Part 1: Humanoid animation (HAnim) architecture
6.1 GeneralThis clause specifies the abstract definitions for the five objects specified by this document. Actual usage of these interfaces requires a binding to a representation system. Two such systems are specified in Annex C VRML binding and Annex D X3D binding. Use of this document in VRML worlds (see ISO/IEC 14772-1) shall use the binding and encoding specified in Annex C VRML binding. Use of this document in an X3D world shall use the binding specified in 26 Humanoid animation (HAnim) component of ISO/IEC 19775-1 and shall be encoded as specified in ISO/IEC 19776. Other bindings are allowed but are not specified in this document.
Table 6.1 lists the topics of this clause.
Each object interface defined below is presented as follows:
interface ObjectName {
datatype field_1_name field_1_default_value field_1_range_limits
... ... ... ...
datatype field_n_name field_n_default_value field_n_range_limits
}
Empty string data types are represented by two consecutive double quotation marks (""). Empty sequences are represented by open and closing brackets with no intervening values ([]).
For fields accepting numerical values, range limits are specified as lower and upper values, enclosed in either parentheses of square brackets. A parenthesis specifies that a limit is exclusive, and a square bracket specifies that the limit is inclusive. For fields that have multiple values the limits apply to each value.
EXAMPLE If a range is specified as [0,∞) then the field may take any value from, and including, 0, up to, but not including, infinity.
For fields that are Object data types, acceptable object types
are listed in square brackets. Objects of any of these listed data types may
be used in any quantity, and in any order, unless otherwise specified for
that particular field.
For fields that are string data types, acceptable values (if any
are specified) are listed in square brackets. Each individual acceptable
value is listed in double quotation marks, with values separated by vertical
bars (|) to indicate that any one of these values is acceptable. If no
acceptable values are specified then any value may be used, provided it
meets the requirements specified in this document for that
particular field.
Any fields that are not listed in this document shall be ignored. Fields may be added in subsequent parts of ISO/IEC 19774, in which case those fields become part of the object definition and shall be supported if those parts of ISO/IEC 19774 are supported by the implementation.
6.2 HumanoidThe Humanoid object is a container for the components that define the skeleton, geometry and landmarks of the HAnim figure. These components, which include the Joint, Segment, Site and Displacer objects, specify the majority of the details of the HAnim figure, while the other fields of the Humanoid object specify information about the author, copyright, and usage restrictions of the model.
This section describes the interface of the Humanoid object and the fields which allow the functionality of the object to be defined and accessed.
interface Humanoid {
float[3] bboxCenter 0 0 0 (-∞,∞)
float[3] bboxSize -1 -1 -1 [0,∞) or -1
float[3] center 0 0 0 (-∞,∞)
string description ""
sequence<string> info []
sequence<float[3]> jointBindingPositions [0 0 0] (-∞,∞)
sequence<float[4]> jointBindingRotations [0 0 1 0] (-∞,∞)
sequence<float[3]> jointBindingScales [1 1 1] (0,∞)
sequence<Object> joints [] [Joint]
integer loa -1 [-1,4]
string name required name
float[4] rotation 0 0 1 0 (-∞,∞)
float[3] scale 1 1 1 (0,∞)
float[4] scaleOrientation 0 0 1 0 (-∞,∞)
sequence<Object> segments [] [Segment]
sequence<Object> sites [] [Site]
string skeletalConfiguration ["BASIC"]
sequence<Object> skeleton [] [Joint, Site]
sequence<Object> skin [] [indexed mesh objects as defined by the representation]
sequence<float[3]> skinBindingCoords [] (-∞,∞)
sequence<float[4]> skinBindingNormals [] (-∞,∞)
sequence<float[3]> skinCoord [] (-∞,∞)
sequence<float[3]> skinNormal [] (-∞,∞)
float[3] translation 0 0 0 (-∞,∞)
string version "2.0" ["2.0"]
sequence<Object> viewpoints [] [Site]
}
The bboxCenter and bboxSize fields specify a bounding box that encloses the children of the Humanoid object. These fields are hints and if defined may be used for optimization purposes. A default bboxSize value, (-1, -1, -1), implies that the bounding box is not specified and, if needed, shall be calculated by the browser.
The center field specifies a translation offset from the origin of the local coordinate system (0,0,0). The rotation field specifies a rotation of the coordinate system of the humanoid figure. The scale field specifies a non-uniform scale of the humanoid figure coordinate system and the scale values shall be greater than zero. The scaleOrientation specifies a rotation of the coordinate system of the humanoid figure before the scale (to specify scales in arbitrary orientations). The scaleOrientation applies only to the scale operation. The translation field specifies a translation to the coordinate system of the entire humanoid figure.
The description field describes the type of humanoid model defined by the Humanoid object.
The info field consists of a sequence of strings, each of which is of the form term=value. The following terms are defined in this document:
authorName authorEmail copyright creationDate usageRestrictions humanoidVersion age gender height weight
The gender term typically has a value of female,
male or neuter. The humanoidVersion
term refers to the version of the humanoid being used, in order to track
revisions to the data. It is not the same as the version field of the
Humanoid object, which refers to the version of
the HAnim specification that was used when building the humanoid.
