gecko-dev/media/libjpeg/jdct.h
Ryan VanderMeulen 50bc583645 Bug 1061240 - Update libjpeg-turbo to version 1.4. r=jmuizelaar, r=glandium, f=rankov
--HG--
rename : media/libjpeg/config.h => media/libjpeg/jconfigint.h
rename : media/libjpeg/simd/jcclrmmx.asm => media/libjpeg/simd/jccolext-mmx.asm
rename : media/libjpeg/simd/jcclrss2-64.asm => media/libjpeg/simd/jccolext-sse2-64.asm
rename : media/libjpeg/simd/jcclrss2.asm => media/libjpeg/simd/jccolext-sse2.asm
rename : media/libjpeg/simd/jccolmmx.asm => media/libjpeg/simd/jccolor-mmx.asm
rename : media/libjpeg/simd/jccolss2-64.asm => media/libjpeg/simd/jccolor-sse2-64.asm
rename : media/libjpeg/simd/jccolss2.asm => media/libjpeg/simd/jccolor-sse2.asm
rename : media/libjpeg/simd/jcgrammx.asm => media/libjpeg/simd/jcgray-mmx.asm
rename : media/libjpeg/simd/jcgrass2-64.asm => media/libjpeg/simd/jcgray-sse2-64.asm
rename : media/libjpeg/simd/jcgrass2.asm => media/libjpeg/simd/jcgray-sse2.asm
rename : media/libjpeg/simd/jcgrymmx.asm => media/libjpeg/simd/jcgryext-mmx.asm
rename : media/libjpeg/simd/jcgryss2-64.asm => media/libjpeg/simd/jcgryext-sse2-64.asm
rename : media/libjpeg/simd/jcgryss2.asm => media/libjpeg/simd/jcgryext-sse2.asm
rename : media/libjpeg/simd/jcsammmx.asm => media/libjpeg/simd/jcsample-mmx.asm
rename : media/libjpeg/simd/jcsamss2-64.asm => media/libjpeg/simd/jcsample-sse2-64.asm
rename : media/libjpeg/simd/jcsamss2.asm => media/libjpeg/simd/jcsample-sse2.asm
rename : media/libjpeg/simd/jdclrmmx.asm => media/libjpeg/simd/jdcolext-mmx.asm
rename : media/libjpeg/simd/jdclrss2-64.asm => media/libjpeg/simd/jdcolext-sse2-64.asm
rename : media/libjpeg/simd/jdclrss2.asm => media/libjpeg/simd/jdcolext-sse2.asm
rename : media/libjpeg/simd/jdcolmmx.asm => media/libjpeg/simd/jdcolor-mmx.asm
rename : media/libjpeg/simd/jdcolss2-64.asm => media/libjpeg/simd/jdcolor-sse2-64.asm
rename : media/libjpeg/simd/jdcolss2.asm => media/libjpeg/simd/jdcolor-sse2.asm
rename : media/libjpeg/simd/jdmermmx.asm => media/libjpeg/simd/jdmerge-mmx.asm
rename : media/libjpeg/simd/jdmerss2-64.asm => media/libjpeg/simd/jdmerge-sse2-64.asm
rename : media/libjpeg/simd/jdmerss2.asm => media/libjpeg/simd/jdmerge-sse2.asm
rename : media/libjpeg/simd/jdmrgmmx.asm => media/libjpeg/simd/jdmrgext-mmx.asm
rename : media/libjpeg/simd/jdmrgss2-64.asm => media/libjpeg/simd/jdmrgext-sse2-64.asm
rename : media/libjpeg/simd/jdmrgss2.asm => media/libjpeg/simd/jdmrgext-sse2.asm
rename : media/libjpeg/simd/jdsammmx.asm => media/libjpeg/simd/jdsample-mmx.asm
rename : media/libjpeg/simd/jdsamss2-64.asm => media/libjpeg/simd/jdsample-sse2-64.asm
rename : media/libjpeg/simd/jdsamss2.asm => media/libjpeg/simd/jdsample-sse2.asm
rename : media/libjpeg/simd/jf3dnflt.asm => media/libjpeg/simd/jfdctflt-3dn.asm
rename : media/libjpeg/simd/jfsseflt-64.asm => media/libjpeg/simd/jfdctflt-sse-64.asm
