂▂ ▁▁ ▃▃ ▃▂ ▂▂ ▃ ──68 Y'A u8
- ▃ ▃ ▂▃▁▂▂▃ ▁·▂ · ▃▃▃ ▃ ▃ ──69 Y'aA u8
-▁ ▃▂▂▃·▁ ▂▂ ▂ ▂▂ ▂▃▂▃▃▁▃▁▁·▃▃·▂ ▁▁▂▃▂▁▂▁▁· ▂ ▁ ▃▂▂ ▂▂▂▂ ▃▂ ▂▂ ▂▂ ▂▂ ▃ ──70 Y' u8
-▃ ▁▂▂▁ ▃ ▃ ▂· ▃▃▂▃▃ ▂▃ ▃ ▃ ▂ ▃ ▃ ──71 Y'CbCr float
-▂ ▁▃▂▁▁▂ ▃▃ ▃ ▃ ·▁▁▂▂▁ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▂▃▂ ▃ ▃ ▃▃▃ ▃▃▃▃ ▂▃ ▃▃ ▃▃ ▃▃ ▃ ──72 Y'CbCrA float
-▁ ▂▂▃▂·▁ ▂▂ ▂ ▂▂ ▁▂▂▃▂▂▂▂▃▂ ▃ ▃ ··▂▁▁▂▁▂▁ ▂ ▂ ▃▂▂ ▂▂▂▂ ▂▂ ▂▃ ▃▂ ▂▂ ▃ ──73 RGBA u16
-▁ ▂▂▂▂·▁ ▂▂ ▂ ▃▃ ▁▂▂▃▃▂▂▂▃▂ ▃ ▃ ·▂▁▁·▁▂▁ ▂ ▂ ▂▂▃ ▂▂▂▂ ▁▂ ▃▂ ▂▂ ▂▂ ▃ ──74 RGBA u8
-▁ ▂▂▃▂·▁ ▂▂ ▂ ▂▂ ▁▂▂▃▃▂ ▂▃▂ ▃ ▃ ▁ ·· ▂▁▂▁ ▂ ▂ ▃▂▂ ▂▂▂▂ ▁▂ ▂▂ ▂▂ ▂▂ ▃ ──75 RaGaBaA float
-▂ ▃▃▂▃▁▂ ▃▃ ▃ ▃▃ ▂▃▃▃▃▃▃▃ ▃ ▃ ▃ ▂▂ ·▂▃▂▃▂ ▃ ▃ ▂▃▃ ▃▃▃▃ ▂▃ ▃▃ ▃▃ ▃▃ ▃ ──76 RaGaBaA u16
-▂ ▃▃▂▃▁▂ ▃▃ ▃ ▃▃ ▂▃▃▂▃▃▃▃ ▃ ▃ ▃ ▂ · ▂ ▂▃▂ ▃ ▃ ▂▃▃ ▃▃▃▃ ▂▃ ▃▃ ▃▃ ▃▃ ▃ ──77 RaGaBaA u8
-▂ ▃▃▂▃▁· ▃▃ ▃ ▃▃ ▂▃▃▃▃▁▃▃ ▃ ▃ ▃ ▂▁▁▂ ·▂▃▃ ▃ ▃ ▃▃▃ ▃▃▃▃ ▂▃ ▃▃ ▃▃ ▃▃ ▃ ──78 RGB u16
-▁ ▂▂▁▂·· ▂▂ ▂ ▃▂ ▁▂▂▃▂▁▃▃▃▂ ▃ ▃ ▁··▁· ▁▂▁ ▂ ▂ ▂▂▂ ▂▂▃▂ ▂▂ ▂▂ ▂▂ ▂▂ ▃ ──79 RGB u8
-▁ ▂▂▃▂·▁ ▂ ▂ ▂▂ ▁▂▂▃▃▂ ·▁▁ ▃ ▁▃▂ ▁▁▁▂▂ ·· · ▁ ▃▂▂ ▂▂▂▂ ▁· ▂▂ ▂▂ ▂▂ ▃ ──80 YA float
-▂ ▃▃▂▃▁▂ ▃▃ ▃ ▃▃ ▂▃▃▃▃▃ ▃▃▃ ▂▃▃ ▂▂·▁▁ ▃ ▂ ▁ ▂ ▃▃▃ ▃▃▃▃ ▂▃ ▃▃ ▃▃ ▃▃ ──81 YaA float
-▂ ▃▃▂▃▁· ▃ ▃ ▃▃ ▂▃▂▃▃▁ ▁▁· ▂▃▁ ▂▂▁▂▂ ▁· ▁ · ▃▃▃ ▃▃▃▃ ▂▁ ▃▃ ▃▃ ▃▃ ▃ ──82 Y float
-▂ ▃▃ ▃▁▂ ▃ ▃ ▃▃ ▂▃▃▃▃▃▃▁▃▂ ▂▃▃ ▂▂▂▃▃▂▃·▁ · ▃ ▃▃ ▃▃▃▃ ▂▁ ▃ ▃▃ ▃▃ ──83 YA u16
-▃ ▃ ▂▃ ▃ ▃▃▃ ▂▃▃ ▃▃▁▂▂▃▃▃·▃ ▃ ▃ ▃▂ ──84 YaA u16
-▃ ▃ ▂▁ ▃ ▃ ▂ ▂▃▁ ▃ ▂ ▃▂▂▃▃▃▃▁▁· ▃ · ▃▂ ──85 Y u16
-▁ ▃▃▃▃·▂ ▂▂ ▂ ▂▃ ▂▃▃▃▃▃▃▁▂▂ ▂▃▃ ▂▁▁▂▂▃▃·▁▁· ▂ ▃▃▃ ▃▃▂▃ ▂▁ ▃▂ ▃▂ ▂▃ ▃ ──86 YA u8
-▃ ▃ ▂▃ ▃ ▃▃▃ ▂▃▃ ▃ ▁▂▂▃ ▁·▂ · ▃ ▃▂ ──87 YaA u8
-▁ ▂▂▃▃·▁ ▂▂ ▂ ▂▂ ▃▂▃▃▃▂▂▂▂▁ ▃ ▁ ▁▁▂▃▂▃▂▁▁·▃ ·▂ ▂▂▂ ▂▂▂▂ ▃▂ ▂▂ ▂▂ ▂▂ ▃ ──88 Y u8
- ──89 B'aG'aR'aA u8
- ──90 cairo-RGB24
-· ▃▁▂▃▁▂ ▁▁ ▁ ▁▁ ▂▃▃▃▃▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▂▃▂ ▃ ▃ ▁ ▁▁▁▁ ▂▃ ▁▁ ▁▁ ▁▁ ──91 HSVA double
-· ▃▁▃▃▁▂ ▁▁ ▁ ▁▁ ▂▃▃▃▃▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▃▃▂ ▃ ▃ ▃ ▁▁▁▁ ▂▃ ▁▁ ▁▁ ▁▁ ──92 HSV double
- ──93 HSVA float
- ──94 HSV float
-· ▃▁▃▃▁▂ ▁▁ ▁ ▁▁ ▂▃▃▃▃▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▂▃▃ ▃ ▃ ▃▁▁ ▁▁▁ ▂▃ ▁▁ ▁▁ ▁▁ ──95 CIE Lab double
-· ▃▁▃▃▁▂ ▁▁ ▁ ▁▁ ▂▃▃▃▃▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▂▃▂ ▃ ▃ ▂▁▁ ▁▁ ▁▁ ▁▁ ▁▁ ──96 CIE Lab alpha double
-· ▃▁▃▃▁▂ ▁▁ ▁ ▁▁ ▂▃▃▃▃▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▃▃▂ ▃ ▃ ▂▁▁ ▁ ▁ ▂▃ ▁▁ ▁▁ ▁▁ ──97 CIE LCH(ab) double
-· ▃▁▃▃▁▂ ▁▁ ▁ ▁▁ ▂▃▃▃▃▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▂▃▂ ▃ ▃ ▂▁▁ ▁▁ ▂▃ ▁▁ ▁▁ ▁▁ ──98 CIE LCH(ab) alpha double
- ──99 CIE Lab float
-▁ ▂▂▃▂·▁ ▂▂ ▂ ▂▂ ▁▂▂▃▃▂ ▂▃▂ ▃ ▃ ▁▁▁▂▂ ▂▁▂▁ ▂ ▂ ▁▂▂ ▂ ▂▂ ▂▂ ▂▂ ▂▂ ▃ ──100 CIE Lab alpha float
- ──101 CIE L alpha float
- ──102 CIE Lab u8
- ──103 CIE Lab u16
- ──104 CIE LCH(ab) float
- ──105 CIE LCH(ab) alpha float
-· ▃▁▂▃▁▂ ▁▁ ▁ ▁▁ ▂▃▃▃▃▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▂▃▂ ▃ ▃ ▂▁▁ ▁▁▁▁ ▂▃ ▁ ▁▁ ▁▁ ──106 CMYK double
-· ▃▁▂▃▁▂ ▁▁ ▁ ▁▁ ▂▃▃▃▃▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▂▃▃ ▃ ▂ ▃▁▁ ▁▁▁▁ ▂▃ ▁▁ ▁▁ ──107 CMY double
- ──108 CMYK float
- ──109 CMY float
- ──110 CMYK u8
-· ▃▁▂▃▁▂ ▁▁ ▁ ▁▁ ▂▃▃▃▃▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▂▃▂ ▃ ▃ ▃▁▁ ▁▁▁▁ ▂▃ ▁▁ ▁ ▁▁ ──111 HSLA double
-· ▃▁▂▃▁▂ ▁▁ ▁ ▁▁ ▂▃▃▃▃▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▃▃▃ ▃ ▂ ▂▁▁ ▁▁▁▁ ▂▃ ▁▁ ▁▁ ──112 HSL double
- ──113 HSLA float
- ──114 HSL float
-· ▃▁▃▃▁▂ ▁▁ ▁ ▁▁ ▂▃▃▃▃▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▂▃▃ ▃ ▃ ▂▁▁ ▁▁▁▁ ▂▃ ▁▁ ▁▁ ▁ ──115 Y'CbCr709 double
-· ▃▁▃▃▁▂ ▁▁ ▁ ▁▁ ▂▃▃▃▃▃ ▃ ▃ ▃ ▃ ▂▂▂▃▃ ▃▂▃▂ ▃ ▃ ▂▁▁ ▁▁▁▁ ▂▃ ▁▁ ▁▁ ──116 Y'CbCrA709 double
- ──117 Y'CbCrA709 float
- ──118 Y'CbCr709 float
- ──119 cairo-ARGB32
- ──120 cairo-A8
-|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
-||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||111111111111111111111
-||||||||||111111111122222222223333333333444444444455555555556666666666777777777788888888889999999999000000000011111111112
-0123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890
-total length: 7788
-total cost : 20663828
+▁ ·▂▁··▁▂▁ ▁· ▂▂ ▃▂ ▂▃ ▃ ▂▂ ··▁▁·▁▂▃▂▃ ▃▃ ▃ ▁▁▁▂▂▂▂▁▂▁▂▃▂▂▃▃▁▂▂ ▂▂ ▂▂▂▂▂▁▂ ▂▂ ▂▂ ▂ ──55 R'G'B'A float
+▂ ▁▃▂·▁▂▃▂ ▂▁ ▃▃ ▃▃ ▃ ▃▃ ·▁▁▁▁▃ ▃ ▃ ▃ ▂▂▃▃▃▂▃▂▃▂▃ ▃▃ ▃▂▃▃ ▃▃ ▃▃▃▃▃▂▃ ▃▃ ▃▃ ▂ ──56 R'G'B'A u16
+▂ ▁ ▂▁▁▂▃▂ ▂▁ ▃▃ ▃▃ ▃ ▃▃ · ··▁▂▃ ▃ ▃ ▃ ▂▂▂▃▃▃▃▂▃▃▃ ▃▃ ▃▂▃▃ ▃▃ ▃▃▃▃▃▂▃ ▃▃ ▃▃ ▁ ──57 R'aG'aB'aA float
+▃ ▂ ▃▂▂▃▃▃ ▃▂ ▃▃ ▃▃ ▃ ▃▃ ▁▂ ·▂▃▃ ▃ ▃ ▃ ▃▃▃▃▃▃▃▃▃▃▃ ▃▃ ▃▃▃▃ ▃▃ ▃▃▃▃▃▃▃ ▃▃ ▃▃ ▁ ──58 R'aG'aB'aA u16
+▃ ▂ ▂▃ ▃ ▃▂ ▁▂· ▂▂ ▃▃▃ ▃▃▃▃ ▃ ▃ ▃ · ──59 R'aG'aB'aA u8
+▂ ▁▃·▁▁·▁▂ ·▁ ▃▃ ▃▃ ▃ ▃▃ ·▁·▁ ·▃ ▃ ▃ ▃ ▂▂▁▃▂▁▁▂▃▂▃ ▃▃ ▃▁▃▃ ▃▃ ▃▃▃▃▁▂▃ ▃▃ ▃▃ ▂ ──60 R'G'B' float
+▃ ▂ ·▁▂▁▂▃ ▁▂ ▁▂▁▂▁ ▃▃▂▃▃▂▂▃ ▃ ▃▁ ▂▃ ▂ ──61 R'G'B' u16
+▂ ▁▃▂▁▁▂▃▂ ▂▁ ▃▃ ▃▁ ▂· ▁ ▃▃ ·▁▁▂▂▁ ···▁▁·▁▁ ▂▂▂▃▃▃▃·▁▁▁▂▂▁▂▂▃▃▃ ▃▃ ▃▃▃▃▃▂▁ ▃▃ ▃▃ ▃ ──62 Y'A float
+▃ ▂ ▃▂▂▃ ▃ ▃▂ ▂ ▃▁ ▂ ▁▂·▁▁▂▃ ▁▁▁▂▁·▂ ▃▃▃ ▃▁▂▂▂▃▃▂▃▃▃ ▃▂ ▂ ──63 Y'aA float
+▃ ▂ ▁▂▂▁▂▃ ▁▂ ▂ ▁▁ · ▁▂▁▂▂·▁· ▁▁·▁▁· ▃▃▂▃▃▂▂▁▁·▂▂▁▂▂▁▂ ▂▃▂ ▃ ──6
Environment
Massimo Valentini
stability fixes
-/babl-0.1.14
+/babl-0.1.16
ECHO_T = @ECHO_T@
EGREP = @EGREP@
EXEEXT = @EXEEXT@
+F16C_EXTRA_CFLAGS = @F16C_EXTRA_CFLAGS@
FGREP = @FGREP@
GREP = @GREP@
INSTALL = @INSTALL@
SHELL = @SHELL@
SHREXT = @SHREXT@
SSE2_EXTRA_CFLAGS = @SSE2_EXTRA_CFLAGS@
+SSE4_1_EXTRA_CFLAGS = @SSE4_1_EXTRA_CFLAGS@
SSE_EXTRA_CFLAGS = @SSE_EXTRA_CFLAGS@
STRIP = @STRIP@
THREAD_LIB = @THREAD_LIB@
enum
{
- ARCH_X86_INTEL_FEATURE_PNI = 1 << 0
+ ARCH_X86_INTEL_FEATURE_PNI = 1 << 0,
+ ARCH_X86_INTEL_FEATURE_SSSE3 = 1 << 9,
+ ARCH_X86_INTEL_FEATURE_SSE4_1 = 1 << 19,
+ ARCH_X86_INTEL_FEATURE_SSE4_2 = 1 << 20,
+ ARCH_X86_INTEL_FEATURE_AVX = 1 << 28,
+ ARCH_X86_INTEL_FEATURE_F16C = 1 << 29,
};
#if !defined(ARCH_X86_64) && (defined(PIC) || defined(__PIC__))
if (ecx & ARCH_X86_INTEL_FEATURE_PNI)
caps |= BABL_CPU_ACCEL_X86_SSE3;
+
+ if (ecx & ARCH_X86_INTEL_FEATURE_SSSE3)
+ caps |= BABL_CPU_ACCEL_X86_SSSE3;
+
+ if (ecx & ARCH_X86_INTEL_FEATURE_SSE4_1)
+ caps |= BABL_CPU_ACCEL_X86_SSE4_1;
+
+ if (ecx & ARCH_X86_INTEL_FEATURE_F16C)
+ caps |= BABL_CPU_ACCEL_X86_F16C;
#endif /* USE_SSE */
}
#endif /* USE_MMX */
#ifdef USE_SSE
if ((caps & BABL_CPU_ACCEL_X86_SSE) && !arch_accel_sse_os_support ())
- caps &= ~(BABL_CPU_ACCEL_X86_SSE | BABL_CPU_ACCEL_X86_SSE2);
+ caps &= ~(BABL_CPU_ACCEL_X86_SSE |
+ BABL_CPU_ACCEL_X86_SSE2 |
+ BABL_CPU_ACCEL_X86_SSE3 |
+ BABL_CPU_ACCEL_X86_SSSE3 |
+ BABL_CPU_ACCEL_X86_SSE4_1);
#endif
return caps;
BABL_CPU_ACCEL_X86_SSE = 0x10000000,
BABL_CPU_ACCEL_X86_SSE2 = 0x08000000,
BABL_CPU_ACCEL_X86_SSE3 = 0x02000000,
+ BABL_CPU_ACCEL_X86_SSSE3 = 0x00800000,
+ BABL_CPU_ACCEL_X86_SSE4_1 = 0x00400000,
+ /* BABL_CPU_ACCEL_X86_SSE4_2 = 0x00200000, */
+ /* BABL_CPU_ACCEL_X86_AVX = 0x00080000, */
+ BABL_CPU_ACCEL_X86_F16C = 0x00040000,
/* powerpc accelerations */
BABL_CPU_ACCEL_PPC_ALTIVEC = 0x04000000,
#define BABL_MAJOR_VERSION 0
#define BABL_MINOR_VERSION 1
-#define BABL_MICRO_VERSION 14
+#define BABL_MICRO_VERSION 16
/** Get the version information on the babl library */
void babl_get_version (int *major,
ECHO_T = @ECHO_T@
EGREP = @EGREP@
EXEEXT = @EXEEXT@
+F16C_EXTRA_CFLAGS = @F16C_EXTRA_CFLAGS@
FGREP = @FGREP@
GREP = @GREP@
INSTALL = @INSTALL@
SHELL = @SHELL@
SHREXT = @SHREXT@
SSE2_EXTRA_CFLAGS = @SSE2_EXTRA_CFLAGS@
+SSE4_1_EXTRA_CFLAGS = @SSE4_1_EXTRA_CFLAGS@
SSE_EXTRA_CFLAGS = @SSE_EXTRA_CFLAGS@
STRIP = @STRIP@
THREAD_LIB = @THREAD_LIB@
}
}
-#if 0
-//-----------------------------------------------------------------------------
-
-static void halfp2singles(void *target, void *source, long numel)
-{
- uint16_t *hp = (uint16_t *) source; // Type pun input as an unsigned 16-bit int
- uint32_t *xp = (uint32_t *) target; // Type pun output as an unsigned 32-bit int
- uint16_t h, hs, he, hm;
- uint32_t xs, xe, xm;
- int32_t xes;
- int e;
-
- if( source == NULL || target == NULL ) // Nothing to convert (e.g., imag part of pure real)
- return;
- while( numel-- ) {
- h = *hp++;
- if( (h & 0x7FFFu) == 0 ) { // Signed zero
- *xp++ = ((uint32_t) h) << 16; // Return the signed zero
- } else { // Not zero
- hs = h & 0x8000u; // Pick off sign bit
- he = h & 0x7C00u; // Pick off exponent bits
- hm = h & 0x03FFu; // Pick off mantissa bits
- if( he == 0 ) { // Denormal will convert to normalized
- e = -1; // The following loop figures out how much extra to adjust the exponent
- do {
- e++;
- hm <<= 1;
- } while( (hm & 0x0400u) == 0 ); // Shift until leading bit overflows into exponent bit
- xs = ((uint32_t) hs) << 16; // Sign bit
- xes = ((int32_t) (he >> 10)) - 15 + 127 - e; // Exponent unbias the halfp, then bias the single
- xe = (uint32_t) (xes << 23); // Exponent
- xm = ((uint32_t) (hm & 0x03FFu)) << 13; // Mantissa
- *xp++ = (xs | xe | xm); // Combine sign bit, exponent bits, and mantissa bits
- } else if( he == 0x7C00u ) { // Inf or NaN (all the exponent bits are set)
- if( hm == 0 ) { // If mantissa is zero ...
