Lens

class Lens

Bases: TypedWritableReferenceCount

A base class for any number of different kinds of lenses, linear and otherwise. Presently, this includes perspective and orthographic lenses.

A Lens object is the main part of a Camera node, which defines the fundamental interface to point-of-view for rendering. Lenses are also used in other contexts, however; for instance, a Spotlight is also defined using a lens.

Inheritance diagram

Inheritance diagram of Lens

enum FromCorners

These flags are passed in as the last parameter to control the behavior of set_frustum_from_corners(). See the documentation for that method for an explanation of each flag.

enumerator FC_roll = 1
enumerator FC_camera_plane = 2
enumerator FC_off_axis = 4
enumerator FC_aspect_ratio = 8
enumerator FC_shear = 16
enumerator FC_keystone = 32
enum StereoChannel
enumerator SC_mono = 0
enumerator SC_left = 1
enumerator SC_right = 2
enumerator SC_stereo = 3

== SC_left | SC_right

void clear(void)

Resets all lens parameters to their initial default settings.

void clear_custom_film_mat(void)

Disables the lens custom_film_mat correction.

void clear_keystone(void)

Disables the lens keystone correction.

void clear_view_mat(void)

Resets the lens transform to identity.

bool extrude(LPoint2 const &point2d, LPoint3 &near_point, LPoint3 &far_point) const
bool extrude(LPoint3 const &point2d, LPoint3 &near_point, LPoint3 &far_point) const

Given a 2-d point in the range (-1,1) in both dimensions, where (0,0) is the center of the lens and (-1,-1) is the lower-left corner, compute the corresponding vector in space that maps to this point, if such a vector can be determined. The vector is returned by indicating the points on the near plane and far plane that both map to the indicated 2-d point.

Returns true if the vector is defined, or false otherwise.

Given a 2-d point in the range (-1,1) in both dimensions, where (0,0) is the center of the lens and (-1,-1) is the lower-left corner, compute the corresponding vector in space that maps to this point, if such a vector can be determined. The vector is returned by indicating the points on the near plane and far plane that both map to the indicated 2-d point.

The z coordinate of the 2-d point is ignored.

Returns true if the vector is defined, or false otherwise.

bool extrude_depth(LPoint3 const &point2d, LPoint3 &point3d) const

Uses the depth component of the 3-d result from project() to compute the original point in 3-d space corresponding to a particular point on the lens. This exactly reverses project(), assuming the point does fall legitimately within the lens.

bool extrude_vec(LPoint2 const &point2d, LVector3 &vec3d) const
bool extrude_vec(LPoint3 const &point2d, LVector3 &vec3d) const

Given a 2-d point in the range (-1,1) in both dimensions, where (0,0) is the center of the lens and (-1,-1) is the lower-left corner, compute the vector that corresponds to the view direction. This will be parallel to the normal on the surface (the far plane) corresponding to the lens shape at this point.

See the comment block on Lens::extrude_vec_impl() for a more in-depth comment on the meaning of this vector.

Returns true if the vector is defined, or false otherwise.

Given a 2-d point in the range (-1,1) in both dimensions, where (0,0) is the center of the lens and (-1,-1) is the lower-left corner, compute the vector that corresponds to the view direction. This will be parallel to the normal on the surface (the far plane) corresponding to the lens shape at this point.

See the comment block on Lens::extrude_vec_impl() for a more in-depth comment on the meaning of this vector.

The z coordinate of the 2-d point is ignored.

Returns true if the vector is defined, or false otherwise.

PN_stdfloat get_aspect_ratio(void) const

Returns the aspect ratio of the Lens. This is determined based on the indicated film size; see set_film_size().

std::string const &get_change_event(void) const

Returns the name of the event that will be generated whenever any properties of this particular Lens have changed.

static TypeHandle get_class_type(void)
PN_stdfloat get_convergence_distance(void) const

See set_convergence_distance().

CoordinateSystem get_coordinate_system(void) const

Returns the coordinate system that all 3-d computations are performed within for this Lens. Normally, this is CS_default.

LMatrix4 const &get_custom_film_mat(void) const

Returns the custom_film_mat specified for the lens.

static PN_stdfloat get_default_far(void)

Returns the default far plane distance that will be assigned to each newly- created lens. This is read from the Config.prc file.

static PN_stdfloat get_default_near(void)

Returns the default near plane distance that will be assigned to each newly-created lens. This is read from the Config.prc file.

PN_stdfloat get_far(void) const

Returns the position of the far plane (or cylinder, sphere, whatever).

LMatrix4 const &get_film_mat(void) const

Returns the matrix that transforms from a point behind the lens to a point on the film.

LMatrix4 const &get_film_mat_inv(void) const

Returns the matrix that transforms from a point on the film to a point behind the lens.