rename : media/libjpeg/simd/jfsseflt.asm => media/libjpeg/simd/jfdctflt-sse.asm
rename : media/libjpeg/simd/jfmmxfst.asm => media/libjpeg/simd/jfdctfst-mmx.asm
rename : media/libjpeg/simd/jfss2fst-64.asm => media/libjpeg/simd/jfdctfst-sse2-64.asm
rename : media/libjpeg/simd/jfss2fst.asm => media/libjpeg/simd/jfdctfst-sse2.asm
rename : media/libjpeg/simd/jfmmxint.asm => media/libjpeg/simd/jfdctint-mmx.asm
rename : media/libjpeg/simd/jfss2int-64.asm => media/libjpeg/simd/jfdctint-sse2-64.asm
rename : media/libjpeg/simd/jfss2int.asm => media/libjpeg/simd/jfdctint-sse2.asm
rename : media/libjpeg/simd/ji3dnflt.asm => media/libjpeg/simd/jidctflt-3dn.asm
rename : media/libjpeg/simd/jisseflt.asm => media/libjpeg/simd/jidctflt-sse.asm
rename : media/libjpeg/simd/jiss2flt-64.asm => media/libjpeg/simd/jidctflt-sse2-64.asm
rename : media/libjpeg/simd/jiss2flt.asm => media/libjpeg/simd/jidctflt-sse2.asm
rename : media/libjpeg/simd/jimmxfst.asm => media/libjpeg/simd/jidctfst-mmx.asm
rename : media/libjpeg/simd/jiss2fst-64.asm => media/libjpeg/simd/jidctfst-sse2-64.asm
rename : media/libjpeg/simd/jiss2fst.asm => media/libjpeg/simd/jidctfst-sse2.asm
rename : media/libjpeg/simd/jimmxint.asm => media/libjpeg/simd/jidctint-mmx.asm
rename : media/libjpeg/simd/jiss2int-64.asm => media/libjpeg/simd/jidctint-sse2-64.asm
rename : media/libjpeg/simd/jiss2int.asm => media/libjpeg/simd/jidctint-sse2.asm
rename : media/libjpeg/simd/jimmxred.asm => media/libjpeg/simd/jidctred-mmx.asm
rename : media/libjpeg/simd/jiss2red-64.asm => media/libjpeg/simd/jidctred-sse2-64.asm
rename : media/libjpeg/simd/jiss2red.asm => media/libjpeg/simd/jidctred-sse2.asm
rename : media/libjpeg/simd/jcqnt3dn.asm => media/libjpeg/simd/jquant-3dn.asm
rename : media/libjpeg/simd/jcqntmmx.asm => media/libjpeg/simd/jquant-mmx.asm
rename : media/libjpeg/simd/jcqntsse.asm => media/libjpeg/simd/jquant-sse.asm
rename : media/libjpeg/simd/jcqnts2f-64.asm => media/libjpeg/simd/jquantf-sse2-64.asm
rename : media/libjpeg/simd/jcqnts2f.asm => media/libjpeg/simd/jquantf-sse2.asm
rename : media/libjpeg/simd/jcqnts2i-64.asm => media/libjpeg/simd/jquanti-sse2-64.asm
rename : media/libjpeg/simd/jcqnts2i.asm => media/libjpeg/simd/jquanti-sse2.asm
2015-01-20 22:19:57 -05:00

208 lines
8.8 KiB
C

/*
* jdct.h
*
* This file was part of the Independent JPEG Group's software:
* Copyright (C) 1994-1996, Thomas G. Lane.
* It was modified by The libjpeg-turbo Project to include only code relevant
* to libjpeg-turbo.
* For conditions of distribution and use, see the accompanying README file.
*
* This include file contains common declarations for the forward and
* inverse DCT modules. These declarations are private to the DCT managers
* (jcdctmgr.c, jddctmgr.c) and the individual DCT algorithms.