- *xp++ = (((uint32_t) hs) << 16) | ((uint32_t) 0x7F800000u); // Signed Inf
- } else {
- *xp++ = (uint32_t) 0xFFC00000u; // NaN, only 1st mantissa bit set
- }
- } else { // Normalized number
- xs = ((uint32_t) hs) << 16; // Sign bit
- xes = ((int32_t) (he >> 10)) - 15 + 127; // Exponent unbias the halfp, then bias the single
- xe = (uint32_t) (xes << 23); // Exponent
- xm = ((uint32_t) hm) << 13; // Mantissa
- *xp++ = (xs | xe | xm); // Combine sign bit, exponent bits, and mantissa bits
- }
- }
- }
-}
-
-static void singles2halfp(void *target, void *source, long numel)
-{
- uint16_t *hp = (uint16_t *) target; // Type pun output as an unsigned 16-bit int
- uint32_t *xp = (uint32_t *) source; // Type pun input as an unsigned 32-bit int
- uint16_t hs, he, hm;
- uint32_t x, xs, xe, xm;
- int hes;
-
- if( source == NULL || target == NULL ) { // Nothing to convert (e.g., imag part of pure real)
- return;
- }
- while( numel-- ) {
- x = *xp++;
- if( (x & 0x7FFFFFFFu) == 0 ) { // Signed zero
- *hp++ = (uint16_t) (x >> 16); // Return the signed zero
- } else { // Not zero
- xs = x & 0x80000000u; // Pick off sign bit
- xe = x & 0x7F800000u; // Pick off exponent bits
- xm = x & 0x007FFFFFu; // Pick off mantissa bits
- if( xe == 0 ) { // Denormal will underflow, return a signed zero
- *hp++ = (uint16_t) (xs >> 16);
- } else if( xe == 0x7F800000u ) { // Inf or NaN (all the exponent bits are set)
- if( xm == 0 ) { // If mantissa is zero ...
- *hp++ = (uint16_t) ((xs >> 16) | 0x7C00u); // Signed Inf
- } else {
- *hp++ = (uint16_t) 0xFE00u; // NaN, only 1st mantissa bit set
- }
- } else { // Normalized number
- hs = (uint16_t) (xs >> 16); // Sign bit
- hes = ((int)(xe >> 23)) - 127 + 15; // Exponent unbias the single, then bias the halfp
- if( hes >= 0x1F ) { // Overflow
- *hp++ = (uint16_t) ((xs >> 16) | 0x7C00u); // Signed Inf
- } else if( hes <= 0 ) { // Underflow
- if( (14 - hes) > 24 ) { // Mantissa shifted all the way off & no rounding possibility
- hm = (uint16_t) 0u; // Set mantissa to zero
- } else {
- xm |= 0x00800000u; // Add the hidden leading bit
- hm = (uint16_t) (xm >> (14 - hes)); // Mantissa
- if( (xm >> (13 - hes)) & 0x00000001u ) // Check for rounding
- hm += (uint16_t) 1u; // Round, might overflow into exp bit, but this is OK
- }
- *hp++ = (hs | hm); // Combine sign bit and mantissa bits, biased exponent is zero
- } else {
- he = (uint16_t) (hes << 10); // Exponent
- hm = (uint16_t) (xm >> 13); // Mantissa
- if( xm & 0x00001000u ) // Check for rounding
- *hp++ = (hs | he | hm) + (uint16_t) 1u; // Round, might overflow to inf, this is OK
- else
- *hp++ = (hs | he | hm); // No rounding
- }
- }
- }
- }
-}
-#endif
-
-//-----------------------------------------------------------------------------
-
static void halfp2doubles(void *target, void *source, long numel)
{
uint16_t *hp = (uint16_t *) source; // Type pun input as an unsigned 16-bit int
uint32_t xs, xe, xm;
int32_t xes;
int e;
-
- xp += next; // Little Endian adjustment if necessary
if( source == NULL || target == NULL ) // Nothing to convert (e.g., imag part of pure real)
return;
while( numel-- ) {
+ uint32_t x;
+
h = *hp++;
if( (h & 0x7FFFu) == 0 ) { // Signed zero
- *xp++ = ((uint32_t) h) << 16; // Return the signed zero
+ x = ((uint32_t) h) << 16; // Return the signed zero
} else { // Not zero
hs = h & 0x8000u; // Pick off sign bit
he = h & 0x7C00u; // Pick off exponent bits
xes = ((int32_t) (he >> 10)) - 15 + 1023 - e; // Exponent unbias the halfp, then bias the double
xe = (uint32_t) (xes << 20); // Exponent
xm = ((uint32_t) (hm & 0x03FFu)) << 10; // Mantissa
- *xp++ = (xs | xe | xm); // Combine sign bit, exponent bits, and mantissa bits
+ x = (xs | xe | xm); // Combine sign bit, exponent bits, and mantissa bits
} else if( he == 0x7C00u ) { // Inf or NaN (all the exponent bits are set)
if( hm == 0 ) { // If mantissa is zero ...
- *xp++ = (((uint32_t) hs) << 16) | ((uint32_t) 0x7FF00000u); // Signed Inf
+ x = (((uint32_t) hs) << 16) | ((uint32_t) 0x7FF00000u); // Signed Inf
} else {
- *xp++ = (uint32_t) 0xFFF80000u; // NaN, only the 1st mantissa bit set
+ x = (uint32_t) 0xFFF80000u; // NaN, only the 1st mantissa bit set
}
} else {
xs = ((uint32_t) hs) << 16; // Sign bit
xes = ((int32_t) (he >> 10)) - 15 + 1023; // Exponent unbias the halfp, then bias the double
xe = (uint32_t) (xes << 20); // Exponent
xm = ((uint32_t) hm) << 10; // Mantissa
- *xp++ = (xs | xe | xm); // Combine sign bit, exponent bits, and mantissa bits
+ x = (xs | xe | xm); // Combine sign bit, exponent bits, and mantissa bits
}
}
- xp++; // Skip over the remaining 32 bits of the mantissa
+
+ xp[1 - next] = 0;
+ xp[next] = x;
+
+ xp += 2;
}
}
/* Define to 1 if AltiVec support is available. */
#undef USE_ALTIVEC
+/* Define to 1 if f16c intrinsics are available. */
+#undef USE_F16C
+
/* Define to 1 if MMX assembly is available. */
#undef USE_MMX
/* Define to 1 if SSE2 assembly is available. */
#undef USE_SSE2
+
+/* Define to 1 if SSE4_1 assembly is available. */
+#undef USE_SSE4_1
#! /bin/sh
# Guess values for system-dependent variables and create Makefiles.
-# Generated by GNU Autoconf 2.69 for babl 0.1.14.
+# Generated by GNU Autoconf 2.69 for babl 0.1.16.
#
#
# Copyright (C) 1992-1996, 1998-2012 Free Software Foundation, Inc.
# Identity of this package.
PACKAGE_NAME='babl'
PACKAGE_TARNAME='babl'
-PACKAGE_VERSION='0.1.14'
-PACKAGE_STRING='babl 0.1.14'
+PACKAGE_VERSION='0.1.16'
+PACKAGE_STRING='babl 0.1.16'
PACKAGE_BUGREPORT=''
PACKAGE_URL=''
HAVE_SRANDOM_FALSE
HAVE_SRANDOM_TRUE
ALTIVEC_EXTRA_CFLAGS
+F16C_EXTRA_CFLAGS
+SSE4_1_EXTRA_CFLAGS
SSE2_EXTRA_CFLAGS
SSE_EXTRA_CFLAGS
MMX_EXTRA_CFLAGS
enable_mmx
enable_sse
enable_sse2
+enable_sse4_1
+enable_f16c
enable_altivec
'
ac_precious_vars='build_alias
# Omit some internal or obsolete options to make the list less imposing.
# This message is too long to be a string in the A/UX 3.1 sh.
cat <<_ACEOF
-\`configure' configures babl 0.1.14 to adapt to many kinds of systems.
+\`configure' configures babl 0.1.16 to adapt to many kinds of systems.
Usage: $0 [OPTION]... [VAR=VALUE]...
if test -n "$ac_init_help"; then
case $ac_init_help in
- short | recursive ) echo "Configuration of babl 0.1.14:";;
+ short | recursive ) echo "Configuration of babl 0.1.16:";;
esac
cat <<\_ACEOF
--enable-mmx enable MMX support (default=auto)
--enable-sse enable SSE support (default=auto)
--enable-sse2 enable SSE2 support (default=auto)
+ --enable-sse4_1 enable SSE4_1 support (default=auto)
+ --enable-f16c enable hardware half-float support (default=auto)
--enable-altivec enable AltiVec support (default=auto)
Optional Packages:
test -n "$ac_init_help" && exit $ac_status
if $ac_init_version; then
cat <<\_ACEOF
-babl configure 0.1.14
+babl configure 0.1.16
generated by GNU Autoconf 2.69
Copyright (C) 2012 Free Software Foundation, Inc.
This file contains any messages produced by compilers while
running configure, to aid debugging if configure makes a mistake.
-It was created by babl $as_me 0.1.14, which was
+It was created by babl $as_me 0.1.16, which was
generated by GNU Autoconf 2.69. Invocation command line was
$ $0 $@
BABL_MAJOR_VERSION=0
BABL_MINOR_VERSION=1
-BABL_MICRO_VERSION=14
+BABL_MICRO_VERSION=16
BABL_INTERFACE_AGE=1
-BABL_BINARY_AGE=114
-BABL_VERSION=0.1.14
-BABL_REAL_VERSION=0.1.14
+BABL_BINARY_AGE=116
+BABL_VERSION=0.1.16
+BABL_REAL_VERSION=0.1.16
BABL_API_VERSION=0.1
-BABL_LIBRARY_VERSION="113:1:113"
+BABL_LIBRARY_VERSION="115:1:115"
BABL_CURRENT_MINUS_AGE=0
# Define the identity of the package.
PACKAGE='babl'
- VERSION='0.1.14'
+ VERSION='0.1.16'
# Some tools Automake needs.
fi
+# Check whether --enable-sse4_1 was given.
+if test "${enable_sse4_1+set}" = set; then :
+ enableval=$enable_sse4_1;
+else
+ enable_sse4_1=$enable_sse
+fi
+
+
+# Check whether --enable-f16c was given.
+if test "${enable_f16c+set}" = set; then :
+ enableval=$enable_f16c;
+else
+ enable_f16c=$enable_sse
+fi
+
+
if test "x$enable_mmx" = xyes; then
MMX_EXTRA_CFLAGS=
SSE_EXTRA_CFLAGS=
SSE2_EXTRA_CFLAGS=
+ SSE4_1_EXTRA_CFLAGS=
{ $as_echo "$as_me:${as_lineno-$LINENO}: checking whether we can compile MMX code" >&5
$as_echo_n "checking whether we can compile MMX code... " >&6; }
{ $as_echo "$as_me:${as_lineno-$LINENO}: WARNING: The assembler does not support the SSE2 command set." >&5
$as_echo "$as_me: WARNING: The assembler does not support the SSE2 command set." >&2;}
+fi
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+
+ if test "x$enable_sse4_1" = xyes; then
+
+ sse4_1_flag=
+ for flag in '-msse4.1'; do
+ if test -z "$sse4_1_flag"; then
+ sse4_1_flag_save_CFLAGS="$CFLAGS"
+ CFLAGS="$CFLAGS $flag"
+ { $as_echo "$as_me:${as_lineno-$LINENO}: checking whether $CC understands $flag" >&5
+$as_echo_n "checking whether $CC understands $flag... " >&6; }
+ cat confdefs.h - <<_ACEOF >conftest.$ac_ext
+/* end confdefs.h. */
+
+int
+main ()
+{
+
+ ;
+ return 0;
+}
+_ACEOF
+if ac_fn_c_try_compile "$LINENO"; then :
+ sse4_1_flag_works=yes
+else
+ sse4_1_flag_works=no
+fi
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+ { $as_echo "$as_me:${as_lineno-$LINENO}: result: $sse4_1_flag_works" >&5
+$as_echo "$sse4_1_flag_works" >&6; }
+ CFLAGS="$sse4_1_flag_save_CFLAGS"
+ if test "x$sse4_1_flag_works" = "xyes"; then
+ sse4_1_flag="$flag"
+ fi
+ fi
+ done
+
+ SSE4_1_EXTRA_CFLAGS="$SSE_EXTRA_CFLAGS $sse4_1_flag"
+
+ { $as_echo "$as_me:${as_lineno-$LINENO}: checking whether we can compile SSE4_1 code" >&5
+$as_echo_n "checking whether we can compile SSE4_1 code... " >&6; }
+
+ CFLAGS="$CFLAGS $sse4_1_flag"
+
+ cat confdefs.h - <<_ACEOF >conftest.$ac_ext
+/* end confdefs.h. */
+
+int
+main ()
+{
+asm ("pmovzxbd %xmm0,%xmm1");
+ ;
+ return 0;
+}
+_ACEOF
+if ac_fn_c_try_compile "$LINENO"; then :
+
+$as_echo "#define USE_SSE4_1 1" >>confdefs.h
+
+ { $as_echo "$as_me:${as_lineno-$LINENO}: result: yes" >&5
+$as_echo "yes" >&6; }
+
+else
+ enable_sse4_1=no
+ { $as_echo "$as_me:${as_lineno-$LINENO}: result: no" >&5
+$as_echo "no" >&6; }
+ { $as_echo "$as_me:${as_lineno-$LINENO}: WARNING: The assembler does not support the SSE4_1 command set." >&5
+$as_echo "$as_me: WARNING: The assembler does not support the SSE4_1 command set." >&2;}
+
+fi
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+ fi
+ fi
+
+ CFLAGS="$mmx_save_CFLAGS"
+
+ if test "x$enable_f16c" = xyes; then
+
+ f16c_flag=
+ for flag in '-mf16c'; do
+ if test -z "$f16c_flag"; then
+ f16c_flag_save_CFLAGS="$CFLAGS"
+ CFLAGS="$CFLAGS $flag"
+ { $as_echo "$as_me:${as_lineno-$LINENO}: checking whether $CC understands $flag" >&5
+$as_echo_n "checking whether $CC understands $flag... " >&6; }
+ cat confdefs.h - <<_ACEOF >conftest.$ac_ext
+/* end confdefs.h. */
+
+int
+main ()
+{
+
+ ;
+ return 0;
+}
+_ACEOF
+if ac_fn_c_try_compile "$LINENO"; then :
+ f16c_flag_works=yes
+else
+ f16c_flag_works=no
+fi
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+ { $as_echo "$as_me:${as_lineno-$LINENO}: result: $f16c_flag_works" >&5
+$as_echo "$f16c_flag_works" >&6; }
+ CFLAGS="$f16c_flag_save_CFLAGS"
+ if test "x$f16c_flag_works" = "xyes"; then
+ f16c_flag="$flag"
+ fi
+ fi
+ done
+
+ F16C_EXTRA_CFLAGS="$SSE_EXTRA_CFLAGS $f16c_flag"
+
+ { $as_echo "$as_me:${as_lineno-$LINENO}: checking whether we can compile half-floating point code" >&5
+$as_echo_n "checking whether we can compile half-floating point code... " >&6; }
+
+ CFLAGS="$CFLAGS $sse_flag $f16c_flag"
+
+ cat confdefs.h - <<_ACEOF >conftest.$ac_ext
+/* end confdefs.h. */
+#include <immintrin.h>
+int
+main ()
+{
+__m128 val = _mm_cvtph_ps ((__m128i)_mm_setzero_ps());
+ ;
+ return 0;
+}
+_ACEOF
+if ac_fn_c_try_compile "$LINENO"; then :
+
+$as_echo "#define USE_F16C 1" >>confdefs.h
+
+ { $as_echo "$as_me:${as_lineno-$LINENO}: result: yes" >&5
+$as_echo "yes" >&6; }
+
+else
+ enable_f16c=no
+ { $as_echo "$as_me:${as_lineno-$LINENO}: result: no" >&5
+$as_echo "no" >&6; }
+ { $as_echo "$as_me:${as_lineno-$LINENO}: WARNING: The compiler does not support f16c intrinsics." >&5
+$as_echo "$as_me: WARNING: The compiler does not support f16c intrinsics." >&2;}