LVector2 const &get_film_offset(void) const

Returns the horizontal and vertical offset amounts of this Lens. See set_film_offset().

LVecBase2 const &get_film_size(void) const

Returns the horizontal and vertical film size of the virtual film. See set_film_size().

PN_stdfloat get_focal_length(void) const

Returns the focal length of the lens. This may have been set explicitly by a previous call to set_focal_length(), or it may be computed based on the lens’ fov and film_size. For certain kinds of lenses, the focal length has no meaning.

LVecBase2 const &get_fov(void) const

Returns the horizontal and vertical film size of the virtual film. See set_fov().

PN_stdfloat get_hfov(void) const

Returns the horizontal component of fov only. See get_fov().

PN_stdfloat get_interocular_distance(void) const

See set_interocular_distance().

LVecBase2 const &get_keystone(void) const

Returns the keystone correction specified for the lens.

UpdateSeq get_last_change(void) const

Returns the UpdateSeq that is incremented whenever the lens properties are changed. As long as this number remains the same, you may assume the lens properties are unchanged.

LMatrix4 const &get_lens_mat(void) const

Returns the matrix that transforms from a point in front of the lens to a point in space.

LMatrix4 const &get_lens_mat_inv(void) const

Returns the matrix that transforms from a point in space to a point in front of the lens.

PN_stdfloat get_min_fov(void) const

Returns the field of view of the narrowest dimension of the window. See set_min_fov().

PN_stdfloat get_near(void) const

Returns the position of the near plane (or cylinder, sphere, whatever).

LPoint3 get_nodal_point(void) const

Returns the center point of the lens: the point from which the lens is viewing.

LMatrix4 const &get_projection_mat(Lens::StereoChannel channel = SC_mono) const

Returns the complete transformation matrix from a 3-d point in space to a point on the film, if such a matrix exists, or the identity matrix if the lens is nonlinear.

LMatrix4 const &get_projection_mat_inv(Lens::StereoChannel channel = SC_mono) const

Returns the matrix that transforms from a 2-d point on the film to a 3-d vector in space, if such a matrix exists.

LVector3 const &get_up_vector(void) const

Returns the axis perpendicular to the camera’s view vector that indicates the “up” direction.

PN_stdfloat get_vfov(void) const

Returns the vertical component of fov only. See get_fov().

LVecBase3 const &get_view_hpr(void) const

Returns the direction in which the lens is facing.

LMatrix4 const &get_view_mat(void) const

Returns the direction in which the lens is facing.

LVector3 const &get_view_vector(void) const

Returns the axis along which the lens is facing.

virtual bool is_linear(void) const

Returns true if the lens represents a linear projection (e.g. PerspectiveLens, OrthographicLens), and therefore there is a valid matrix returned by get_projection_mat(), or false otherwise.

virtual bool is_orthographic(void) const

Returns true if the lens represents a orthographic projection (i.e. it is a OrthographicLens), false otherwise.

virtual bool is_perspective(void) const

Returns true if the lens represents a perspective projection (i.e. it is a PerspectiveLens), false otherwise.

virtual PointerTo<BoundingVolume> make_bounds(void) const

Allocates and returns a new BoundingVolume that encloses the frustum used for this kind of lens, if possible. If a suitable bounding volume cannot be created, returns NULL.

virtual PointerTo<Lens> make_copy(void) const = 0
virtual PointerTo<Geom> make_geometry(void)

Allocates and returns a new Geom that can be rendered to show a visible representation of the frustum used for this kind of lens, if it makes sense to do so. If a visible representation cannot be created, returns NULL.

virtual void output(std::ostream &out) const
bool project(LPoint3 const &point3d, LPoint3 &point2d) const
bool project(LPoint3 const &point3d, LPoint2 &point2d) const

Given a 3-d point in space, determine the 2-d point this maps to, in the range (-1,1) in both dimensions, where (0,0) is the center of the lens and (-1,-1) is the lower-left corner.

Returns true if the 3-d point is in front of the lens and within the viewing frustum (in which case point2d is filled in), or false otherwise (in which case point2d will be filled in with something, which may or may not be meaningful).

Given a 3-d point in space, determine the 2-d point this maps to, in the range (-1,1) in both dimensions, where (0,0) is the center of the lens and (-1,-1) is the lower-left corner.

The z coordinate will also be set to a value in the range (-1, 1), where 1 represents a point on the near plane, and -1 represents a point on the far plane.

Returns true if the 3-d point is in front of the lens and within the viewing frustum (in which case point2d is filled in), or false otherwise (in which case point2d will be filled in with something, which may or may not be meaningful).

void recompute_all(void)

Forces all internal parameters of the Lens to be recomputed. Normally, this should never need to be called; it is provided only to assist in debugging.

void set_aspect_ratio(PN_stdfloat aspect_ratio)

Sets the aspect ratio of the lens. This is the ratio of the height to the width of the generated image. Setting this overrides the two-parameter fov or film size setting.

void set_change_event(std::string const &event)

Sets the name of the event that will be generated whenever any properties of the Lens have changed. If this is not set for a particular lens, no event will be generated.