* The individual DCT algorithms are kept in separate files to ease
* machine-dependent tuning (e.g., assembly coding).
*/
/*
* A forward DCT routine is given a pointer to a work area of type DCTELEM[];
* the DCT is to be performed in-place in that buffer. Type DCTELEM is int
* for 8-bit samples, INT32 for 12-bit samples. (NOTE: Floating-point DCT
* implementations use an array of type FAST_FLOAT, instead.)
* The DCT inputs are expected to be signed (range +-CENTERJSAMPLE).
* The DCT outputs are returned scaled up by a factor of 8; they therefore
* have a range of +-8K for 8-bit data, +-128K for 12-bit data. This
* convention improves accuracy in integer implementations and saves some
* work in floating-point ones.
* Quantization of the output coefficients is done by jcdctmgr.c. This
* step requires an unsigned type and also one with twice the bits.
*/
#if BITS_IN_JSAMPLE == 8
#ifndef WITH_SIMD
typedef int DCTELEM; /* 16 or 32 bits is fine */
typedef unsigned int UDCTELEM;
typedef unsigned long long UDCTELEM2;
#else
typedef short DCTELEM; /* prefer 16 bit with SIMD for parellelism */
typedef unsigned short UDCTELEM;
typedef unsigned int UDCTELEM2;
#endif
#else
typedef INT32 DCTELEM; /* must have 32 bits */
typedef unsigned long long UDCTELEM2;
#endif
/*
* An inverse DCT routine is given a pointer to the input JBLOCK and a pointer
* to an output sample array. The routine must dequantize the input data as
* well as perform the IDCT; for dequantization, it uses the multiplier table
* pointed to by compptr->dct_table. The output data is to be placed into the
* sample array starting at a specified column. (Any row offset needed will
* be applied to the array pointer before it is passed to the IDCT code.)
* Note that the number of samples emitted by the IDCT routine is
* DCT_scaled_size * DCT_scaled_size.
*/
/* typedef inverse_DCT_method_ptr is declared in jpegint.h */
/*
* Each IDCT routine has its own ideas about the best dct_table element type.
*/
typedef MULTIPLIER ISLOW_MULT_TYPE; /* short or int, whichever is faster */
#if BITS_IN_JSAMPLE == 8
typedef MULTIPLIER IFAST_MULT_TYPE; /* 16 bits is OK, use short if faster */
#define IFAST_SCALE_BITS 2 /* fractional bits in scale factors */
#else
typedef INT32 IFAST_MULT_TYPE; /* need 32 bits for scaled quantizers */
#define IFAST_SCALE_BITS 13 /* fractional bits in scale factors */
#endif
typedef FAST_FLOAT FLOAT_MULT_TYPE; /* preferred floating type */
/*
* Each IDCT routine is responsible for range-limiting its results and
* converting them to unsigned form (0..MAXJSAMPLE). The raw outputs could
* be quite far out of range if the input data is corrupt, so a bulletproof
* range-limiting step is required. We use a mask-and-table-lookup method
* to do the combined operations quickly. See the comments with
* prepare_range_limit_table (in jdmaster.c) for more info.