+
fi
rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
fi
+
+
fi
# report actual input values of CONFIG_FILES etc. instead of their
# values after options handling.
ac_log="
-This file was extended by babl $as_me 0.1.14, which was
+This file was extended by babl $as_me 0.1.16, which was
generated by GNU Autoconf 2.69. Invocation command line was
CONFIG_FILES = $CONFIG_FILES
cat >>$CONFIG_STATUS <<_ACEOF || ac_write_fail=1
ac_cs_config="`$as_echo "$ac_configure_args" | sed 's/^ //; s/[\\""\`\$]/\\\\&/g'`"
ac_cs_version="\\
-babl config.status 0.1.14
+babl config.status 0.1.16
configured by $0, generated by GNU Autoconf 2.69,
with options \\"\$ac_cs_config\\"
m4_define([babl_major_version], [0])
m4_define([babl_minor_version], [1])
-m4_define([babl_micro_version], [14])
+m4_define([babl_micro_version], [16])
m4_define([babl_real_version],
[babl_major_version.babl_minor_version.babl_micro_version])
m4_define([babl_version], [babl_real_version])
[ --enable-sse2 enable SSE2 support (default=auto)],,
enable_sse2=$enable_sse)
+AC_ARG_ENABLE(sse4_1,
+ [ --enable-sse4_1 enable SSE4_1 support (default=auto)],,
+ enable_sse4_1=$enable_sse)
+
+AC_ARG_ENABLE(f16c,
+ [ --enable-f16c enable hardware half-float support (default=auto)],,
+ enable_f16c=$enable_sse)
+
if test "x$enable_mmx" = xyes; then
BABL_DETECT_CFLAGS(MMX_EXTRA_CFLAGS, '-mmmx')
SSE_EXTRA_CFLAGS=
SSE2_EXTRA_CFLAGS=
+ SSE4_1_EXTRA_CFLAGS=
AC_MSG_CHECKING(whether we can compile MMX code)
AC_MSG_RESULT(no)
AC_MSG_WARN([The assembler does not support the SSE2 command set.])
)
+
+ if test "x$enable_sse4_1" = xyes; then
+ BABL_DETECT_CFLAGS(sse4_1_flag, '-msse4.1')
+ SSE4_1_EXTRA_CFLAGS="$SSE_EXTRA_CFLAGS $sse4_1_flag"
+
+ AC_MSG_CHECKING(whether we can compile SSE4_1 code)
+
+ CFLAGS="$CFLAGS $sse4_1_flag"
+
+ AC_COMPILE_IFELSE([AC_LANG_PROGRAM(,[asm ("pmovzxbd %xmm0,%xmm1");])],
+ AC_DEFINE(USE_SSE4_1, 1, [Define to 1 if SSE4_1 assembly is available.])
+ AC_MSG_RESULT(yes)
+ ,
+ enable_sse4_1=no
+ AC_MSG_RESULT(no)
+ AC_MSG_WARN([The assembler does not support the SSE4_1 command set.])
+ )
+ fi
+ fi
+
+ CFLAGS="$mmx_save_CFLAGS"
+
+ if test "x$enable_f16c" = xyes; then
+ BABL_DETECT_CFLAGS(f16c_flag, '-mf16c')
+ F16C_EXTRA_CFLAGS="$SSE_EXTRA_CFLAGS $f16c_flag"
+
+ AC_MSG_CHECKING(whether we can compile half-floating point code)
+
+ CFLAGS="$CFLAGS $sse_flag $f16c_flag"
+
+ AC_COMPILE_IFELSE([AC_LANG_PROGRAM([#include <immintrin.h>],[__m128 val = _mm_cvtph_ps ((__m128i)_mm_setzero_ps());])],
+ AC_DEFINE(USE_F16C, 1, [Define to 1 if f16c intrinsics are available.])
+ AC_MSG_RESULT(yes)
+ ,
+ enable_f16c=no
+ AC_MSG_RESULT(no)
+ AC_MSG_WARN([The compiler does not support f16c intrinsics.])
+ )
fi
fi
AC_SUBST(MMX_EXTRA_CFLAGS)
AC_SUBST(SSE_EXTRA_CFLAGS)
AC_SUBST(SSE2_EXTRA_CFLAGS)
+ AC_SUBST(SSE4_1_EXTRA_CFLAGS)
+ AC_SUBST(F16C_EXTRA_CFLAGS)
fi
ECHO_T = @ECHO_T@
EGREP = @EGREP@
EXEEXT = @EXEEXT@
+F16C_EXTRA_CFLAGS = @F16C_EXTRA_CFLAGS@
FGREP = @FGREP@
GREP = @GREP@
INSTALL = @INSTALL@
SHELL = @SHELL@
SHREXT = @SHREXT@
SSE2_EXTRA_CFLAGS = @SSE2_EXTRA_CFLAGS@
+SSE4_1_EXTRA_CFLAGS = @SSE4_1_EXTRA_CFLAGS@
SSE_EXTRA_CFLAGS = @SSE_EXTRA_CFLAGS@
STRIP = @STRIP@
THREAD_LIB = @THREAD_LIB@
ECHO_T = @ECHO_T@
EGREP = @EGREP@
EXEEXT = @EXEEXT@
+F16C_EXTRA_CFLAGS = @F16C_EXTRA_CFLAGS@
FGREP = @FGREP@
GREP = @GREP@
INSTALL = @INSTALL@
SHELL = @SHELL@
SHREXT = @SHREXT@
SSE2_EXTRA_CFLAGS = @SSE2_EXTRA_CFLAGS@
+SSE4_1_EXTRA_CFLAGS = @SSE4_1_EXTRA_CFLAGS@
SSE_EXTRA_CFLAGS = @SSE_EXTRA_CFLAGS@
STRIP = @STRIP@
THREAD_LIB = @THREAD_LIB@
return samples;
}
+static long
+rgbf_to_Labf (float *src,
+ float *dst,
+ long samples)
+{
+ long n = samples;
+
+ while (n--)
+ {
+ float r = src[0];
+ float g = src[1];
+ float b = src[2];
+
+ float xr = 0.43603516f / D50_WHITE_REF_X * r + 0.38511658f / D50_WHITE_REF_X * g + 0.14305115f / D50_WHITE_REF_X * b;
+ float yr = 0.22248840f / D50_WHITE_REF_Y * r + 0.71690369f / D50_WHITE_REF_Y * g + 0.06060791f / D50_WHITE_REF_Y * b;
+ float zr = 0.01391602f / D50_WHITE_REF_Z * r + 0.09706116f / D50_WHITE_REF_Z * g + 0.71392822f / D50_WHITE_REF_Z * b;
+
+ float fx = xr > LAB_EPSILON ? cbrtf (xr) : (LAB_KAPPA * xr + 16.0f) / 116.0f;
+ float fy = yr > LAB_EPSILON ? cbrtf (yr) : (LAB_KAPPA * yr + 16.0f) / 116.0f;
+ float fz = zr > LAB_EPSILON ? cbrtf (zr) : (LAB_KAPPA * zr + 16.0f) / 116.0f;
+
+ float L = 116.0f * fy - 16.0f;
+ float A = 500.0f * (fx - fy);
+ float B = 200.0f * (fy - fz);
+
+ dst[0] = L;
+ dst[1] = A;
+ dst[2] = B;
+
+ src += 3;
+ dst += 3;
+ }
+
+ return samples;
+}
+
static long
rgbaf_to_Labaf (float *src,
float *dst,
return samples;
}
+static long
+Labf_to_rgbf (float *src,
+ float *dst,
+ long samples)
+{
+ long n = samples;
+
+ while (n--)
+ {
+ float L = src[0];
+ float A = src[1];
+ float B = src[2];
+
+ float fy = (L + 16.0f) / 116.0f;
+ float fx = fy + A / 500.0f;
+ float fz = fy - B / 200.0f;
+
+ float yr = L > LAB_KAPPA * LAB_EPSILON ? cubef (fy) : L / LAB_KAPPA;
+ float xr = cubef (fx) > LAB_EPSILON ? cubef (fx) : (fx * 116.0f - 16.0f) / LAB_KAPPA;
+ float zr = cubef (fz) > LAB_EPSILON ? cubef (fz) : (fz * 116.0f - 16.0f) / LAB_KAPPA;
+
+ float r = 3.134274799724f * D50_WHITE_REF_X * xr -1.617275708956f * D50_WHITE_REF_Y * yr -0.490724283042f * D50_WHITE_REF_Z * zr;
+ float g = -0.978795575994f * D50_WHITE_REF_X * xr +1.916161689117f * D50_WHITE_REF_Y * yr +0.033453331711f * D50_WHITE_REF_Z * zr;
+ float b = 0.071976988401f * D50_WHITE_REF_X * xr -0.228984974402f * D50_WHITE_REF_Y * yr +1.405718224383f * D50_WHITE_REF_Z * zr;
+
+ dst[0] = r;
+ dst[1] = g;
+ dst[2] = b;
+
+ src += 3;
+ dst += 3;
+ }
+
+ return samples;
+}
+
static long
Labaf_to_rgbaf (float *src,
float *dst,
"linear", laba_to_rgba,
NULL
);
+ babl_conversion_new (
+ babl_format ("RGB float"),
+ babl_format ("CIE Lab float"),
+ "linear", rgbf_to_Labf,
+ NULL
+ );
+ babl_conversion_new (
+ babl_format ("CIE Lab float"),
+ babl_format ("RGB float"),
+ "linear", Labf_to_rgbf,
+ NULL
+ );
babl_conversion_new (
babl_format ("RGBA float"),
babl_format ("CIE Lab alpha float"),
ext_LTLIBRARIES = \
cairo.la \
CIE.la \
+ float-half.la \
gegl-fixups.la \
gggl-lies.la \
+ gggl-table.la \
+ gggl-table-lies.la \
gggl.la \
gimp-8bit.la \
grey.la \
sse2-float.la \
sse2-int8.la \
sse2-int16.la \
+ sse4-int8.la \
+ sse-half.la \
two-table.la \
ycbcr.la
cairo_la_SOURCES = cairo.c cairo-tables.h
CIE_la_SOURCES = CIE.c
simple_la_SOURCES = simple.c
+float_half_la_SOURCES = float-half.c
gegl_fixups_la_SOURCES = gegl-fixups.c
gggl_lies_la_SOURCES = gggl-lies.c
+gggl_table_lies_la_SOURCES = gggl-table-lies.c
+gggl_table_la_SOURCES = gggl-table.c
gggl_la_SOURCES = gggl.c
gimp_8bit_la_SOURCES = gimp-8bit.c
grey_la_SOURCES = grey.c
sse2_float_la_SOURCES = sse2-float.c
sse2_int8_la_SOURCES = sse2-int8.c
sse2_int16_la_SOURCES = sse2-int16.c
+sse4_int8_la_SOURCES = sse4-int8.c
+sse_half_la_SOURCES = sse-half.c
two_table_la_SOURCES = two-table.c two-table-tables.h
ycbcr_la_SOURCES = ycbcr.c
float_la_SOURCES = float.c
sse2_float_la_CFLAGS = $(SSE2_EXTRA_CFLAGS)
sse2_int8_la_CFLAGS = $(SSE2_EXTRA_CFLAGS)
sse2_int16_la_CFLAGS = $(SSE2_EXTRA_CFLAGS)
+sse4_int8_la_CFLAGS = $(SSE4_1_EXTRA_CFLAGS)
+sse_half_la_CFLAGS = $(SSE4_1_EXTRA_CFLAGS) $(F16C_EXTRA_CFLAGS)
fast_float_la_LIBADD =
am_fast_float_la_OBJECTS = fast-float.lo
fast_float_la_OBJECTS = $(am_fast_float_la_OBJECTS)
+float_half_la_LIBADD =
+am_float_half_la_OBJECTS = float-half.lo
+float_half_la_OBJECTS = $(am_float_half_la_OBJECTS)
float_la_LIBADD =
am_float_la_OBJECTS = float.lo
float_la_OBJECTS = $(am_float_la_OBJECTS)
gggl_lies_la_LIBADD =
am_gggl_lies_la_OBJECTS = gggl-lies.lo
gggl_lies_la_OBJECTS = $(am_gggl_lies_la_OBJECTS)
+gggl_table_lies_la_LIBADD =
+am_gggl_table_lies_la_OBJECTS = gggl-table-lies.lo
+gggl_table_lies_la_OBJECTS = $(am_gggl_table_lies_la_OBJECTS)
+gggl_table_la_LIBADD =
+am_gggl_table_la_OBJECTS = gggl-table.lo
+gggl_table_la_OBJECTS = $(am_gggl_table_la_OBJECTS)
gggl_la_LIBADD =
am_gggl_la_OBJECTS = gggl.lo
gggl_la_OBJECTS = $(am_gggl_la_OBJECTS)
simple_la_LIBADD =
am_simple_la_OBJECTS = simple.lo
simple_la_OBJECTS = $(am_simple_la_OBJECTS)
+sse_half_la_LIBADD =
+am_sse_half_la_OBJECTS = sse_half_la-sse-half.lo
+sse_half_la_OBJECTS = $(am_sse_half_la_OBJECTS)
+sse_half_la_LINK = $(LIBTOOL) $(AM_V_lt) --tag=CC $(AM_LIBTOOLFLAGS) \
+ $(LIBTOOLFLAGS) --mode=link $(CCLD) $(sse_half_la_CFLAGS) \
+ $(CFLAGS) $(AM_LDFLAGS) $(LDFLAGS) -o $@
sse2_float_la_LIBADD =
am_sse2_float_la_OBJECTS = sse2_float_la-sse2-float.lo
sse2_float_la_OBJECTS = $(am_sse2_float_la_OBJECTS)
sse2_int8_la_LINK = $(LIBTOOL) $(AM_V_lt) --tag=CC $(AM_LIBTOOLFLAGS) \
$(LIBTOOLFLAGS) --mode=link $(CCLD) $(sse2_int8_la_CFLAGS) \
$(CFLAGS) $(AM_LDFLAGS) $(LDFLAGS) -o $@
+sse4_int8_la_LIBADD =
+am_sse4_int8_la_OBJECTS = sse4_int8_la-sse4-int8.lo
+sse4_int8_la_OBJECTS = $(am_sse4_int8_la_OBJECTS)
+sse4_int8_la_LINK = $(LIBTOOL) $(AM_V_lt) --tag=CC $(AM_LIBTOOLFLAGS) \
+ $(LIBTOOLFLAGS) --mode=link $(CCLD) $(sse4_int8_la_CFLAGS) \
+ $(CFLAGS) $(AM_LDFLAGS) $(LDFLAGS) -o $@
two_table_la_LIBADD =
am_two_table_la_OBJECTS = two-table.lo
two_table_la_OBJECTS = $(am_two_table_la_OBJECTS)
am__v_CCLD_1 =
SOURCES = $(CIE_la_SOURCES) $(HSL_la_SOURCES) $(HSV_la_SOURCES) \
$(cairo_la_SOURCES) $(fast_float_la_SOURCES) \
- $(float_la_SOURCES) $(gegl_fixups_la_SOURCES) \
- $(gggl_lies_la_SOURCES) $(gggl_la_SOURCES) \
- $(gimp_8bit_la_SOURCES) $(grey_la_SOURCES) \
+ $(float_half_la_SOURCES) $(float_la_SOURCES) \
+ $(gegl_fixups_la_SOURCES) $(gggl_lies_la_SOURCES) \
+ $(gggl_table_lies_la_SOURCES) $(gggl_table_la_SOURCES) \
+ $(gggl_la_SOURCES) $(gimp_8bit_la_SOURCES) $(grey_la_SOURCES) \
$(naive_CMYK_la_SOURCES) $(simple_la_SOURCES) \
- $(sse2_float_la_SOURCES) $(sse2_int16_la_SOURCES) \
- $(sse2_int8_la_SOURCES) $(two_table_la_SOURCES) \
+ $(sse_half_la_SOURCES) $(sse2_float_la_SOURCES) \
+ $(sse2_int16_la_SOURCES) $(sse2_int8_la_SOURCES) \
+ $(sse4_int8_la_SOURCES) $(two_table_la_SOURCES) \
$(ycbcr_la_SOURCES)
DIST_SOURCES = $(CIE_la_SOURCES) $(HSL_la_SOURCES) $(HSV_la_SOURCES) \
$(cairo_la_SOURCES) $(fast_float_la_SOURCES) \
- $(float_la_SOURCES) $(gegl_fixups_la_SOURCES) \
- $(gggl_lies_la_SOURCES) $(gggl_la_SOURCES) \
- $(gimp_8bit_la_SOURCES) $(grey_la_SOURCES) \
+ $(float_half_la_SOURCES) $(float_la_SOURCES) \
+ $(gegl_fixups_la_SOURCES) $(gggl_lies_la_SOURCES) \
