Base class for rendering a 3D scene to an image. More...

Inherited by DD::Image::DepthMapRender, and DD::Image::RenderScene.

Classes
class	Span
Public Types
enum	{ SINGLE_SAMPLE, MULTISAMPLE_LOW, MULTISAMPLE_MED, MULTISAMPLE_HIGH }
Public Member Functions
virtual CameraOp *	render_camera (int sample=0)
virtual GeoOp *	render_geo (int sample=0)=0
virtual Matrix4	camera_matrix (int sample=0)
virtual Matrix4	projection_matrix (int sample=0)
virtual void	format_matrix (int sample=0)
Matrix4	get_format_matrix (float cx, float cy)
void	knobs (Knob_Callback)
void	display_knobs (Knob_Callback)
void	texture_knobs (Knob_Callback)
void	sampling_knobs (Knob_Callback)
void	jitter_knobs (Knob_Callback)
void	render_knobs (Knob_Callback)
	All of the above.
unsigned	node_color () const
double	overscanX ()
double	overscanY ()
int	projection_mode () const
int	blend_operation () const
const Filter *	texture_filter () const
int	multisampling () const
float	hitThresholdDistance () const
virtual int	multisamples () const
virtual Scene *	scene (int n=0) const
virtual double	shutter () const
virtual double	offset () const
virtual unsigned	samples () const
HandlesMode	doAnyHandles (ViewerContext *ctx)
void	build_handles (ViewerContext *)
void	buildGeoInfoHandle (ViewerContext *ctx, DrawHandleCallbackFunc cb)
virtual void	probe (const Vector3 &center, const VertexContext &vtx, Pixel &out)
	Sample the frame buffer.
Static Public Member Functions
static Matrix4	get_format_matrix (const Format &format, float cx, float cy)
static const Intersection *	multisample_array (int n)
static unsigned int	jitter_array_size (int samples)
static const Intersection *	jitter_array (int samples)
Static Public Attributes
static const char *	multisampling_table []
	Output sampling modes:
Protected Member Functions
	Render (Node *node, bool enable_mip_filter=false)
void	_validate (bool)
virtual void	overrideBBox (Box &f) const
void	_request (int x, int y, int r, int t, ChannelMask, int)
void	engine (int y, int x, int r, ChannelMask, Row &)
virtual bool	generate_render_primitives ()
virtual bool	evaluate_lights (Scene *scene)
virtual void	initialize ()
void	draw_primitives (int y, int x, int r, ChannelMask, Row &, const Row *bg)
void	_close ()
Protected Attributes
TextureFilter	texture_filter_
	Default filter to use in texture filtering.
int	projection_mode_
	Point projection mode.
int	max_tessellation_
	Threshold to limit recursive subdivision.
int	blend_operation_
	Default blending operation for textures.
int	multisampling_
	Output sampling modes.
int	samples_
	User set number of samples to render.
ShutterControls	shutter_
	shutter controls
double	spatial_jitter_
	User set jittering of sample in raster.
double	temporal_jitter_
	User set jittering of sample in time.
const Intersection *	jitter
	XY jitter array to use - based on total samples.
unsigned	input_scenes
	Number of scenes renderer gets from it's input.
unsigned	render_scenes
	Number of scenes to render.
const Intersection *	multisample_array_
	Multisampling array to use.
double	overscanX_
double	overscanY_
	overscan size of the frame buffer
Format	overscanFormat_
	overscan format
int	junk
Box3	world_bbox
	Intersected world-space bbox of all scenes.
Box	screen_bbox
	Intersected screen-space bbox of all scenes.
int	output_motion_vectors
	Whether renderer outputs motion vector channels.
Channel	MB_channel [3]
	Motion vector channels to output.
bool	generate_primitives
	Threading lock to keep other threads waiting.
std::vector< Scene * >	scenes_
	List of Scenes, one per sample.

Detailed Description

Base class for rendering a 3D scene to an image.

Constructor & Destructor Documentation

Render::Render	(	Node *	node,
		bool	enable_mip_filter = `false`
	)		`[protected]`

Base class constructor.