*/
#define IDCT_range_limit(cinfo) ((cinfo)->sample_range_limit + CENTERJSAMPLE)
#define RANGE_MASK (MAXJSAMPLE * 4 + 3) /* 2 bits wider than legal samples */
/* Extern declarations for the forward and inverse DCT routines. */
EXTERN(void) jpeg_fdct_islow (DCTELEM * data);
EXTERN(void) jpeg_fdct_ifast (DCTELEM * data);
EXTERN(void) jpeg_fdct_float (FAST_FLOAT * data);
EXTERN(void) jpeg_idct_islow
(j_decompress_ptr cinfo, jpeg_component_info * compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col);
EXTERN(void) jpeg_idct_ifast
(j_decompress_ptr cinfo, jpeg_component_info * compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col);
EXTERN(void) jpeg_idct_float
(j_decompress_ptr cinfo, jpeg_component_info * compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col);
EXTERN(void) jpeg_idct_7x7
(j_decompress_ptr cinfo, jpeg_component_info * compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col);
EXTERN(void) jpeg_idct_6x6
(j_decompress_ptr cinfo, jpeg_component_info * compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col);
EXTERN(void) jpeg_idct_5x5
(j_decompress_ptr cinfo, jpeg_component_info * compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col);
EXTERN(void) jpeg_idct_4x4
(j_decompress_ptr cinfo, jpeg_component_info * compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col);
EXTERN(void) jpeg_idct_3x3
(j_decompress_ptr cinfo, jpeg_component_info * compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col);
EXTERN(void) jpeg_idct_2x2
(j_decompress_ptr cinfo, jpeg_component_info * compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col);
EXTERN(void) jpeg_idct_1x1
(j_decompress_ptr cinfo, jpeg_component_info * compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col);
EXTERN(void) jpeg_idct_9x9
(j_decompress_ptr cinfo, jpeg_component_info * compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col);
EXTERN(void) jpeg_idct_10x10
(j_decompress_ptr cinfo, jpeg_component_info * compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col);
EXTERN(void) jpeg_idct_11x11
(j_decompress_ptr cinfo, jpeg_component_info * compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col);
EXTERN(void) jpeg_idct_12x12
(j_decompress_ptr cinfo, jpeg_component_info * compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col);
EXTERN(void) jpeg_idct_13x13
(j_decompress_ptr cinfo, jpeg_component_info * compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col);
EXTERN(void) jpeg_idct_14x14
(j_decompress_ptr cinfo, jpeg_component_info * compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col);
EXTERN(void) jpeg_idct_15x15
(j_decompress_ptr cinfo, jpeg_component_info * compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col);
EXTERN(void) jpeg_idct_16x16
(j_decompress_ptr cinfo, jpeg_component_info * compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col);
/*
* Macros for handling fixed-point arithmetic; these are used by many
* but not all of the DCT/IDCT modules.
*
* All values are expected to be of type INT32.
* Fractional constants are scaled left by CONST_BITS bits.
* CONST_BITS is defined within each module using these macros,
* and may differ from one module to the next.
*/
#define ONE ((INT32) 1)
#define CONST_SCALE (ONE << CONST_BITS)
/* Convert a positive real constant to an integer scaled by CONST_SCALE.
* Caution: some C compilers fail to reduce "FIX(constant)" at compile time,
* thus causing a lot of useless floating-point operations at run time.
*/
#define FIX(x) ((INT32) ((x) * CONST_SCALE + 0.5))
/* Descale and correctly round an INT32 value that's scaled by N bits.
* We assume RIGHT_SHIFT rounds towards minus infinity, so adding
* the fudge factor is correct for either sign of X.
*/
#define DESCALE(x,n) RIGHT_SHIFT((x) + (ONE << ((n)-1)), n)
/* Multiply an INT32 variable by an INT32 constant to yield an INT32 result.
* This macro is used only when the two inputs will actually be no more than
* 16 bits wide, so that a 16x16->32 bit multiply can be used instead of a
* full 32x32 multiply. This provides a useful speedup on many machines.
* Unfortunately there is no way to specify a 16x16->32 multiply portably
* in C, but some C compilers will do the right thing if you provide the
* correct combination of casts.
*/
#ifdef SHORTxSHORT_32 /* may work if 'int' is 32 bits */
#define MULTIPLY16C16(var,const) (((INT16) (var)) * ((INT16) (const)))
#endif
#ifdef SHORTxLCONST_32 /* known to work with Microsoft C 6.0 */
#define MULTIPLY16C16(var,const) (((INT16) (var)) * ((INT32) (const)))
#endif
#ifndef MULTIPLY16C16 /* default definition */
#define MULTIPLY16C16(var,const) ((var) * (const))
#endif
/* Same except both inputs are variables. */
#ifdef SHORTxSHORT_32 /* may work if 'int' is 32 bits */
#define MULTIPLY16V16(var1,var2) (((INT16) (var1)) * ((INT16) (var2)))
#endif
#ifndef MULTIPLY16V16 /* default definition */
#define MULTIPLY16V16(var1,var2) ((var1) * (var2))
#endif