+ $(gggl_table_lies_la_SOURCES) $(gggl_table_la_SOURCES) \
+ $(gggl_la_SOURCES) $(gimp_8bit_la_SOURCES) $(grey_la_SOURCES) \
$(naive_CMYK_la_SOURCES) $(simple_la_SOURCES) \
- $(sse2_float_la_SOURCES) $(sse2_int16_la_SOURCES) \
- $(sse2_int8_la_SOURCES) $(two_table_la_SOURCES) \
+ $(sse_half_la_SOURCES) $(sse2_float_la_SOURCES) \
+ $(sse2_int16_la_SOURCES) $(sse2_int8_la_SOURCES) \
+ $(sse4_int8_la_SOURCES) $(two_table_la_SOURCES) \
$(ycbcr_la_SOURCES)
am__can_run_installinfo = \
case $$AM_UPDATE_INFO_DIR in \
ECHO_T = @ECHO_T@
EGREP = @EGREP@
EXEEXT = @EXEEXT@
+F16C_EXTRA_CFLAGS = @F16C_EXTRA_CFLAGS@
FGREP = @FGREP@
GREP = @GREP@
INSTALL = @INSTALL@
SHELL = @SHELL@
SHREXT = @SHREXT@
SSE2_EXTRA_CFLAGS = @SSE2_EXTRA_CFLAGS@
+SSE4_1_EXTRA_CFLAGS = @SSE4_1_EXTRA_CFLAGS@
SSE_EXTRA_CFLAGS = @SSE_EXTRA_CFLAGS@
STRIP = @STRIP@
THREAD_LIB = @THREAD_LIB@
ext_LTLIBRARIES = \
cairo.la \
CIE.la \
+ float-half.la \
gegl-fixups.la \
gggl-lies.la \
+ gggl-table.la \
+ gggl-table-lies.la \
gggl.la \
gimp-8bit.la \
grey.la \
sse2-float.la \
sse2-int8.la \
sse2-int16.la \
+ sse4-int8.la \
+ sse-half.la \
two-table.la \
ycbcr.la
cairo_la_SOURCES = cairo.c cairo-tables.h
CIE_la_SOURCES = CIE.c
simple_la_SOURCES = simple.c
+float_half_la_SOURCES = float-half.c
gegl_fixups_la_SOURCES = gegl-fixups.c
gggl_lies_la_SOURCES = gggl-lies.c
+gggl_table_lies_la_SOURCES = gggl-table-lies.c
+gggl_table_la_SOURCES = gggl-table.c
gggl_la_SOURCES = gggl.c
gimp_8bit_la_SOURCES = gimp-8bit.c
grey_la_SOURCES = grey.c
sse2_float_la_SOURCES = sse2-float.c
sse2_int8_la_SOURCES = sse2-int8.c
sse2_int16_la_SOURCES = sse2-int16.c
+sse4_int8_la_SOURCES = sse4-int8.c
+sse_half_la_SOURCES = sse-half.c
two_table_la_SOURCES = two-table.c two-table-tables.h
ycbcr_la_SOURCES = ycbcr.c
float_la_SOURCES = float.c
sse2_float_la_CFLAGS = $(SSE2_EXTRA_CFLAGS)
sse2_int8_la_CFLAGS = $(SSE2_EXTRA_CFLAGS)
sse2_int16_la_CFLAGS = $(SSE2_EXTRA_CFLAGS)
+sse4_int8_la_CFLAGS = $(SSE4_1_EXTRA_CFLAGS)
+sse_half_la_CFLAGS = $(SSE4_1_EXTRA_CFLAGS) $(F16C_EXTRA_CFLAGS)
all: all-am
.SUFFIXES:
fast-float.la: $(fast_float_la_OBJECTS) $(fast_float_la_DEPENDENCIES) $(EXTRA_fast_float_la_DEPENDENCIES)
$(AM_V_CCLD)$(LINK) -rpath $(extdir) $(fast_float_la_OBJECTS) $(fast_float_la_LIBADD) $(LIBS)
+float-half.la: $(float_half_la_OBJECTS) $(float_half_la_DEPENDENCIES) $(EXTRA_float_half_la_DEPENDENCIES)
+ $(AM_V_CCLD)$(LINK) -rpath $(extdir) $(float_half_la_OBJECTS) $(float_half_la_LIBADD) $(LIBS)
+
float.la: $(float_la_OBJECTS) $(float_la_DEPENDENCIES) $(EXTRA_float_la_DEPENDENCIES)
$(AM_V_CCLD)$(LINK) -rpath $(extdir) $(float_la_OBJECTS) $(float_la_LIBADD) $(LIBS)
gggl-lies.la: $(gggl_lies_la_OBJECTS) $(gggl_lies_la_DEPENDENCIES) $(EXTRA_gggl_lies_la_DEPENDENCIES)
$(AM_V_CCLD)$(LINK) -rpath $(extdir) $(gggl_lies_la_OBJECTS) $(gggl_lies_la_LIBADD) $(LIBS)
+gggl-table-lies.la: $(gggl_table_lies_la_OBJECTS) $(gggl_table_lies_la_DEPENDENCIES) $(EXTRA_gggl_table_lies_la_DEPENDENCIES)
+ $(AM_V_CCLD)$(LINK) -rpath $(extdir) $(gggl_table_lies_la_OBJECTS) $(gggl_table_lies_la_LIBADD) $(LIBS)
+
+gggl-table.la: $(gggl_table_la_OBJECTS) $(gggl_table_la_DEPENDENCIES) $(EXTRA_gggl_table_la_DEPENDENCIES)
+ $(AM_V_CCLD)$(LINK) -rpath $(extdir) $(gggl_table_la_OBJECTS) $(gggl_table_la_LIBADD) $(LIBS)
+
gggl.la: $(gggl_la_OBJECTS) $(gggl_la_DEPENDENCIES) $(EXTRA_gggl_la_DEPENDENCIES)
$(AM_V_CCLD)$(LINK) -rpath $(extdir) $(gggl_la_OBJECTS) $(gggl_la_LIBADD) $(LIBS)
simple.la: $(simple_la_OBJECTS) $(simple_la_DEPENDENCIES) $(EXTRA_simple_la_DEPENDENCIES)
$(AM_V_CCLD)$(LINK) -rpath $(extdir) $(simple_la_OBJECTS) $(simple_la_LIBADD) $(LIBS)
+sse-half.la: $(sse_half_la_OBJECTS) $(sse_half_la_DEPENDENCIES) $(EXTRA_sse_half_la_DEPENDENCIES)
+ $(AM_V_CCLD)$(sse_half_la_LINK) -rpath $(extdir) $(sse_half_la_OBJECTS) $(sse_half_la_LIBADD) $(LIBS)
+
sse2-float.la: $(sse2_float_la_OBJECTS) $(sse2_float_la_DEPENDENCIES) $(EXTRA_sse2_float_la_DEPENDENCIES)
$(AM_V_CCLD)$(sse2_float_la_LINK) -rpath $(extdir) $(sse2_float_la_OBJECTS) $(sse2_float_la_LIBADD) $(LIBS)
sse2-int8.la: $(sse2_int8_la_OBJECTS) $(sse2_int8_la_DEPENDENCIES) $(EXTRA_sse2_int8_la_DEPENDENCIES)
$(AM_V_CCLD)$(sse2_int8_la_LINK) -rpath $(extdir) $(sse2_int8_la_OBJECTS) $(sse2_int8_la_LIBADD) $(LIBS)
+sse4-int8.la: $(sse4_int8_la_OBJECTS) $(sse4_int8_la_DEPENDENCIES) $(EXTRA_sse4_int8_la_DEPENDENCIES)
+ $(AM_V_CCLD)$(sse4_int8_la_LINK) -rpath $(extdir) $(sse4_int8_la_OBJECTS) $(sse4_int8_la_LIBADD) $(LIBS)
+
two-table.la: $(two_table_la_OBJECTS) $(two_table_la_DEPENDENCIES) $(EXTRA_two_table_la_DEPENDENCIES)
$(AM_V_CCLD)$(LINK) -rpath $(extdir) $(two_table_la_OBJECTS) $(two_table_la_LIBADD) $(LIBS)
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/HSV.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/cairo.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/fast-float.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/float-half.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/float.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/gegl-fixups.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/gggl-lies.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/gggl-table-lies.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/gggl-table.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/gggl.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/gimp-8bit.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/grey.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/sse2_float_la-sse2-float.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/sse2_int16_la-sse2-int16.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/sse2_int8_la-sse2-int8.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/sse4_int8_la-sse4-int8.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/sse_half_la-sse-half.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/two-table.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/ycbcr.Plo@am__quote@
@AMDEP_TRUE@@am__fastdepCC_FALSE@ DEPDIR=$(DEPDIR) $(CCDEPMODE) $(depcomp) @AMDEPBACKSLASH@
@am__fastdepCC_FALSE@ $(AM_V_CC@am__nodep@)$(LTCOMPILE) -c -o $@ $<
+sse_half_la-sse-half.lo: sse-half.c
+@am__fastdepCC_TRUE@ $(AM_V_CC)$(LIBTOOL) $(AM_V_lt) --tag=CC $(AM_LIBTOOLFLAGS) $(LIBTOOLFLAGS) --mode=compile $(CC) $(DEFS) $(DEFAULT_INCLUDES) $(INCLUDES) $(AM_CPPFLAGS) $(CPPFLAGS) $(sse_half_la_CFLAGS) $(CFLAGS) -MT sse_half_la-sse-half.lo -MD -MP -MF $(DEPDIR)/sse_half_la-sse-half.Tpo -c -o sse_half_la-sse-half.lo `test -f 'sse-half.c' || echo '$(srcdir)/'`sse-half.c
+@am__fastdepCC_TRUE@ $(AM_V_at)$(am__mv) $(DEPDIR)/sse_half_la-sse-half.Tpo $(DEPDIR)/sse_half_la-sse-half.Plo
+@AMDEP_TRUE@@am__fastdepCC_FALSE@ $(AM_V_CC)source='sse-half.c' object='sse_half_la-sse-half.lo' libtool=yes @AMDEPBACKSLASH@
+@AMDEP_TRUE@@am__fastdepCC_FALSE@ DEPDIR=$(DEPDIR) $(CCDEPMODE) $(depcomp) @AMDEPBACKSLASH@
+@am__fastdepCC_FALSE@ $(AM_V_CC@am__nodep@)$(LIBTOOL) $(AM_V_lt) --tag=CC $(AM_LIBTOOLFLAGS) $(LIBTOOLFLAGS) --mode=compile $(CC) $(DEFS) $(DEFAULT_INCLUDES) $(INCLUDES) $(AM_CPPFLAGS) $(CPPFLAGS) $(sse_half_la_CFLAGS) $(CFLAGS) -c -o sse_half_la-sse-half.lo `test -f 'sse-half.c' || echo '$(srcdir)/'`sse-half.c
+
sse2_float_la-sse2-float.lo: sse2-float.c
@am__fastdepCC_TRUE@ $(AM_V_CC)$(LIBTOOL) $(AM_V_lt) --tag=CC $(AM_LIBTOOLFLAGS) $(LIBTOOLFLAGS) --mode=compile $(CC) $(DEFS) $(DEFAULT_INCLUDES) $(INCLUDES) $(AM_CPPFLAGS) $(CPPFLAGS) $(sse2_float_la_CFLAGS) $(CFLAGS) -MT sse2_float_la-sse2-float.lo -MD -MP -MF $(DEPDIR)/sse2_float_la-sse2-float.Tpo -c -o sse2_float_la-sse2-float.lo `test -f 'sse2-float.c' || echo '$(srcdir)/'`sse2-float.c
@am__fastdepCC_TRUE@ $(AM_V_at)$(am__mv) $(DEPDIR)/sse2_float_la-sse2-float.Tpo $(DEPDIR)/sse2_float_la-sse2-float.Plo
@AMDEP_TRUE@@am__fastdepCC_FALSE@ DEPDIR=$(DEPDIR) $(CCDEPMODE) $(depcomp) @AMDEPBACKSLASH@
@am__fastdepCC_FALSE@ $(AM_V_CC@am__nodep@)$(LIBTOOL) $(AM_V_lt) --tag=CC $(AM_LIBTOOLFLAGS) $(LIBTOOLFLAGS) --mode=compile $(CC) $(DEFS) $(DEFAULT_INCLUDES) $(INCLUDES) $(AM_CPPFLAGS) $(CPPFLAGS) $(sse2_int8_la_CFLAGS) $(CFLAGS) -c -o sse2_int8_la-sse2-int8.lo `test -f 'sse2-int8.c' || echo '$(srcdir)/'`sse2-int8.c
+sse4_int8_la-sse4-int8.lo: sse4-int8.c
+@am__fastdepCC_TRUE@ $(AM_V_CC)$(LIBTOOL) $(AM_V_lt) --tag=CC $(AM_LIBTOOLFLAGS) $(LIBTOOLFLAGS) --mode=compile $(CC) $(DEFS) $(DEFAULT_INCLUDES) $(INCLUDES) $(AM_CPPFLAGS) $(CPPFLAGS) $(sse4_int8_la_CFLAGS) $(CFLAGS) -MT sse4_int8_la-sse4-int8.lo -MD -MP -MF $(DEPDIR)/sse4_int8_la-sse4-int8.Tpo -c -o sse4_int8_la-sse4-int8.lo `test -f 'sse4-int8.c' || echo '$(srcdir)/'`sse4-int8.c
+@am__fastdepCC_TRUE@ $(AM_V_at)$(am__mv) $(DEPDIR)/sse4_int8_la-sse4-int8.Tpo $(DEPDIR)/sse4_int8_la-sse4-int8.Plo
+@AMDEP_TRUE@@am__fastdepCC_FALSE@ $(AM_V_CC)source='sse4-int8.c' object='sse4_int8_la-sse4-int8.lo' libtool=yes @AMDEPBACKSLASH@
+@AMDEP_TRUE@@am__fastdepCC_FALSE@ DEPDIR=$(DEPDIR) $(CCDEPMODE) $(depcomp) @AMDEPBACKSLASH@
+@am__fastdepCC_FALSE@ $(AM_V_CC@am__nodep@)$(LIBTOOL) $(AM_V_lt) --tag=CC $(AM_LIBTOOLFLAGS) $(LIBTOOLFLAGS) --mode=compile $(CC) $(DEFS) $(DEFAULT_INCLUDES) $(INCLUDES) $(AM_CPPFLAGS) $(CPPFLAGS) $(sse4_int8_la_CFLAGS) $(CFLAGS) -c -o sse4_int8_la-sse4-int8.lo `test -f 'sse4-int8.c' || echo '$(srcdir)/'`sse4-int8.c
+
mostlyclean-libtool:
-rm -f *.lo
--- /dev/null
+/* babl - dynamically extendable universal pixel conversion library.
+ * Copyright (C) 2015 Daniel Sabo
+ * 2016 Øyvind Kolås
+ *
+ * This library 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 3 of the License, or (at your option) any later version.
+ *
+ * This library 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 this library; if not, see
+ * <http://www.gnu.org/licenses/>.
+ */
+
+/* Copyright: (c) 2009 by James Tursa, All Rights Reserved
+ *
+ * This code uses the BSD License:
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are
+ * met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in
+ * the documentation and/or other materials provided with the distribution
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ *
+ * halfprecision converts the input argument to/from a half precision floating
+ * point bit pattern corresponding to IEEE 754r. The bit pattern is stored in a
+ * uint16 class variable. Please note that halfprecision is *not* a class. That
+ * is, you cannot do any arithmetic with the half precision bit patterns.