The event is thrown with one parameter, the lens itself. This can be used to automatically track changes to camera fov, etc. in the application.

void set_convergence_distance(PN_stdfloat convergence_distance)

Sets the distance between between the camera plane and the point in the distance that the left and right eyes are both looking at. This distance is used to apply a stereo effect when the lens is rendered on a stereo display region. It only has an effect on a PerspectiveLens.

This parameter must be greater than 0, but may be as large as you like. It controls the distance at which the two stereo images will appear to converge, which is a normal property of stereo vision. Normally this should be set to the distance from the camera to the area of interest in your scene. Anything beyond this distance will appear to go into the screen, and anything closer will appear to come out of the screen. If you want to simulate parallel stereo, set this to infinity.

Note that this creates an off-axis frustum, which means that the lenses are still pointing in the same direction, which is usually more desirable than the more naive toe-in approach, where the two lenses are simply tilted toward each other.

Prior to Panda3D 1.9.0, the convergence was being calculated incorrectly. It has since been corrected. To restore the legacy behavior you can set the stereo-lens-old-convergence variable to true.

Also see set_interocular_distance(), which relates.

void set_coordinate_system(CoordinateSystem cs)

Specifies the coordinate system that all 3-d computations are performed within for this Lens. Normally, this is CS_default.

void set_custom_film_mat(LMatrix4 const &custom_film_mat)

Specifies a custom matrix to transform the points on the film after they have been converted into nominal film space (-1 .. 1 in U and V). This can be used to introduce arbitrary scales, rotations, or other linear transforms to the media plane. This is normally a 2-d matrix, but a full 4x4 matrix may be specified. This is applied on top of any film size, lens shift, and/or keystone correction.

void set_far(PN_stdfloat far_distance)

Defines the position of the far plane (or cylinder, sphere, whatever). Points farther from the lens than this may not be rendered.

void set_film_offset(PN_stdfloat x, PN_stdfloat y)
void set_film_offset(LVecBase2 const &film_offset)

Sets the horizontal and vertical offset amounts of this Lens. These are both in the same units specified in set_film_size().

This can be used to establish an off-axis lens.

void set_film_size(PN_stdfloat width)
void set_film_size(PN_stdfloat width, PN_stdfloat height)
void set_film_size(LVecBase2 const &film_size)

Sets the horizontal size of the film without changing its shape. The aspect ratio remains unchanged; this computes the vertical size of the film to automatically maintain the aspect ratio.

Sets the size and shape of the “film” within the lens. This both establishes the units used by calls like set_focal_length(), and establishes the aspect ratio of the frame.

In a physical camera, the field of view of a lens is determined by the lens’ focal length and by the size of the film area exposed by the lens. For instance, a 35mm camera exposes a rectangle on the film about 24mm x 36mm, which means a 50mm lens gives about a 40-degree horizontal field of view.

In the virtual camera, you may set the film size to any units here, and specify a focal length in the same units to simulate the same effect. Or, you may ignore this parameter, and specify the field of view and aspect ratio of the lens directly.

Sets the size and shape of the “film” within the lens. This both establishes the units used by calls like set_focal_length(), and establishes the aspect ratio of the frame.

In a physical camera, the field of view of a lens is determined by the lens’ focal length and by the size of the film area exposed by the lens. For instance, a 35mm camera exposes a rectangle on the film about 24mm x 36mm, which means a 50mm lens gives about a 40-degree horizontal field of view.

In the virtual camera, you may set the film size to any units here, and specify a focal length in the same units to simulate the same effect. Or, you may ignore this parameter, and specify the field of view and aspect ratio of the lens directly.

void set_focal_length(PN_stdfloat focal_length)

Sets the focal length of the lens. This may adjust the field-of-view correspondingly, and is an alternate way to specify field of view.

For certain kinds of lenses (e.g. OrthographicLens), the focal length has no meaning.

void set_fov(PN_stdfloat fov)
void set_fov(PN_stdfloat hfov, PN_stdfloat vfov)
void set_fov(LVecBase2 const &fov)

Sets the horizontal field of view of the lens without changing the aspect ratio. The vertical field of view is adjusted to maintain the same aspect ratio.

Sets the field of view of the lens in both dimensions. This establishes both the field of view and the aspect ratio of the lens. This is one way to specify the field of view of a lens; set_focal_length() is another way.

For certain kinds of lenses (like OrthoLens), the field of view has no meaning.

Sets the field of view of the lens in both dimensions. This establishes both the field of view and the aspect ratio of the lens. This is one way to specify the field of view of a lens; set_focal_length() is another way.