References DD::Image::Chan_U, DD::Image::eDisableMotionVector, generate_primitives, input_scenes, jitter, jitter_array(), DD::Image::CameraOp::LENS_RENDER_CAMERA, max_tessellation_, MB_channel, multisampling_, output_motion_vectors, overscanY_, projection_mode_, render_scenes, samples_, spatial_jitter_, and temporal_jitter_.

Member Function Documentation

CameraOp * Render::render_camera ( int sample = 0 ) [virtual]

Return a CameraOp for 'sample'. The base class returns null.

Reimplemented in DD::Image::RenderScene.

Referenced by _validate().

GeoOp * DD::Image::Render::render_geo ( int sample = 0 ) [pure virtual]

Return a GeoOp for 'sample'. This GeoOp will have it's cache enabled which will be created during _validate(). The base class returns 0.

Implemented in DD::Image::RenderScene.

Referenced by _validate().

Matrix4 Render::camera_matrix ( int sample = 0 ) [virtual]

Calculate the camera's inverse local matrix for 'sample'. Base class pulls the camera away in Z by the format's width, and XY centered. Remember, this is the camera's *inverse* matrix, so translations are reversed.

Reimplemented in DD::Image::DepthMapRender, and DD::Image::RenderScene.

References DD::Image::IopInfoOwner::format(), DD::Image::Format::height(), DD::Image::Matrix4::makeIdentity(), DD::Image::Matrix4::translate(), and DD::Image::Format::width().

Referenced by _validate().

Matrix4 Render::projection_matrix ( int sample = 0 ) [virtual]

Calculate the camera projection matrix for 'sample'. Base class implementation sets up a simple projection with a lens of 1.

Reimplemented in DD::Image::RenderScene.

References DD::Image::IopInfoOwner::format(), DD::Image::IopInfoOwner::h(), DD::Image::Format::height(), DD::Image::Matrix4::projection(), DD::Image::Matrix4::scaling(), DD::Image::IopInfoOwner::w(), and DD::Image::Format::width().

Referenced by _validate().

void Render::format_matrix ( int sample = 0 ) [virtual]

obsolete virtual function, do not use. Call get_format_matrix() instead

Matrix4 Render::get_format_matrix	(	float	cx,
		float	cy
	)

Calculate the transformation from post-projection NDC to pixel space. cx,cy are where in the pixel the -1,-1 corner should be.

References DD::Image::IopInfoOwner::info_.

Referenced by _validate(), and DD::Image::RenderScene::build_handles().

Matrix4 Render::get_format_matrix	(	const Format &	format,
		float	cx,
		float	cy
	)		`[static]`

Calculate the transformation from post-projection NDC to pixel space. cx,cy are where in the pixel the -1,-1 corner should be.

References DD::Image::IopInfoOwner::h(), DD::Image::Format::height(), DD::Image::Matrix4::makeIdentity(), DD::Image::Matrix4::scale(), DD::Image::Matrix4::translate(), DD::Image::IopInfoOwner::w(), and DD::Image::Format::width().

void Render::_validate ( bool for_real ) [protected, virtual]

Base-class implementation allocates a Scene structure for each time sample and enables the generation of primitives in the first engine() call.

Reimplemented from DD::Image::Iop.

Reimplemented in DD::Image::RenderScene.

References DD::Image::Scene::_id, DD::Image::Scene::_time, DD::Image::Op::aborted(), DD::Image::GeoOp::build_scene(), DD::Image::Scene::camera, camera_matrix(), DD::Image::IopInfoOwner::channels(), DD::Image::Box::clear(), DD::Image::Iop::copy_info(), DD::Image::CameraOp::default_camera(), DD::Image::eClassicMotionVector, DD::Image::eDisableMotionVector, evaluate_lights(), DD::Image::Box3::expand(), DD::Image::Scene::format_, generate_primitives, get_format_matrix(), DD::Image::Format::height(), DD::Image::IopInfoOwner::info_, DD::Image::Iop::input0(), input_scenes, DD::Image::Op::inputContext(), DD::Image::Box::intersect(), jitter, jitter_array(), DD::Image::Scene::lens_func, DD::Image::CameraOp::LENS_PERSPECTIVE, DD::Image::CameraOp::LENS_RENDER_CAMERA, DD::Image::CameraOp::lensNfunction(), MAX(), max_tessellation_, DD::Image::Scene::max_tessellation_, DD::Image::Scene::mb_scene_, DD::Image::Scene::mb_type_, DD::Image::Box::merge(), multisample_array(), multisample_array_, multisampling_, output_motion_vectors, overrideBBox(), overscanFormat_, overscanY_, DD::Image::Box::pad(), projection_matrix(), DD::Image::Scene::projection_mode_, projection_mode_, render_camera(), render_geo(), render_scenes, scenes_, screen_bbox, DD::Image::Box::set(), DD::Image::MatrixArray::set_camera_matrix(), DD::Image::MatrixArray::set_format_matrix(), DD::Image::Iop::set_out_channels(), DD::Image::MatrixArray::set_projection_matrix(), DD::Image::Op::slowness(), spatial_jitter_, texture_filter_, DD::Image::Scene::validate(), DD::Image::Op::validate(), DD::Image::Format::width(), and world_bbox.