+ * halfprecision is simply a function that converts the IEEE 754r half precision
+ * bit pattern to/from other numeric MATLAB variables. You can, however, take
+ * the half precision bit patterns, convert them to single or double, do the
+ * operation, and then convert the result back manually.
+ *
+ * 1 bit sign bit
+ * 5 bits exponent, biased by 15
+ * 10 bits mantissa, hidden leading bit, normalized to 1.0
+ *
+ * Special floating point bit patterns recognized and supported:
+ *
+ * All exponent bits zero:
+ * - If all mantissa bits are zero, then number is zero (possibly signed)
+ * - Otherwise, number is a denormalized bit pattern
+ *
+ * All exponent bits set to 1:
+ * - If all mantissa bits are zero, then number is +Infinity or -Infinity
+ * - Otherwise, number is NaN (Not a Number)
+ */
+
+#include "config.h"
+
+#include <stdint.h>
+#include <stdlib.h>
+
+#include "babl.h"
+#include "extensions/util.h"
+
+static void halfp2singles(void *target, const void *source, long numel)
+{
+ uint16_t *hp = (uint16_t *) source; // Type pun input as an unsigned 16-bit int
+ uint32_t *xp = (uint32_t *) target; // Type pun output as an unsigned 32-bit int
+ uint16_t h, hs, he, hm;
+ uint32_t xs, xe, xm;
+ int32_t xes;
+ int e;
+
+ if( source == NULL || target == NULL ) // Nothing to convert (e.g., imag part of pure real)
+ return;
+ while( numel-- ) {
+ h = *hp++;
+ if( (h & 0x7FFFu) == 0 ) { // Signed zero
+ *xp++ = ((uint32_t) h) << 16; // Return the signed zero
+ } else { // Not zero
+ hs = h & 0x8000u; // Pick off sign bit
+ he = h & 0x7C00u; // Pick off exponent bits
+ hm = h & 0x03FFu; // Pick off mantissa bits
+ if( he == 0 ) { // Denormal will convert to normalized
+ e = -1; // The following loop figures out how much extra to adjust the exponent
+ do {
+ e++;
+ hm <<= 1;
+ } while( (hm & 0x0400u) == 0 ); // Shift until leading bit overflows into exponent bit
+ xs = ((uint32_t) hs) << 16; // Sign bit
+ xes = ((int32_t) (he >> 10)) - 15 + 127 - e; // Exponent unbias the halfp, then bias the single
+ xe = (uint32_t) (xes << 23); // Exponent
+ xm = ((uint32_t) (hm & 0x03FFu)) << 13; // Mantissa
+ *xp++ = (xs | xe | xm); // Combine sign bit, exponent bits, and mantissa bits
+ } else if( he == 0x7C00u ) { // Inf or NaN (all the exponent bits are set)
+ if( hm == 0 ) { // If mantissa is zero ...
+ *xp++ = (((uint32_t) hs) << 16) | ((uint32_t) 0x7F800000u); // Signed Inf
+ } else {
+ *xp++ = (uint32_t) 0xFFC00000u; // NaN, only 1st mantissa bit set
+ }
+ } else { // Normalized number
+ xs = ((uint32_t) hs) << 16; // Sign bit
+ xes = ((int32_t) (he >> 10)) - 15 + 127; // Exponent unbias the halfp, then bias the single
+ xe = (uint32_t) (xes << 23); // Exponent
+ xm = ((uint32_t) hm) << 13; // Mantissa
+ *xp++ = (xs | xe | xm); // Combine sign bit, exponent bits, and mantissa bits
+ }
+ }
+ }
+}
+
+static void singles2halfp(void *target, const void *source, long numel)
+{
+ uint16_t *hp = (uint16_t *) target; // Type pun output as an unsigned 16-bit int
+ uint32_t *xp = (uint32_t *) source; // Type pun input as an unsigned 32-bit int
+ uint16_t hs, he, hm;
+ uint32_t x, xs, xe, xm;
+ int hes;
+
+ if( source == NULL || target == NULL ) { // Nothing to convert (e.g., imag part of pure real)
+ return;
+ }
+ while( numel-- ) {
+ x = *xp++;
+ if( (x & 0x7FFFFFFFu) == 0 ) { // Signed zero
+ *hp++ = (uint16_t) (x >> 16); // Return the signed zero
+ } else { // Not zero
+ xs = x & 0x80000000u; // Pick off sign bit
+ xe = x & 0x7F800000u; // Pick off exponent bits
+ xm = x & 0x007FFFFFu; // Pick off mantissa bits
+ if( xe == 0 ) { // Denormal will underflow, return a signed zero
+ *hp++ = (uint16_t) (xs >> 16);
+ } else if( xe == 0x7F800000u ) { // Inf or NaN (all the exponent bits are set)
+ if( xm == 0 ) { // If mantissa is zero ...
+ *hp++ = (uint16_t) ((xs >> 16) | 0x7C00u); // Signed Inf
+ } else {
+ *hp++ = (uint16_t) 0xFE00u; // NaN, only 1st mantissa bit set
+ }
+ } else { // Normalized number
+ hs = (uint16_t) (xs >> 16); // Sign bit
+ hes = ((int)(xe >> 23)) - 127 + 15; // Exponent unbias the single, then bias the halfp
+ if( hes >= 0x1F ) { // Overflow
+ *hp++ = (uint16_t) ((xs >> 16) | 0x7C00u); // Signed Inf
+ } else if( hes <= 0 ) { // Underflow
+ if( (14 - hes) > 24 ) { // Mantissa shifted all the way off & no rounding possibility
+ hm = (uint16_t) 0u; // Set mantissa to zero
+ } else {
+ xm |= 0x00800000u; // Add the hidden leading bit
+ hm = (uint16_t) (xm >> (14 - hes)); // Mantissa
+ if( (xm >> (13 - hes)) & 0x00000001u ) // Check for rounding
+ hm += (uint16_t) 1u; // Round, might overflow into exp bit, but this is OK
+ }
+ *hp++ = (hs | hm); // Combine sign bit and mantissa bits, biased exponent is zero
+ } else {
+ he = (uint16_t) (hes << 10); // Exponent
+ hm = (uint16_t) (xm >> 13); // Mantissa
+ if( xm & 0x00001000u ) // Check for rounding
+ *hp++ = (hs | he | hm) + (uint16_t) 1u; // Round, might overflow to inf, this is OK
+ else
+ *hp++ = (hs | he | hm); // No rounding
+ }
+ }
+ }
+ }
+}
+
+static inline long
+conv_yHalf_yF (const uint16_t *src, float *dst, long samples)
+{
+ halfp2singles(dst, src, samples);
+ return samples;
+}
+
+static long
+conv_yaHalf_yaF (const uint16_t *src, float *dst, long samples)
+{
+ return conv_yHalf_yF (src, dst, samples * 2) / 2;
+}
+
+static long
+conv_rgbHalf_rgbF (const uint16_t *src, float *dst, long samples)
+{
+ return conv_yHalf_yF (src, dst, samples * 3) / 3;
+}
+
+static long
+conv_rgbaHalf_rgbaF (const uint16_t *src, float *dst, long samples)
+{
+ return conv_yHalf_yF (src, dst, samples * 4) / 4;
+}
+
+static inline long
+conv_yF_yHalf (const float *src, uint16_t *dst, long samples)
+{
+ singles2halfp (dst, src, samples);
+ return samples;
+}
+
+static long
+conv_yaF_yaHalf (const float *src, uint16_t *dst, long samples)
+{
+ return conv_yF_yHalf (src, dst, samples * 2) / 2;
+}
+
+static long
+conv_rgbF_rgbHalf (const float *src, uint16_t *dst, long samples)
+{
+ return conv_yF_yHalf (src, dst, samples * 3) / 3;
+}
+
+static long
+conv_rgbaF_rgbaHalf (const float *src, uint16_t *dst, long samples)
+{
+ return conv_yF_yHalf (src, dst, samples * 4) / 4;
+}
+
+int init (void);
+
+int
+init (void)
+{
+ const Babl *rgbaF_linear = babl_format_new (
+ babl_model ("RGBA"),
+ babl_type ("float"),
+ babl_component ("R"),
+ babl_component ("G"),
+ babl_component ("B"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgbaHalf_linear = babl_format_new (
+ babl_model ("RGBA"),
+ babl_type ("half"),
+ babl_component ("R"),
+ babl_component ("G"),
+ babl_component ("B"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgbaF_gamma = babl_format_new (
+ babl_model ("R'G'B'A"),
+ babl_type ("float"),
+ babl_component ("R'"),
+ babl_component ("G'"),
+ babl_component ("B'"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgbaHalf_gamma = babl_format_new (
+ babl_model ("R'G'B'A"),
+ babl_type ("half"),
+ babl_component ("R'"),
+ babl_component ("G'"),
+ babl_component ("B'"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgbF_linear = babl_format_new (
+ babl_model ("RGB"),
+ babl_type ("float"),
+ babl_component ("R"),
+ babl_component ("G"),
+ babl_component ("B"),
+ NULL);
+ const Babl *rgbHalf_linear = babl_format_new (
+ babl_model ("RGB"),
+ babl_type ("half"),
+ babl_component ("R"),
+ babl_component ("G"),
+ babl_component ("B"),
+ NULL);
+ const Babl *rgbF_gamma = babl_format_new (
+ babl_model ("R'G'B'"),
+ babl_type ("float"),
+ babl_component ("R'"),
+ babl_component ("G'"),
+ babl_component ("B'"),
+ NULL);
+ const Babl *rgbHalf_gamma = babl_format_new (
+ babl_model ("R'G'B'"),
+ babl_type ("half"),
+ babl_component ("R'"),
+ babl_component ("G'"),
+ babl_component ("B'"),
+ NULL);
+ const Babl *yaF_linear = babl_format_new (
+ babl_model ("YA"),
+ babl_type ("float"),
+ babl_component ("Y"),
+ babl_component ("A"),
+ NULL);
+ const Babl *yaHalf_linear = babl_format_new (
+ babl_model ("YA"),
+ babl_type ("half"),
+ babl_component ("Y"),
+ babl_component ("A"),
+ NULL);
+ const Babl *yaF_gamma = babl_format_new (
+ babl_model ("Y'A"),
+ babl_type ("float"),
+ babl_component ("Y'"),
+ babl_component ("A"),
+ NULL);
+ const Babl *yaHalf_gamma = babl_format_new (
+ babl_model ("Y'A"),
+ babl_type ("half"),
+ babl_component ("Y'"),
+ babl_component ("A"),
+ NULL);
+ const Babl *yF_linear = babl_format_new (
+ babl_model ("Y"),
+ babl_type ("float"),
+ babl_component ("Y"),
+ NULL);
+ const Babl *yHalf_linear = babl_format_new (
+ babl_model ("Y"),
+ babl_type ("half"),
+ babl_component ("Y"),
+ NULL);
+ const Babl *yF_gamma = babl_format_new (
+ babl_model ("Y'"),
+ babl_type ("float"),
+ babl_component ("Y'"),
+ NULL);
+ const Babl *yHalf_gamma = babl_format_new (
+ babl_model ("Y'"),
+ babl_type ("half"),
+ babl_component ("Y'"),
+ NULL);
+
+#define CONV(src, dst) \
+{ \
+ babl_conversion_new (src ## _linear, dst ## _linear, "linear", conv_ ## src ## _ ## dst, NULL); \
+ babl_conversion_new (src ## _gamma, dst ## _gamma, "linear", conv_ ## src ## _ ## dst, NULL); \
+}
+
+ CONV(rgbaHalf, rgbaF);
+ CONV(rgbHalf, rgbF);
+ CONV(yaHalf, yaF);
+ CONV(yHalf, yF);
+ CONV(rgbaF, rgbaHalf);
+ CONV(rgbF, rgbHalf);
+ CONV(yaF, yaHalf);
+ CONV(yF, yHalf);
+
+ return 0;
+}
* gamma correction (not really,. gamma correction belongs in seperate ops,.
*/
-#define USE_TABLES
-#ifdef USE_TABLES
-
-/* lookup tables used in conversion */
-
-static float table_8_F[1 << 8];
-static float table_16_F[1 << 16];
-static unsigned char table_F_8[1 << 16];
-static unsigned short table_F_16[1 << 16];
-
-
-static int table_inited = 0;
-
-static void
-table_init (void)
-{
- int i;
-
- if (table_inited)
- return;
- table_inited = 1;
-
- /* fill tables for conversion from integer to float */
- for (i = 0; i < 1 << 8; i++)
- {
- table_8_F[i] = (i * 1.0) / 255.0;
- }
- for (i = 0; i < 1 << 16; i++)
- {
- table_16_F[i] = (i * 1.0) / 65535.0;
- }
- /* fill tables for conversion from float to integer */
- {
- union
- {
- float f;
- unsigned short s[2];
- } u;
- u.f = 0.0;
-
- u.s[0] = 0x8000;
-
- for (i = 0; i < 1 << 16; i++)
- {
- unsigned char c;
- unsigned short s;
-
- u.s[1] = i;
-
- if (u.f <= 0.0)
- {
- c = 0;
- s = 0;
- }
- else if (u.f >= 1.0)
- {
- c = 255;
- s = 65535;
- }
- else
- {
- c = lrint (u.f * 255.0);
- s = lrint (u.f * 65535.0);
- }
-
- /*fprintf (stderr, "%2.3f=%03i %05i ", f, c, (*hi));
- / if (! ((*hi)%9))
- / fprintf (stderr, "\n"); */
-
- table_F_8[u.s[1]] = c;
- table_F_16[u.s[1]] = s;
- }
- }
- /* fix tables to ensure 1:1 conversions back and forth */
- if (0)
- { /*FIXME: probably not the right way to do it,.. must sit down and scribble on paper */
- int i;
- for (i = 0; i < 256; i++)
- {
- float f = table_8_F[i];
- unsigned short *hi = ((unsigned short *) (void *) &f);
- unsigned short *lo = ((unsigned short *) (void *) &f);
- *lo = 0;
- table_F_8[(*hi)] = i;
- }
- }
-}
-
-/* function to find the index in table for a float */
-static unsigned int
-gggl_float_to_index16 (float f)
-{
- union
- {
- float f;
- unsigned short s[2];
- } u;
- u.f = f;
- return u.s[1];
-}
-
-
-static long
-conv_F_8 (unsigned char *src, unsigned char *dst, long samples)
-{
- long n = samples;
-
- if (!table_inited)
- table_init ();
- while (n--)
- {
- register float f = (*(float *) src);
- *(unsigned char *) dst = table_F_8[gggl_float_to_index16 (f)];
- dst += 1;
- src += 4;
- }
- return samples;
-}
-
-static long
-conv_F_16 (unsigned char *src, unsigned char *dst, long samples)
-{
- long n = samples;
-
- if (!table_inited)
- table_init ();
- while (n--)
- {
- register float f = (*(float *) src);
- *(unsigned short *) dst = table_F_16[gggl_float_to_index16 (f)];
- dst += 2;
- src += 4;
- }
- return samples;
-}
-
-static long
-conv_8_F (unsigned char *src, unsigned char *dst, long samples)
-{
- long n = samples;
-
- if (!table_inited)
- table_init ();
- while (n--)
- {
- (*(float *) dst) = table_8_F[*(unsigned char *) src];
- dst += 4;
- src += 1;
- }
- return samples;
-}
-
-static long
-conv_16_F (unsigned char *src, unsigned char *dst, long samples)
-{
- long n = samples;
-
- if (!table_inited)
- table_init ();
- while (n--)
- {
- (*(float *) dst) = table_16_F[*(unsigned short *) src];
- dst += 4;
- src += 2;
- }
- return samples;
-}
-
-#else
-
static long
conv_F_8 (unsigned char *src, unsigned char *dst, long samples)
{
return samples;
}
-
-#endif
-
static long
conv_F_D (unsigned char *src, unsigned char *dst, long samples)
{
return samples;
}
-
static long
conv_D_F (unsigned char *src, unsigned char *dst, long samples)
{
o (rgba8, rgb8);
o (rgbaF, rgbA8);
-#ifdef USE_TABLES
- if (!table_inited)
- table_init ();
-#endif
-
return 0;
}
--- /dev/null
+/*
+ * This file was part of gggl, it implements a variety of pixel conversion
+ * functions that are usable with babl, the file needs more cleanup, and
+ * doesn't return the number of samples processed as a long, like it's
+ * supposed to.
+ *
+ * GGGL is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * GGGL is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with GGGL; if not, see <http://www.gnu.org/licenses/>.
+ *
+ * Rights are granted to use this shared object in libraries covered by
+ * LGPL. (exception added, during import into babl CVS.)
+ *
+ * Copyright 2003, 2004, 2005 Øyvind Kolås <pippin@gimp.org>
+ */
+
+#define _POSIX_C_SOURCE 200112L
+
+#include "config.h"
+#include <math.h>
+#include <string.h>
+
+#include "babl.h"
+#include "extensions/util.h"
+
+/*
+ * Implemented according to information read from:
+ *
+ * http://www.cinenet.net/~spitzak/conversion/sketches_0265.pdf
+ *
+ * initially ignoring any diffusion, to keep the implementation
+ * smaller, and interchangeable with the non optimized version.