For certain kinds of lenses (like OrthographicLens), the field of view has no meaning.

void set_frustum_from_corners(LVecBase3 const &ul, LVecBase3 const &ur, LVecBase3 const &ll, LVecBase3 const &lr, int flags)

Sets up the lens to use the frustum defined by the four indicated points. This is most useful for a PerspectiveLens, but it may be called for other kinds of lenses as well.

The frustum will be rooted at the origin (or by whatever translation might have been specified in a previous call to set_view_mat).

It is legal for the four points not to be arranged in a rectangle; if this is the case, the frustum will be fitted as tightly as possible to cover all four points.

The flags parameter contains the union of one or more of the following bits to control the behavior of this function:

FC_roll - If this is included, the camera may be rotated so that its up vector is perpendicular to the top line. Otherwise, the standard up vector is used.

FC_camera_plane - This allows the camera plane to be adjusted to be as nearly perpendicular to the center of the frustum as possible. Without this bit, the orientation camera plane is defined by position of the four points (which should all be coplanar). With this bit, the camera plane is arbitrary, and may be chosen so that the four points do not themselves lie in the camera plane (but the points will still be within the frustum).

FC_off_axis - This allows the resulting frustum to be off-axis to get the tightest possible fit. Without this bit, the viewing axis will be centered within the frustum, but there may be more wasted space along the edges.

FC_aspect_ratio - This allows the frustum to be scaled non-proportionately in the vertical and horizontal dimensions, if necessary, to get a tighter fit. Without this bit, the current aspect ratio will be preserved.

FC_shear - This allows the frustum to be sheared, if necessary, to get the tightest possible fit. This may result in a parallelogram-based frustum, which will give a slanted appearance to the rendered image. Without this bit, the frustum will be rectangle-based.

In general, if 0 is passed in as the value for flags, the generated frustum will be a loose fit but sane; if -1 is passed in, it will be a tighter fit and possibly screwy.

void set_interocular_distance(PN_stdfloat interocular_distance)

Sets the distance between the left and right eyes of a stereo camera. This distance is used to apply a stereo effect when the lens is rendered on a stereo display region. It only has an effect on a PerspectiveLens.

The left eye and the right eye are each offset along the X axis by half of this distance, so that this parameter specifies the total distance between them.

Also see set_convergence_distance(), which relates.

void set_keystone(LVecBase2 const &keystone)

Indicates the ratio of keystone correction to perform on the lens, in each of three axes. This will build a special non-affine scale factor into the projection matrix that will compensate for keystoning of a projected image; this can be used to compensate for a projector that for physical reasons cannot be aimed directly at its screen.

The default value is taken from the default-keystone Config variable. 0, 0 indicates no keystone correction; specify a small value (usually in the range -1 .. 1) in either the x or y position to generate a keystone correction in that axis.

void set_min_fov(PN_stdfloat min_fov)

Sets the field of view of the smallest dimension of the window. If the window is wider than it is tall, this specifies the vertical field of view; if it is taller than it is wide, this specifies the horizontal field of view.

In many cases, this is preferable to setting either the horizontal or vertical field of view explicitly. Setting this parameter means that pulling the window wider will widen the field of view, which is usually what you expect to happen.

void set_near(PN_stdfloat near_distance)

Defines the position of the near plane (or cylinder, sphere, whatever). Points closer to the lens than this may not be rendered.

void set_near_far(PN_stdfloat near_distance, PN_stdfloat far_distance)

Simultaneously changes the near and far planes.

void set_view_hpr(PN_stdfloat h, PN_stdfloat p, PN_stdfloat r)
void set_view_hpr(LVecBase3 const &view_hpr)

Sets the direction in which the lens is facing. Normally, this is down the forward axis (usually the Y axis), but it may be rotated. This is only one way of specifying the rotation; you may also specify an explicit vector in which to look, or you may give a complete transformation matrix.

void set_view_mat(LMatrix4 const &view_mat)

Sets an arbitrary transformation on the lens. This replaces the individual transformation components like set_view_hpr().

Setting a transformation here will have a slightly different effect than putting one on the LensNode that contains this lens. In particular, lighting and other effects computations will still be performed on the lens in its untransformed (facing forward) position, but the actual projection matrix will be transformed by this matrix.

void set_view_vector(PN_stdfloat x, PN_stdfloat y, PN_stdfloat z, PN_stdfloat i, PN_stdfloat j, PN_stdfloat k)
void set_view_vector(LVector3 const &view_vector, LVector3 const &up_vector)

Specifies the direction in which the lens is facing by giving an axis to look along, and a perpendicular (or at least non-parallel) up axis.

See also set_view_hpr().

virtual void write(std::ostream &out, int indent_level = 0) const