DD::Image::Blur Class Reference

Inherits DD::Image::Iop.

List of all members.

Public Types
enum	{ BOX, TRIANGLE, QUADRATIC, GAUSSIAN }
Public Member Functions
	Blur (Node *node)
void	_validate (bool)
void	_request (int x, int y, int r, int t, ChannelMask, int count)
void	engine (int y, int x, int r, ChannelMask, Row &)
void	_close ()
double	size () const
void	size (double v)
void	size (double v, double w)
double	size_x () const
void	size_x (double v)
double	size_y () const
void	size_y (double v)
int	type () const
void	type (int t)
virtual void	knobs (Knob_Callback)
const char *	Class () const
const char *	node_help () const
Static Public Member Functions
static void	make_filter (int type, float *array, int size, double diameter, double gamma=1.0)
static int	filter_size (int type, double diameter)
Static Public Attributes
static const char *const	filter_names []
static const Description	d
Protected Attributes
double	size_x_
double	size_y_
int	quality_
int	type_
double	falloff

---

Detailed Description

The "fast blur". Does a 2-pass convolution with a gaussian filter. If the amount of blur is large enough it uses DownRezIop and UpRezIop to scale it down so a smaller filter can be used.

Odd integer sizes when the filter is set to BOX make a box filter of exactly that size. Non-integer values cause the outer edge of the next larger box to gradually fade in. The filters used when the type is not BOX are larger and chosen to result in approximately the same amount of blurring as when box is used.

---

Member Function Documentation

void Blur::make_filter	(	int	type,
		float *	array,
		int	size,
		double	diameter,
		double	gamma = 1.0
	)		[static]

Fill in an array with the coefficients of a filter of the given type (one of BOX, TRIANGLE, QUADRATIC, or GAUSSIAN, possibly others will be added in the future). The size is how big the array is, if this is smaller than what filter_size() returned the filter has the ends truncated and is renormalized so the portion adds to one. If it is bigger the rest if filled with zero. size must be an odd number. diameter controls the filter width, for box filters integers indicate a box of exactly that size, all other numbers and filter types set the coefficients to resemble and smoothly interpolate those box filters.

int Blur::filter_size	(	int	type,
		double	diameter
	)		[static]

Return the recommended size of an array to pass to make_filter(). The returned value is always odd, so that the filter has a center value.

References DD::Image::IopInfoOwner::r().

---

Member Data Documentation

const char *const Blur::filter_names [static]

This array may be passed to Enumeration_knob if you want to make a control where the user selects one of the filter types that the Blur can generate with filter_size() and make_filter().