+ *
+ * due to ability to be able to relicence gggl under a different
+ * licence than GPL, I avoided the temptation to look at the
+ * source files in the same location, in case I was going to
+ * need this piece of code for projects where GPL compatibility
+ * was a must.
+ *
+ * TODO: error diffusion,
+ * gamma correction (not really,. gamma correction belongs in seperate ops,.
+ */
+
+/* lookup tables used in conversion */
+
+static float table_8_F[1 << 8];
+static float table_16_F[1 << 16];
+static unsigned char table_F_8[1 << 16];
+static unsigned short table_F_16[1 << 16];
+
+
+static int table_inited = 0;
+
+static void
+table_init (void)
+{
+ int i;
+
+ if (table_inited)
+ return;
+ table_inited = 1;
+
+ /* fill tables for conversion from integer to float */
+ for (i = 0; i < 1 << 8; i++)
+ {
+ table_8_F[i] = (i * 1.0) / 255.0;
+ }
+ for (i = 0; i < 1 << 16; i++)
+ {
+ table_16_F[i] = (i * 1.0) / 65535.0;
+ }
+ /* fill tables for conversion from float to integer */
+ {
+ union
+ {
+ float f;
+ unsigned short s[2];
+ } u;
+ u.f = 0.0;
+
+ u.s[0] = 0x8000;
+
+ for (i = 0; i < 1 << 16; i++)
+ {
+ unsigned char c;
+ unsigned short s;
+
+ u.s[1] = i;
+
+ if (u.f <= 0.0)
+ {
+ c = 0;
+ s = 0;
+ }
+ else if (u.f >= 1.0)
+ {
+ c = 255;
+ s = 65535;
+ }
+ else
+ {
+ c = lrint (u.f * 255.0);
+ s = lrint (u.f * 65535.0);
+ }
+
+ /*fprintf (stderr, "%2.3f=%03i %05i ", f, c, (*hi));
+ / if (! ((*hi)%9))
+ / fprintf (stderr, "\n"); */
+
+ table_F_8[u.s[1]] = c;
+ table_F_16[u.s[1]] = s;
+ }
+ }
+ /* fix tables to ensure 1:1 conversions back and forth */
+ if (0)
+ { /*FIXME: probably not the right way to do it,.. must sit down and scribble on paper */
+ int i;
+ for (i = 0; i < 256; i++)
+ {
+ float f = table_8_F[i];
+ unsigned short *hi = ((unsigned short *) (void *) &f);
+ unsigned short *lo = ((unsigned short *) (void *) &f);
+ *lo = 0;
+ table_F_8[(*hi)] = i;
+ }
+ }
+}
+
+/* function to find the index in table for a float */
+static unsigned int
+gggl_float_to_index16 (float f)
+{
+ union
+ {
+ float f;
+ unsigned short s[2];
+ } u;
+ u.f = f;
+ return u.s[1];
+}
+
+
+static long
+conv_F_8 (unsigned char *src, unsigned char *dst, long samples)
+{
+ long n = samples;
+
+ if (!table_inited)
+ table_init ();
+ while (n--)
+ {
+ register float f = (*(float *) src);
+ *(unsigned char *) dst = table_F_8[gggl_float_to_index16 (f)];
+ dst += 1;
+ src += 4;
+ }
+ return samples;
+}
+
+static long
+conv_F_16 (unsigned char *src, unsigned char *dst, long samples)
+{
+ long n = samples;
+
+ if (!table_inited)
+ table_init ();
+ while (n--)
+ {
+ register float f = (*(float *) src);
+ *(unsigned short *) dst = table_F_16[gggl_float_to_index16 (f)];
+ dst += 2;
+ src += 4;
+ }
+ return samples;
+}
+
+static long
+conv_8_F (unsigned char *src, unsigned char *dst, long samples)
+{
+ long n = samples;
+
+ if (!table_inited)
+ table_init ();
+ while (n--)
+ {
+ (*(float *) dst) = table_8_F[*(unsigned char *) src];
+ dst += 4;
+ src += 1;
+ }
+ return samples;
+}
+
+static long
+conv_16_F (unsigned char *src, unsigned char *dst, long samples)
+{
+ long n = samples;
+
+ if (!table_inited)
+ table_init ();
+ while (n--)
+ {
+ (*(float *) dst) = table_16_F[*(unsigned short *) src];
+ dst += 4;
+ src += 2;
+ }
+ return samples;
+}
+
+/*********/
+static long
+conv_rgbaF_rgba8 (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_F_8 (src, dst, samples * 4);
+ return samples;
+}
+
+#define conv_rgbaF_rgbP8 conv_rgbaF_rgba8
+
+static long
+conv_rgbF_rgb8 (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_F_8 (src, dst, samples * 3);
+ return samples;
+}
+
+static long
+conv_gaF_ga8 (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_F_8 (src, dst, samples * 2);
+ return samples;
+}
+
+#define conv_rgbAF_rgbA8 conv_rgbaF_rgba8
+#define conv_gF_g8 conv_F_8
+#define conv_gAF_gA8 conv_gaF_ga8
+
+static long
+conv_rgbaF_rgba16 (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_F_16 (src, dst, samples * 4);
+ return samples;
+}
+
+static long
+conv_rgbF_rgb16 (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_F_16 (src, dst, samples * 3);
+ return samples;
+}
+
+static long
+conv_gaF_ga16 (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_F_16 (src, dst, samples * 2);
+ return samples;
+}
+
+#define conv_rgbAF_rgbA16 conv_rgbaF_rgba16
+#define conv_gF_g16 conv_F_16
+#define conv_gAF_gA16 conv_gaF_ga16
+
+static long
+conv_rgba8_rgbaF (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_8_F (src, dst, samples * 4);
+ return samples;
+}
+
+static long
+conv_rgb8_rgbF (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_8_F (src, dst, samples * 3);
+ return samples;
+}
+
+static long
+conv_ga8_gaF (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_8_F (src, dst, samples * 2);
+ return samples;
+}
+
+#define conv_rgbA8_rgbAF conv_rgba8_rgbaF
+#define conv_gA8_gAF conv_ga8_gaF
+#define conv_g8_gF conv_8_F
+
+static long
+conv_rgba16_rgbaF (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_16_F (src, dst, samples * 4);
+ return samples;
+}
+
+static long
+conv_rgb16_rgbF (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_16_F (src, dst, samples * 3);
+ return samples;
+}
+
+static long
+conv_ga16_gaF (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_16_F (src, dst, samples * 2);
+ return samples;
+}
+
+#define conv_rgbA16_rgbAF conv_rgba16_rgbaF
+#define conv_gA16_gAF conv_ga16_gaF
+#define conv_g16_gF conv_16_F
+
+int init (void);
+
+int
+init (void)
+{
+ const Babl *rgbaF = babl_format_new (
+ babl_model ("RGBA"),
+ babl_type ("float"),
+ babl_component ("R"),
+ babl_component ("G"),
+ babl_component ("B"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgba16 = babl_format_new (
+ babl_model ("RGBA"),
+ babl_type ("u16"),
+ babl_component ("R"),
+ babl_component ("G"),
+ babl_component ("B"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgba8 = babl_format_new (
+ babl_model ("RGBA"),
+ babl_type ("u8"),
+ babl_component ("R"),
+ babl_component ("G"),
+ babl_component ("B"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgbAF = babl_format_new (
+ babl_model ("RaGaBaA"),
+ babl_type ("float"),
+ babl_component ("Ra"),
+ babl_component ("Ga"),
+ babl_component ("Ba"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgbA16 = babl_format_new (
+ babl_model ("RaGaBaA"),
+ babl_type ("u16"),
+ babl_component ("Ra"),
+ babl_component ("Ga"),
+ babl_component ("Ba"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgbA8 = babl_format_new (
+ babl_model ("RaGaBaA"),
+ babl_type ("u8"),
+ babl_component ("Ra"),
+ babl_component ("Ga"),
+ babl_component ("Ba"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgbF = babl_format_new (
+ babl_model ("RGB"),
+ babl_type ("float"),
+ babl_component ("R"),
+ babl_component ("G"),
+ babl_component ("B"),
+ NULL);
+ const Babl *rgb16 = babl_format_new (
+ babl_model ("RGB"),
+ babl_type ("u16"),
+ babl_component ("R"),
+ babl_component ("G"),
+ babl_component ("B"),
+ NULL);
+ const Babl *rgb8 = babl_format_new (
+ babl_model ("RGB"),
+ babl_type ("u8"),
+ babl_component ("R"),
+ babl_component ("G"),
+ babl_component ("B"),
+ NULL);
+ const Babl *gaF = babl_format_new (
+ babl_model ("YA"),
+ babl_type ("float"),
+ babl_component ("Y"),
+ babl_component ("A"),
+ NULL);
+ const Babl *gAF = babl_format_new (
+ babl_model ("YaA"),
+ babl_type ("float"),
+ babl_component ("Ya"),
+ babl_component ("A"),
+ NULL);
+ const Babl *gF = babl_format_new (
+ babl_model ("Y"),
+ babl_type ("float"),
+ babl_component ("Y"),
+ NULL);
+ const Babl *ga16 = babl_format_new (
+ babl_model ("YA"),
+ babl_type ("u16"),
+ babl_component ("Y"),
+ babl_component ("A"),
+ NULL);
+ const Babl *gA16 = babl_format_new (
+ babl_model ("YaA"),
+ babl_type ("u16"),
+ babl_component ("Ya"),
+ babl_component ("A"),
+ NULL);
+ const Babl *g16 = babl_format_new (
+ babl_model ("Y"),
+ babl_type ("u16"),
+ babl_component ("Y"),
+ NULL);
+ const Babl *ga8 = babl_format_new (
+ babl_model ("YA"),
+ babl_type ("u8"),
+ babl_component ("Y"),
+ babl_component ("A"),
+ NULL);
+ const Babl *gA8 = babl_format_new (
+ babl_model ("YaA"),
+ babl_type ("u8"),
+ babl_component ("Ya"),
+ babl_component ("A"),
+ NULL);
+ const Babl *g8 = babl_format_new (
+ babl_model ("Y"),
+ babl_type ("u8"),
+ babl_component ("Y"),
+ NULL);
+
+#define o(src, dst) \
+ babl_conversion_new (src, dst, "linear", conv_ ## src ## _ ## dst, NULL)
+
+ o (rgbaF, rgba8);
+ o (rgba8, rgbaF);
+ o (rgbaF, rgba16);
+ o (rgba16, rgbaF);
+ o (rgbAF, rgbA8);
+ o (rgbA8, rgbAF);
+ o (rgbAF, rgbA16);
+ o (rgbA16, rgbAF);
+ o (rgbF, rgb8);
+ o (rgb8, rgbF);
+ o (rgbF, rgb16);
+ o (rgb16, rgbF);
+ o (gaF, ga8);
+ o (gAF, gA8);
+ o (gF, g8);
+ o (ga8, gaF);
+ o (gA8, gAF);
+ o (g8, gF);
+ o (gaF, ga16);
+ o (gAF, gA16);
+ o (gF, g16);
+ o (ga16, gaF);
+ o (gA16, gAF);
+ o (g16, gF);
+
+ if (!table_inited)
+ table_init ();
+
+ return 0;
+}
--- /dev/null
+/*
+ * This file was part of gggl, it implements a variety of pixel conversion
+ * functions that are usable with babl, the file needs more cleanup.
+ *
+ * GGGL is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * GGGL is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with GGGL; if not, see <http://www.gnu.org/licenses/>.
+ *
+ * Rights are granted to use this shared object in libraries covered by
+ * LGPL. (exception added, during import into babl CVS.)
+ *
+ * Copyright 2003, 2004, 2005 Øyvind Kolås <pippin@gimp.org>
+ */
+
+#define _POSIX_C_SOURCE 200112L
+
+#include "config.h"
+#include <math.h>
+#include <string.h>
+#include <stdint.h>
+
+#include "babl.h"
+#include "extensions/util.h"
+
+/*
+ * Implemented according to information read from:
+ *
+ * http://www.cinenet.net/~spitzak/conversion/sketches_0265.pdf
+ *
+ * initially ignoring any diffusion, to keep the implementation
+ * smaller, and interchangeable with the non optimized version.
+ *
+ * due to ability to be able to relicence gggl under a different
+ * licence than GPL, I avoided the temptation to look at the
+ * source files in the same location, in case I was going to
+ * need this piece of code for projects where GPL compatibility
+ * was a must.
+ *
+ * TODO: error diffusion,
+ * gamma correction (not really,. gamma correction belongs in seperate ops,.