virtual void DD::Image::Render::overrideBBox ( Box & f ) const [inline, protected, virtual]

called in Render::_validate, to allow the deriving class to override the bbox.

Referenced by _validate().

void Render::_request	(	int	x,
		int	y,
		int	r,
		int	t,
		ChannelMask	out_channels,
		int	count
	)		`[protected, virtual]`

Render just texture-maps the geometry with input0(). So the request figures out what uv's will be requested and that is the area needed.

Reimplemented from DD::Image::Iop.

Reimplemented in DD::Image::RenderScene.

References DD::Image::Box3::expand(), fast_floor(), DD::Image::IopInfoOwner::format(), DD::Image::GeoInfo::get_group_attribute(), DD::Image::Box::h(), DD::Image::IopInfoOwner::info(), DD::Image::Iop::input0(), DD::Image::Box3::r(), DD::Image::Iop::request(), DD::Image::Scene::request(), scenes_, DD::Image::GeometryList::size(), DD::Image::Box3::t(), DD::Image::Box::w(), DD::Image::Box::x(), and DD::Image::Box::y().

void Render::engine	(	int	y,
		int	x,
		int	r,
		ChannelMask	channels,
		Row &	out
	)		`[protected, virtual]`

Default engine calls initialize() (which calls generate_render_primitives()) and then calls draw_primitives() with bg=0.

Implements DD::Image::Iop.

References draw_primitives(), and initialize().

bool Render::generate_render_primitives ( ) [protected, virtual]

Initialize each sample Scene object and generate it's renderable primitives.

Reimplemented in DD::Image::DepthMapRender.

References DD::Image::Op::aborted(), DD::Image::Scene::delete_render_primitives(), DD::Image::Scene::generate_render_primitives(), DD::Image::GeoInfo::matrix, DD::Image::Scene::object_transforms_, scenes_, and DD::Image::Scene::transforms_.

Referenced by initialize().

bool Render::evaluate_lights ( Scene * scene ) [protected, virtual]

Evaluate all the scene lights.

References DD::Image::Scene::evaluate_lights().

Referenced by _validate().

void Render::initialize ( ) [protected, virtual]

engine() must call this before it can look at the primitives

References DD::Image::Op::aborted(), generate_primitives, generate_render_primitives(), and texture_filter_.

Referenced by engine().

void Render::draw_primitives	(	int	y,
		int	x,
		int	r,
		ChannelMask	channels,
		Row &	out_row,
		const Row *	bg
	)		`[protected]`

Flat renderer used by gridwarp. If bg is non-zero, it is a row containing a background image which is used anywhere there are no primitives (note that it does not do an OVER, instead the alpha channel is replaced by the rendered ones just like all other channels).

References DD::Image::Row::copy(), DD::Image::QuadtreeNode< T >::data, DD::Image::Row::erase(), DD::Image::Render::Span::initialize(), DD::Image::Iop::input0(), DD::Image::Quadtree< T >::intersections(), MAX(), MIN(), DD::Image::Op::node(), DD::Image::IopInfoOwner::r(), DD::Image::Render::Span::r, DD::Image::Box3::r(), DD::Image::Render::Span::rprimitive(), DD::Image::VertexContext::sample(), scenes_, DD::Image::Box3::t(), DD::Image::VertexContext::vdX, DD::Image::Render::Span::vtx, DD::Image::Render::Span::vX, DD::Image::Row::writable(), and DD::Image::IopInfoOwner::x().