dimg_p.h File Reference

Go to the source code of this file.

Classes
class	Digikam::DImg
class	Digikam::DImgStaticPriv

Namespaces
	Digikam

Macros
#define	LANCZOS_DATA_ONE   4096
#define	LANCZOS_DATA_TYPE   int
#define	LANCZOS_SUPPORT   2
#define	LANCZOS_TABLE_RES   256

Typedefs
typedef int64_t	llong
typedef uint64_t	ullong

Macro Definition Documentation

LANCZOS_DATA_ONE

#define LANCZOS_DATA_ONE   4096

LANCZOS_DATA_TYPE

#define LANCZOS_DATA_TYPE   int

Define this to use a floating-point implementation of Lanczos interpolation. The integer implementation is a little bit less accurate, but MUCH faster (even on machines with FPU - ~2.5 times faster on Core2); besides, it will run a hell lot faster on computers without a FPU (e.g. PDAs).

LANCZOS_SUPPORT

#define LANCZOS_SUPPORT   2

A support of 3 gives an overall sharper looking image, but it is a) slower b) gives more sharpening artifacts

LANCZOS_TABLE_RES

#define LANCZOS_TABLE_RES   256

Lanczos kernel is precomputed in a table with this resolution The value below seems to be enough for HQ upscaling up to eight times

Typedef Documentation

llong

typedef int64_t llong

ullong

typedef uint64_t ullong