+ */
+
+/* lookup tables used in conversion */
+
+static float table_8_F[1 << 8];
+static float table_16_F[1 << 16];
+static unsigned char table_F_8[1 << 16];
+static unsigned short table_F_16[1 << 16];
+
+static uint32_t *table_8_F_int = NULL;
+
+static int table_inited = 0;
+
+static void
+table_init (void)
+{
+ if (table_inited)
+ return;
+
+ table_8_F_int = (void*)(table_8_F);
+
+ table_inited = 1;
+
+ /* fill tables for conversion from integer to float */
+ {
+ int i;
+ for (i = 0; i < 1 << 8; i++)
+ {
+ table_8_F[i] = (i * 1.0) / 255.0;
+ }
+ for (i = 0; i < 1 << 16; i++)
+ table_16_F[i] = (i * 1.0) / 65535.0;
+ }
+ /* fill tables for conversion from float to integer */
+ {
+ union
+ {
+ float f;
+ unsigned short s[2];
+ } u;
+ u.f = 0.0;
+
+ u.s[0] = 0x8000;
+
+ for (u.s[1] = 0; u.s[1] < 65535; u.s[1] += 1)
+ {
+ unsigned char c;
+ unsigned short s;
+
+ if (u.f <= 0.0)
+ {
+ c = 0;
+ s = 0;
+ }
+ else if (u.f >= 1.0)
+ {
+ c = 255;
+ s = 65535;
+ }
+ else
+ {
+ c = lrint (u.f * 255.0);
+ s = lrint (u.f * 65535.0);
+ }
+
+ /*fprintf (stderr, "%2.3f=%03i %05i ", f, c, (*hi));
+ / if (! ((*hi)%9))
+ / fprintf (stderr, "\n"); */
+
+ table_F_8[u.s[1]] = c;
+ table_F_16[u.s[1]] = s;
+ }
+ }
+ /* fix tables to ensure 1:1 conversions back and forth */
+ if (0)
+ { /*FIXME: probably not the right way to do it,.. must sit down and scribble on paper */
+ int i;
+ for (i = 0; i < 256; i++)
+ {
+ float f = table_8_F[i];
+ unsigned short *hi = ((unsigned short *) (void *) &f);
+ unsigned short *lo = ((unsigned short *) (void *) &f);
+ *lo = 0;
+ table_F_8[(*hi)] = i;
+ }
+ }
+}
+
+/* function to find the index in table for a float */
+static unsigned int
+gggl_float_to_index16 (float f)
+{
+ union
+ {
+ float f;
+ unsigned short s[2];
+ } u;
+ u.f = f;
+ return u.s[1];
+}
+
+static long
+conv_F_8 (unsigned char *src, unsigned char *dst, long samples)
+{
+ long n = samples;
+
+ if (!table_inited)
+ table_init ();
+ while (n--)
+ {
+ register float f = (*(float *) src);
+ *(unsigned char *) dst = table_F_8[gggl_float_to_index16 (f)];
+ dst += 1;
+ src += 4;
+ }
+ return samples;
+}
+
+static long
+conv_F_16 (unsigned char *src, unsigned char *dst, long samples)
+{
+ long n = samples;
+
+ if (!table_inited)
+ table_init ();
+ while (n--)
+ {
+ register float f = (*(float *) src);
+ *(unsigned short *) dst = table_F_16[gggl_float_to_index16 (f)];
+ dst += 2;
+ src += 4;
+ }
+ return samples;
+}
+
+static long
+conv_8_F (unsigned char *src, unsigned char *dst, long samples)
+{
+ long n = samples;
+
+ if (!table_inited)
+ table_init ();
+ while (n--)
+ {
+ (*(uint32_t *) dst) = table_8_F_int[*(unsigned char *) src];
+ dst += 4;
+ src += 1;
+ }
+ return samples;
+}
+
+static long
+conv_rgb8_rgbaF (unsigned char *src, unsigned char *dst, long samples)
+{
+ long n = samples;
+
+ if (!table_inited)
+ table_init ();
+ while (n--)
+ {
+ (*(uint32_t *) dst) = table_8_F_int[*(unsigned char *) src];
+ dst += 4;
+ src += 1;
+ (*(uint32_t *) dst) = table_8_F_int[*(unsigned char *) src];
+ dst += 4;
+ src += 1;
+ (*(uint32_t *) dst) = table_8_F_int[*(unsigned char *) src];
+ dst += 4;
+ src += 1;
+ (*(float *) dst) = 1.0;
+ dst += 4;
+ }
+ return samples;
+}
+
+static long
+conv_16_F (unsigned char *src, unsigned char *dst, long samples)
+{
+ long n = samples;
+
+ if (!table_inited)
+ table_init ();
+ while (n--)
+ {
+ (*(float *) dst) = table_16_F[*(unsigned short *) src];
+ dst += 4;
+ src += 2;
+ }
+ return samples;
+}
+
+static long
+conv_rgbaF_rgb8 (unsigned char *src, unsigned char *dst, long samples)
+{
+ long n = samples;
+
+ while (n--)
+ {
+ register float f = (*(float *) src);
+ *(unsigned char *) dst = table_F_8[gggl_float_to_index16 (f)];
+ src += 4;
+ dst += 1;
+
+ f = (*(float *) src);
+ *(unsigned char *) dst = table_F_8[gggl_float_to_index16 (f)];
+ src += 4;
+ dst += 1;
+
+ f = (*(float *) src);
+ *(unsigned char *) dst = table_F_8[gggl_float_to_index16 (f)];
+ src += 4;
+ dst += 1;
+
+ src += 4;
+ }
+ return samples;
+}
+
+
+/*********/
+static long
+conv_rgbaF_rgba8 (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_F_8 (src, dst, samples * 4);
+ return samples;
+}
+
+static long
+conv_rgbF_rgb8 (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_F_8 (src, dst, samples * 3);
+ return samples;
+}
+
+static long
+conv_gaF_ga8 (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_F_8 (src, dst, samples * 2);
+ return samples;
+}
+
+#define conv_rgbAF_rgbA8 conv_rgbaF_rgba8
+#define conv_gF_g8 conv_F_8
+#define conv_gAF_gA8 conv_gaF_ga8
+
+static long
+conv_rgbaF_rgba16 (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_F_16 (src, dst, samples * 4);
+ return samples;
+}
+
+static long
+conv_rgbF_rgb16 (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_F_16 (src, dst, samples * 3);
+ return samples;
+}
+
+static long
+conv_gaF_ga16 (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_F_16 (src, dst, samples * 2);
+ return samples;
+}
+
+#define conv_rgbAF_rgbA16 conv_rgbaF_rgba16
+#define conv_gF_g16 conv_F_16
+#define conv_gAF_gA16 conv_gaF_ga16
+
+static long
+conv_rgba8_rgbaF (unsigned char *src, unsigned char *dst, long samples)
+{
+ return conv_8_F (src, dst, samples * 4) / 4;
+}
+
+
+static long
+conv_rgb8_rgbF (unsigned char *src, unsigned char *dst, long samples)
+{
+ return conv_8_F (src, dst, samples * 3) / 3;
+}
+
+static long
+conv_ga8_gaF (unsigned char *src, unsigned char *dst, long samples)
+{
+ return conv_8_F (src, dst, samples * 2) / 2;
+}
+
+#define conv_rgbA8_rgbAF conv_rgba8_rgbaF
+#define conv_gA8_gAF conv_ga8_gaF
+#define conv_g8_gF conv_8_F
+
+static long
+conv_rgba16_rgbaF (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_16_F (src, dst, samples * 4);
+ return samples;
+}
+
+static long
+conv_rgb16_rgbF (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_16_F (src, dst, samples * 3);
+ return samples;
+}
+
+static long
+conv_ga16_gaF (unsigned char *src, unsigned char *dst, long samples)
+{
+ conv_16_F (src, dst, samples * 2);
+ return samples;
+}
+
+#define conv_rgbA16_rgbAF conv_rgba16_rgbaF
+#define conv_gA16_gAF conv_ga16_gaF
+#define conv_g16_gF conv_16_F
+
+int init (void);
+
+int
+init (void)
+{
+ const Babl *rgbaF = babl_format_new (
+ babl_model ("R'G'B'A"),
+ babl_type ("float"),
+ babl_component ("R'"),
+ babl_component ("G'"),
+ babl_component ("B'"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgba16 = babl_format_new (
+ babl_model ("R'G'B'A"),
+ babl_type ("u16"),
+ babl_component ("R'"),
+ babl_component ("G'"),
+ babl_component ("B'"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgba8 = babl_format_new (
+ babl_model ("R'G'B'A"),
+ babl_type ("u8"),
+ babl_component ("R'"),
+ babl_component ("G'"),
+ babl_component ("B'"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgbAF = babl_format_new (
+ babl_model ("R'aG'aB'aA"),
+ babl_type ("float"),
+ babl_component ("R'a"),
+ babl_component ("G'a"),
+ babl_component ("B'a"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgbA16 = babl_format_new (
+ babl_model ("R'aG'aB'aA"),
+ babl_type ("u16"),
+ babl_component ("R'a"),
+ babl_component ("G'a"),
+ babl_component ("B'a"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgbA8 = babl_format_new (
+ babl_model ("R'aG'aB'aA"),
+ babl_type ("u8"),
+ babl_component ("R'a"),
+ babl_component ("G'a"),
+ babl_component ("B'a"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgbF = babl_format_new (
+ babl_model ("R'G'B'"),
+ babl_type ("float"),
+ babl_component ("R'"),
+ babl_component ("G'"),
+ babl_component ("B'"),
+ NULL);
+ const Babl *rgb16 = babl_format_new (
+ babl_model ("R'G'B'"),
+ babl_type ("u16"),
+ babl_component ("R'"),
+ babl_component ("G'"),
+ babl_component ("B'"),
+ NULL);
+ const Babl *rgb8 = babl_format_new (
+ babl_model ("R'G'B'"),
+ babl_type ("u8"),
+ babl_component ("R'"),
+ babl_component ("G'"),
+ babl_component ("B'"),
+ NULL);
+ const Babl *gaF = babl_format_new (
+ babl_model ("Y'A"),
+ babl_type ("float"),
+ babl_component ("Y'"),
+ babl_component ("A"),
+ NULL);
+ const Babl *gAF = babl_format_new (
+ babl_model ("Y'aA"),
+ babl_type ("float"),
+ babl_component ("Y'a"),
+ babl_component ("A"),
+ NULL);
+ const Babl *gF = babl_format_new (
+ babl_model ("Y'"),
+ babl_type ("float"),
+ babl_component ("Y'"),
+ NULL);
+ const Babl *ga16 = babl_format_new (
+ babl_model ("Y'A"),
+ babl_type ("u16"),
+ babl_component ("Y'"),
+ babl_component ("A"),
+ NULL);
+ const Babl *gA16 = babl_format_new (
+ babl_model ("Y'aA"),
+ babl_type ("u16"),
+ babl_component ("Y'a"),
+ babl_component ("A"),
+ NULL);
+ const Babl *g16 = babl_format_new (
+ babl_model ("Y'"),
+ babl_type ("u16"),
+ babl_component ("Y'"),
+ NULL);
+ const Babl *ga8 = babl_format_new (
+ babl_model ("Y'A"),
+ babl_type ("u8"),
+ babl_component ("Y'"),
+ babl_component ("A"),
+ NULL);
+ const Babl *gA8 = babl_format_new (
+ babl_model ("Y'aA"),
+ babl_type ("u8"),
+ babl_component ("Y'a"),
+ babl_component ("A"),
+ NULL);
+ const Babl *g8 = babl_format_new (
+ babl_model ("Y'"),
+ babl_type ("u8"),
+ babl_component ("Y'"),
+ NULL);
+
+#define o(src, dst) \
+ babl_conversion_new (src, dst, "linear", conv_ ## src ## _ ## dst, NULL)
+
+ o (rgbaF, rgba8);
+ o (rgba8, rgbaF);
+ o (rgbaF, rgba16);
+ o (rgba16, rgbaF);
+ o (rgbAF, rgbA8);
+ o (rgbA8, rgbAF);
+ o (rgbAF, rgbA16);
+ o (rgbA16, rgbAF);
+ o (rgbF, rgb8);
+ o (rgb8, rgbF);
+ o (rgbF, rgb16);
+ o (rgb16, rgbF);
+ o (gaF, ga8);
+ o (gAF, gA8);
+ o (gF, g8);
+ o (ga8, gaF);
+ o (gA8, gAF);
+ o (g8, gF);
+ o (gaF, ga16);
+ o (gAF, gA16);
+ o (gF, g16);
+ o (ga16, gaF);
+ o (gA16, gAF);
+ o (g16, gF);
+ o (rgbaF, rgb8);
+ o (rgb8, rgbaF);
+
+ if (!table_inited)
+ table_init ();
+
+ return 0;
+}
* gamma correction (not really,. gamma correction belongs in seperate ops,.
*/
-#define USE_TABLES
-#ifdef USE_TABLES
-
-/* lookup tables used in conversion */
-
-static float table_8_F[1 << 8];
-static float table_16_F[1 << 16];
-static unsigned char table_F_8[1 << 16];
-static unsigned short table_F_16[1 << 16];
-
-static uint32_t *table_8_F_int = NULL;
-
-static int table_inited = 0;
-
-static void
-table_init (void)
-{
- if (table_inited)
- return;
-
- table_8_F_int = (void*)(table_8_F);
-
- table_inited = 1;
-
- /* fill tables for conversion from integer to float */
- {
- int i;
- for (i = 0; i < 1 << 8; i++)
- {
- table_8_F[i] = (i * 1.0) / 255.0;
- }
- for (i = 0; i < 1 << 16; i++)
- table_16_F[i] = (i * 1.0) / 65535.0;
- }
- /* fill tables for conversion from float to integer */
- {
- union
- {
- float f;
- unsigned short s[2];
- } u;
- u.f = 0.0;
-
- u.s[0] = 0x8000;
-
- for (u.s[1] = 0; u.s[1] < 65535; u.s[1] += 1)
- {
- unsigned char c;
- unsigned short s;
-
- if (u.f <= 0.0)
- {
- c = 0;
- s = 0;
- }
- else if (u.f >= 1.0)
- {
- c = 255;
- s = 65535;
- }
- else
- {
- c = lrint (u.f * 255.0);
- s = lrint (u.f * 65535.0);
- }
-
- /*fprintf (stderr, "%2.3f=%03i %05i ", f, c, (*hi));
- / if (! ((*hi)%9))
- / fprintf (stderr, "\n"); */
-
- table_F_8[u.s[1]] = c;
- table_F_16[u.s[1]] = s;
- }
- }
- /* fix tables to ensure 1:1 conversions back and forth */
- if (0)
- { /*FIXME: probably not the right way to do it,.. must sit down and scribble on paper */
- int i;
- for (i = 0; i < 256; i++)
- {
- float f = table_8_F[i];
- unsigned short *hi = ((unsigned short *) (void *) &f);
- unsigned short *lo = ((unsigned short *) (void *) &f);
- *lo = 0;
- table_F_8[(*hi)] = i;
- }
- }
-}
-
-/* function to find the index in table for a float */
-static unsigned int
-gggl_float_to_index16 (float f)
-{
- union
- {
- float f;
- unsigned short s[2];
- } u;
- u.f = f;
- return u.s[1];
-}
-
-static long
-conv_F_8 (unsigned char *src, unsigned char *dst, long samples)
-{
- long n = samples;
-
- if (!table_inited)
- table_init ();
- while (n--)
- {
- register float f = (*(float *) src);
- *(unsigned char *) dst = table_F_8[gggl_float_to_index16 (f)];
- dst += 1;
- src += 4;
- }
- return samples;
-}
-
-static long
-conv_F_16 (unsigned char *src, unsigned char *dst, long samples)
-{
- long n = samples;
-
- if (!table_inited)
- table_init ();
- while (n--)
- {
- register float f = (*(float *) src);
- *(unsigned short *) dst = table_F_16[gggl_float_to_index16 (f)];
- dst += 2;
- src += 4;
- }
- return samples;
-}
-
-static long
-conv_8_F (unsigned char *src, unsigned char *dst, long samples)
-{
- long n = samples;
-
- if (!table_inited)
- table_init ();
- while (n--)
- {
- (*(uint32_t *) dst) = table_8_F_int[*(unsigned char *) src];
- dst += 4;
- src += 1;
- }
- return samples;
-}
-
-static long
-conv_rgb8_rgbaF (unsigned char *src, unsigned char *dst, long samples)
-{
- long n = samples;
-
- if (!table_inited)
- table_init ();
- while (n--)
- {
- (*(uint32_t *) dst) = table_8_F_int[*(unsigned char *) src];
- dst += 4;
- src += 1;
- (*(uint32_t *) dst) = table_8_F_int[*(unsigned char *) src];
- dst += 4;
- src += 1;
- (*(uint32_t *) dst) = table_8_F_int[*(unsigned char *) src];
- dst += 4;
- src += 1;
- (*(float *) dst) = 1.0;
- dst += 4;
- }
- return samples;
-}
-
-static long
-conv_16_F (unsigned char *src, unsigned char *dst, long samples)
-{
- long n = samples;
-
- if (!table_inited)
- table_init ();
- while (n--)
- {
- (*(float *) dst) = table_16_F[*(unsigned short *) src];
- dst += 4;
- src += 2;
- }
- return samples;
-}
-
-static long
-conv_rgbaF_rgb8 (unsigned char *src, unsigned char *dst, long samples)
-{
- long n = samples;
-
- while (n--)
- {
- register float f = (*(float *) src);
- *(unsigned char *) dst = table_F_8[gggl_float_to_index16 (f)];
- src += 4;
- dst += 1;
-
- f = (*(float *) src);
- *(unsigned char *) dst = table_F_8[gggl_float_to_index16 (f)];
- src += 4;
- dst += 1;
-
- f = (*(float *) src);
- *(unsigned char *) dst = table_F_8[gggl_float_to_index16 (f)];
- src += 4;
- dst += 1;
-
- src += 4;
- }
- return samples;
-}
-
-#else
-
static long
conv_F_8 (unsigned char *src, unsigned char *dst, long samples)
{
return samples;
}
-#endif
-
-
static long
conv_F_D (unsigned char *src, unsigned char *dst, long samples)
{
*(uint32_t *) dst = (*(uint32_t *) src);
src += 4;
dst += 4;
- *(uint32_t *) dst = 1.0;
+ *(float *) dst = 1.0;
dst += 4;
}
return samples;
o (rgbA16, rgbaF);
o (rgbaF, rgbaD);
o (rgbaD, rgbaF);
- o (rgb8, rgbaF);
-
-#ifdef USE_TABLES
- if (!table_inited)
- table_init ();
-#endif
return 0;
}
--- /dev/null
+/* babl - dynamically extendable universal pixel conversion library.
+ * Copyright (C) 2015 Daniel Sabo
+ *
+ * This library 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 3 of the License, or (at your option) any later version.
+ *
+ * This library 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 this library; if not, see
+ * <http://www.gnu.org/licenses/>.
+ */
+
+#include "config.h"
+
+#if defined(USE_SSE4_1) && defined(USE_F16C) && defined(ARCH_X86_64)
+
+#include <immintrin.h>
+
+#include <stdint.h>
+#include <stdlib.h>
+
+#include "babl.h"
+#include "babl-cpuaccel.h"
+#include "extensions/util.h"
+
+static inline long
+conv_yHalf_yF (const uint16_t *src, float *dst, long samples)
+{
+ const uint64_t *s_vec;
+ __v4sf *d_vec;
+
+ long n = samples;
+
+ s_vec = (const uint64_t *)src;
+ d_vec = (__v4sf *)dst;
+
+ while (n >= 4)
+ {
+ __m128i in_val = _mm_insert_epi64((__m128i)_mm_setzero_ps(), *s_vec++, 0);
+ __v4sf out_val = (__v4sf)_mm_cvtph_ps(in_val);
+ _mm_storeu_ps((float *)d_vec++, out_val);
+ n -= 4;
+ }
+
+ src = (const uint16_t *)s_vec;
+ dst = (float *)d_vec;
+
+ while (n)
+ {
+ __m128i in_val = _mm_insert_epi16((__m128i)_mm_setzero_ps(), *src++, 0);
+ __v4sf out_val = (__v4sf)_mm_cvtph_ps(in_val);
+ _mm_store_ss(dst++, out_val);
+ n -= 1;
+ }
+
+ return samples;
+}
+
+static long
+conv_yaHalf_yaF (const uint16_t *src, float *dst, long samples)
+{
+ return conv_yHalf_yF (src, dst, samples * 2) / 2;
+}
+
+static long
+conv_rgbHalf_rgbF (const uint16_t *src, float *dst, long samples)
+{
+ return conv_yHalf_yF (src, dst, samples * 3) / 3;
+}
+
+static long
+conv_rgbaHalf_rgbaF (const uint16_t *src, float *dst, long samples)
+{
+ return conv_yHalf_yF (src, dst, samples * 4) / 4;
+}
+
+static inline long
+conv_yF_yHalf (const float *src, uint16_t *dst, long samples)
+{
+ const __v4sf *s_vec;
+ uint64_t *d_vec;
+
+ long n = samples;
+
+ s_vec = (const __v4sf *)src;
+ d_vec = (uint64_t *)dst;
+
+ while (n >= 4)
+ {
+ __m128 in_val = _mm_loadu_ps((float *)s_vec++);
+ __m128i out_val = _mm_cvtps_ph(in_val, _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC);
+ _mm_storel_epi64((__m128i *)d_vec++, out_val);
+ n -= 4;
+ }
+
+ src = (const float *)s_vec;
+ dst = (uint16_t *)d_vec;
+
+ while (n)
+ {
+ __m128 in_val = _mm_load_ss(src++);
+ __m128i out_val = _mm_cvtps_ph(in_val, _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC);
+ *dst++ = _mm_extract_epi16(out_val, 0);
+ n -= 1;
+ }
+
+ return samples;
+}
+
+static long
+conv_yaF_yaHalf (const float *src, uint16_t *dst, long samples)
+{
+ return conv_yF_yHalf (src, dst, samples * 2) / 2;
+}
+
+static long
+conv_rgbF_rgbHalf (const float *src, uint16_t *dst, long samples)
+{
+ return conv_yF_yHalf (src, dst, samples * 3) / 3;
+}
+
+static long
+conv_rgbaF_rgbaHalf (const float *src, uint16_t *dst, long samples)
+{
+ return conv_yF_yHalf (src, dst, samples * 4) / 4;
+}
+
+#endif /* defined(USE_SSE4_1) && defined(USE_F16C) && defined(ARCH_X86_64) */
+
+int init (void);
+
+int
+init (void)
+{
+#if defined(USE_SSE4_1) && defined(USE_F16C) && defined(ARCH_X86_64)
+ const Babl *rgbaF_linear = babl_format_new (
+ babl_model ("RGBA"),
+ babl_type ("float"),
+ babl_component ("R"),
+ babl_component ("G"),
+ babl_component ("B"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgbaHalf_linear = babl_format_new (
+ babl_model ("RGBA"),
+ babl_type ("half"),
+ babl_component ("R"),
+ babl_component ("G"),
+ babl_component ("B"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgbaF_gamma = babl_format_new (
+ babl_model ("R'G'B'A"),
+ babl_type ("float"),
+ babl_component ("R'"),
+ babl_component ("G'"),
+ babl_component ("B'"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgbaHalf_gamma = babl_format_new (
+ babl_model ("R'G'B'A"),
+ babl_type ("half"),
+ babl_component ("R'"),
+ babl_component ("G'"),
+ babl_component ("B'"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgbF_linear = babl_format_new (
+ babl_model ("RGB"),
+ babl_type ("float"),
+ babl_component ("R"),
+ babl_component ("G"),
+ babl_component ("B"),
+ NULL);
+ const Babl *rgbHalf_linear = babl_format_new (
+ babl_model ("RGB"),
+ babl_type ("half"),
+ babl_component ("R"),
+ babl_component ("G"),
+ babl_component ("B"),
+ NULL);
+ const Babl *rgbF_gamma = babl_format_new (
+ babl_model ("R'G'B'"),
+ babl_type ("float"),
+ babl_component ("R'"),
+ babl_component ("G'"),
+ babl_component ("B'"),
+ NULL);
+ const Babl *rgbHalf_gamma = babl_format_new (
+ babl_model ("R'G'B'"),
+ babl_type ("half"),
+ babl_component ("R'"),
+ babl_component ("G'"),
+ babl_component ("B'"),
+ NULL);
+ const Babl *yaF_linear = babl_format_new (
+ babl_model ("YA"),
+ babl_type ("float"),
+ babl_component ("Y"),
+ babl_component ("A"),
+ NULL);
+ const Babl *yaHalf_linear = babl_format_new (
+ babl_model ("YA"),
+ babl_type ("half"),
+ babl_component ("Y"),
+ babl_component ("A"),
+ NULL);
+ const Babl *yaF_gamma = babl_format_new (
+ babl_model ("Y'A"),
+ babl_type ("float"),
+ babl_component ("Y'"),
+ babl_component ("A"),
+ NULL);
+ const Babl *yaHalf_gamma = babl_format_new (
+ babl_model ("Y'A"),
+ babl_type ("half"),
+ babl_component ("Y'"),
+ babl_component ("A"),
+ NULL);
+ const Babl *yF_linear = babl_format_new (
+ babl_model ("Y"),
+ babl_type ("float"),
+ babl_component ("Y"),
+ NULL);
+ const Babl *yHalf_linear = babl_format_new (
+ babl_model ("Y"),
+ babl_type ("half"),
+ babl_component ("Y"),
+ NULL);
+ const Babl *yF_gamma = babl_format_new (
+ babl_model ("Y'"),
+ babl_type ("float"),
+ babl_component ("Y'"),
+ NULL);
+ const Babl *yHalf_gamma = babl_format_new (
+ babl_model ("Y'"),
+ babl_type ("half"),
+ babl_component ("Y'"),
+ NULL);
+
+#define CONV(src, dst) \
+{ \
+ babl_conversion_new (src ## _linear, dst ## _linear, "linear", conv_ ## src ## _ ## dst, NULL); \
+ babl_conversion_new (src ## _gamma, dst ## _gamma, "linear", conv_ ## src ## _ ## dst, NULL); \
+}
+
+ if ((babl_cpu_accel_get_support () & BABL_CPU_ACCEL_X86_SSE4_1) &&
+ (babl_cpu_accel_get_support () & BABL_CPU_ACCEL_X86_F16C))
+ {
+ CONV(rgbaHalf, rgbaF);
+ CONV(rgbHalf, rgbF);
+ CONV(yaHalf, yaF);
+ CONV(yHalf, yF);
+ CONV(rgbaF, rgbaHalf);
+ CONV(rgbF, rgbHalf);
+ CONV(yaF, yaHalf);
+ CONV(yF, yHalf);
+ }
+
+#endif /* defined(USE_SSE4_1) && defined(USE_F16C) && defined(ARCH_X86_64) */
+
+ return 0;
+}
+
babl_component ("Ba"),
babl_component ("A"),
NULL);
+ const Babl *rgbAF_gamma = babl_format_new (
+ babl_model ("R'aG'aB'aA"),
+ babl_type ("float"),
+ babl_component ("R'a"),
+ babl_component ("G'a"),
+ babl_component ("B'a"),
+ babl_component ("A"),
+ NULL);
const Babl *rgbaF_gamma = babl_format_new (
babl_model ("R'G'B'A"),
babl_type ("float"),
"linear",
conv_rgbaF_linear_rgbAF_linear,
NULL);
+
+ babl_conversion_new(rgbaF_gamma,
+ rgbAF_gamma,
+ "linear",
+ conv_rgbaF_linear_rgbAF_linear,
+ NULL);
/* Which of these is faster varies by CPU, and the difference
* is big enough that it's worthwhile to include both and
"linear",
conv_rgbAF_linear_rgbaF_linear_shuffle,
NULL);
+ babl_conversion_new(rgbAF_gamma,
+ rgbaF_gamma,
+ "linear",
+ conv_rgbAF_linear_rgbaF_linear_shuffle,
+ NULL);
+
babl_conversion_new(rgbAF_linear,
rgbaF_linear,
"linear",
--- /dev/null
+/* babl - dynamically extendable universal pixel conversion library.
+ * Copyright (C) 2013 Daniel Sabo
+ *
+ * This library 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 3 of the License, or (at your option) any later version.
+ *
+ * This library 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 this library; if not, see
+ * <http://www.gnu.org/licenses/>.
+ */
+
+#include "config.h"
+
+#if defined(USE_SSE4_1)
+
+/* SSE 4 */
+#include <smmintrin.h>
+
+#include <stdint.h>
+#include <stdlib.h>
+
+#include "babl.h"
+#include "babl-cpuaccel.h"
+#include "extensions/util.h"
+
+static inline long
+conv_y8_yF (const uint8_t *src, float *dst, long samples)
+{
+ const float factor = 1.0f / 255.0f;
+ const __v4sf factor_vec = {1.0f / 255.0f, 1.0f / 255.0f, 1.0f / 255.0f, 1.0f / 255.0f};
+ const uint32_t *s_vec;
+ __v4sf *d_vec;
+
+ long n = samples;
+
+ s_vec = (const uint32_t *)src;
+ d_vec = (__v4sf *)dst;
+
+ while (n >= 4)
+ {
+ __m128i in_val;
+ __v4sf out_val;
+ in_val = _mm_insert_epi32 (in_val, *s_vec++, 0);
+ in_val = _mm_cvtepu8_epi32 (in_val);
+ out_val = _mm_cvtepi32_ps (in_val) * factor_vec;
+ _mm_storeu_ps ((float *)d_vec++, out_val);
+ n -= 4;
+ }
+
+ src = (const uint8_t *)s_vec;
+ dst = (float *)d_vec;
+
+ while (n)
+ {
+ *dst++ = (float)(*src++) * factor;
+ n -= 1;
+ }
+
+ return samples;
+}
+
+static long
+conv_ya8_yaF (const uint8_t *src, float *dst, long samples)
+{
+ return conv_y8_yF (src, dst, samples * 2) / 2;
+}
+
+static long
+conv_rgb8_rgbF (const uint8_t *src, float *dst, long samples)
+{
+ return conv_y8_yF (src, dst, samples * 3) / 3;
+}
+
+static long
+conv_rgba8_rgbaF (const uint8_t *src, float *dst, long samples)
+{
+ return conv_y8_yF (src, dst, samples * 4) / 4;
+}
+
+#endif
+
+int init (void);
+
+int
+init (void)
+{
+#if defined(USE_SSE4_1)
+ const Babl *rgbaF_linear = babl_format_new (
+ babl_model ("RGBA"),
+ babl_type ("float"),
+ babl_component ("R"),
+ babl_component ("G"),
+ babl_component ("B"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgba8_linear = babl_format_new (
+ babl_model ("RGBA"),
+ babl_type ("u8"),
+ babl_component ("R"),
+ babl_component ("G"),
+ babl_component ("B"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgbaF_gamma = babl_format_new (
+ babl_model ("R'G'B'A"),
+ babl_type ("float"),
+ babl_component ("R'"),
+ babl_component ("G'"),
+ babl_component ("B'"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgba8_gamma = babl_format_new (
+ babl_model ("R'G'B'A"),
+ babl_type ("u8"),
+ babl_component ("R'"),
+ babl_component ("G'"),
+ babl_component ("B'"),
+ babl_component ("A"),
+ NULL);
+ const Babl *rgbF_linear = babl_format_new (
+ babl_model ("RGB"),
+ babl_type ("float"),
+ babl_component ("R"),
+ babl_component ("G"),
+ babl_component ("B"),
+ NULL);
+ const Babl *rgb8_linear = babl_format_new (
+ babl_model ("RGB"),
+ babl_type ("u8"),
+ babl_component ("R"),
+ babl_component ("G"),
+ babl_component ("B"),
+ NULL);
+ const Babl *rgbF_gamma = babl_format_new (
+ babl_model ("R'G'B'"),
+ babl_type ("float"),
+ babl_component ("R'"),
+ babl_component ("G'"),
+ babl_component ("B'"),
+ NULL);
+ const Babl *rgb8_gamma = babl_format_new (
+ babl_model ("R'G'B'"),
+ babl_type ("u8"),
+ babl_component ("R'"),
+ babl_component ("G'"),
+ babl_component ("B'"),
+ NULL);
+ const Babl *yaF_linear = babl_format_new (
+ babl_model ("YA"),
+ babl_type ("float"),
+ babl_component ("Y"),
+ babl_component ("A"),
+ NULL);
+ const Babl *ya8_linear = babl_format_new (
+ babl_model ("YA"),
+ babl_type ("u8"),
+ babl_component ("Y"),
+ babl_component ("A"),
+ NULL);
+ const Babl *yaF_gamma = babl_format_new (
+ babl_model ("Y'A"),
+ babl_type ("float"),
+ babl_component ("Y'"),
+ babl_component ("A"),
+ NULL);
+ const Babl *ya8_gamma = babl_format_new (
+ babl_model ("Y'A"),
+ babl_type ("u8"),
+ babl_component ("Y'"),
+ babl_component ("A"),
+ NULL);
+ const Babl *yF_linear = babl_format_new (
+ babl_model ("Y"),
+ babl_type ("float"),
+ babl_component ("Y"),
+ NULL);
+ const Babl *y8_linear = babl_format_new (
+ babl_model ("Y"),
+ babl_type ("u8"),
+ babl_component ("Y"),
+ NULL);
+ const Babl *yF_gamma = babl_format_new (
+ babl_model ("Y'"),
+ babl_type ("float"),
+ babl_component ("Y'"),
+ NULL);
+ const Babl *y8_gamma = babl_format_new (
+ babl_model ("Y'"),
+ babl_type ("u8"),
+ babl_component ("Y'"),
+ NULL);
+
+#define CONV(src, dst) \
+{ \
+ babl_conversion_new (src ## _linear, dst ## _linear, "linear", conv_ ## src ## _ ## dst, NULL); \
+ babl_conversion_new (src ## _gamma, dst ## _gamma, "linear", conv_ ## src ## _ ## dst, NULL); \
+}
+
+ if ((babl_cpu_accel_get_support () & BABL_CPU_ACCEL_X86_SSE4_1))
+ {
+ CONV(rgba8, rgbaF);
+ CONV(rgb8, rgbF);
+ CONV(ya8, yaF);
+ CONV(y8, yF);
+ }
+
+#endif
+
+ return 0;
+}
+
ECHO_T = @ECHO_T@
EGREP = @EGREP@
EXEEXT = @EXEEXT@
+F16C_EXTRA_CFLAGS = @F16C_EXTRA_CFLAGS@
FGREP = @FGREP@
GREP = @GREP@
INSTALL = @INSTALL@
SHELL = @SHELL@
SHREXT = @SHREXT@
SSE2_EXTRA_CFLAGS = @SSE2_EXTRA_CFLAGS@
+SSE4_1_EXTRA_CFLAGS = @SSE4_1_EXTRA_CFLAGS@
SSE_EXTRA_CFLAGS = @SSE_EXTRA_CFLAGS@
STRIP = @STRIP@
THREAD_LIB = @THREAD_LIB@
ECHO_T = @ECHO_T@
EGREP = @EGREP@
EXEEXT = @EXEEXT@
+F16C_EXTRA_CFLAGS = @F16C_EXTRA_CFLAGS@
FGREP = @FGREP@
GREP = @GREP@
INSTALL = @INSTALL@
SHELL = @SHELL@
SHREXT = @SHREXT@
SSE2_EXTRA_CFLAGS = @SSE2_EXTRA_CFLAGS@
+SSE4_1_EXTRA_CFLAGS = @SSE4_1_EXTRA_CFLAGS@
SSE_EXTRA_CFLAGS = @SSE_EXTRA_CFLAGS@
STRIP = @STRIP@
THREAD_LIB = @THREAD_LIB@
projects / babl.git / commitdiff
![docs/graphics/babl-16x16.png](/"http:/www.example.com/cgi/"https:/www.github.com/"bookmarks://"cascade.html?p=babl.git;a=blob;f=docs/graphics/babl-16x16.png;h=57ad178affd0a31a67e5f99115c15214199d729c;hb=35fc3506e93d79b0ed7a3c352f88f9b5c94120db){kind=link}
![history](/"http:/www.example.com/cgi/"https:/www.github.com/"bookmarks://"cascade.html?p=babl.git;a=history;f=docs/graphics/babl-16x16.png;hb=35fc3506e93d79b0ed7a3c352f88f9b5c94120db){kind=link}
![docs/graphics/babl-48x48.png](/"http:/www.example.com/cgi/"https:/www.github.com/"bookmarks://"cascade.html?p=babl.git;a=blob;f=docs/graphics/babl-48x48.png;h=717e95d4d517f0e086773a1c6171219d880379b5;hb=35fc3506e93d79b0ed7a3c352f88f9b5c94120db){kind=link}
![history](/"http:/www.example.com/cgi/"https:/www.github.com/"bookmarks://"cascade.html?p=babl.git;a=history;f=docs/graphics/babl-48x48.png;hb=35fc3506e93d79b0ed7a3c352f88f9b5c94120db){kind=link}
![docs/graphics/babl-a4poster.png](/"http:/www.example.com/cgi/"https:/www.github.com/"bookmarks://"cascade.html?p=babl.git;a=blob;f=docs/graphics/babl-a4poster.png;h=8b21a08aa7d02e535f4f8aebd2c2346c2e8b84a9;hb=35fc3506e93d79b0ed7a3c352f88f9b5c94120db){kind=link}
![history](/"http:/www.example.com/cgi/"https:/www.github.com/"bookmarks://"cascade.html?p=babl.git;a=history;f=docs/graphics/babl-a4poster.png;hb=35fc3506e93d79b0ed7a3c352f88f9b5c94120db){kind=link}
![a/docs/graphics/babl-16x16.png](/"http:/www.example.com/cgi/"https:/www.github.com/"bookmarks://"cascade.html?p=babl.git;a=blob;f=docs/graphics/babl-16x16.png;h=aea6726cedd077078db077698a41f224fabb7c54){kind=link}
![a/docs/graphics/babl-48x48.png](/"http:/www.example.com/cgi/"https:/www.github.com/"bookmarks://"cascade.html?p=babl.git;a=blob;f=docs/graphics/babl-48x48.png;h=7a8e1cc885e3c42b0f99694743a3f18edc8a0b8b){kind=link}
![a/docs/graphics/babl-a4poster.png](/"http:/www.example.com/cgi/"https:/www.github.com/"bookmarks://"cascade.html?p=babl.git;a=blob;f=docs/graphics/babl-a4poster.png;h=85c4d4d5a8949ac54b3aa4e630011ebc44b77497){kind=link}