Additional term=value pairs may be included as
needed for specific applications.
The loa field describes the Level Of Articulation to which the contained Humanoid object complies (the given LOA number, 0 through 4). A value of -1 indicates that no such correspondence occurs.
The name field, which shall be present, and which shall not have the empty string value, stores the name of the humanoid defined by the Humanoid object. Of the other fields, only the skeleton and version fields are required. The remaining fields are optional.
The skeleton field contains the humanoid_root
Joint object. The
Humanoid object is considered the parent object
of the humanoid_root Joint object
and defines a coordinate space for the humanoid_root
Joint object. Thus, the
Humanoid object's transformation affects the
Joint object hierarchy in the skeleton
field. A hierarchy of Joint objects is defined
for each HAnim humanoid figure within the skeleton field of the
Humanoid object and a hierarchical definition of
joints is present even when the geometry of the humanoid figure is not
defined within the skeleton field. The skeleton field can also
contain zero or more Site objects which allow
landmarks to be established with respect to the overall humanoid figure, but
which are not affected by the movement of any of the
Joint objects.
The joints field contains a list of references, one for each Joint object defined within the skeleton field hierarchy of the Humanoid object. The order in which the joints are listed is irrelevant since the names of the joints are stored in the Joint objects themselves.
The segments field contains a list of references, one for each Segment object defined within the skeleton field hierarchy of the Humanoid object. The order in which the segments are listed is irrelevant since the names of the segments are stored in the Segment objects themselves.
The sites field contains a list of references, one for each Site object defined within the skeleton field hierarchy of the Humanoid object. The order in which the sites are listed is irrelevant since the names of the sites are stored in the Site objects themselves.
The references to contained Joint, Segment and Site objects may be used to support inverse kinematics (IK) and other motion tools that animate nodes found within the skeleton field.
The skeletalConfiguration field identifies the skeletal configuration
of this model. Any models that share the same skeletalConfiguration
value can also share skeletal animation data. The only specific skeletal
configuration defined in this document is the
"BASIC" skeletal configuration specified in
4.8.1 Overview. Since the
default value of the skeletal configuration is
"BASIC", a humanoid model that specifies a
skeletalConfiguration value of "BASIC" shall
conform to the restrictive skeletal model specified in
4.8.1 Overview. If
skeletalConfiguration is set to any value other than
"BASIC", the model shall also contain one or more of
the five fields that are required for models that are not
"BASIC" as described below.
The skin field contains one or more indexed mesh definitions. Those indexed mesh definitions utilize the point and normal data that is defined within the skinCoord and skinNormal fields, respectively, of the Humanoid object. This field is defined as an generic type for which the specific representation is defined by each binding to a presentation system. Annex C VRML binding contains a binding for VRML. Annex D X3D binding describes the Humanoid animation component of X3D specified in ISO/IEC 19775-1.
The skinCoord field contains a single sequence of points which are used by the internal mechanisms of the Humanoid object to create the appropriate surface deformations as well as the indexed mesh definitions within the skin field which do the actual rendering of the surface geometry.
The skinNormal field contains the normal data definition which is used by the internal mechanisms of the Humanoid object to create the appropriate surface deformations as well as the indexed mesh definitions within the skin field, which contains the actual surface geometry that is rendered.
The version field stores the version of this document to which the Humanoid object conforms. The version value for this document is "2.0".
The viewpoints field has a different functionality and behaviour than the joints, segments and sites fields. The viewpoints field can contain zero or more Site object definitions. The Site objects defined within this field are affected by the transformations applied to the Humanoid object, but are not affected by any of the transformations performed on the skeletal hierarchy defined within the skeleton field. The purpose of these Site objects is to define locations for virtual cameras in the reference frame of the Humanoid object. Site objects that are defined in this manner are intended to be used as attachment points from which a certain viewing perspective can be seen (such as viewing the face or viewing a profile of the human figure).
Since the translations and rotations of all the joints is zero at the binding pose, there is no transformation required to calculate the positions and normals of the vertices in the local coordinate system of the Joint objects. Models that contain non-zero values for the translations and rotations of the Joint objects at the binding pose shall also provide the additional fields necessary for the implementation to calculate the positions and normals of the vertices in the local coordinate system of the Joint objects.
The following five fields provide information needed for each binding pose of a humanoid or non-human model, as specified in 4.8.3 Modelling of non-human HAnim figures.
None of these five fields shall be used in a model that has a
skeletalConfiguration value of "BASIC".
The jointBindingPositions, jointBindingRotations, and jointBindingScales fields specify arrays of positions, rotations, and scale values, respectively. These sets of attributes are associated with the array of Joint objects contained in the joints field. If only one value is provided (such as the default value) then it is applied to all listed Joint objects equivalently. Applying each set of these translation, rotation, and scale values, in order, to the corresponding Joint objects maps a skeleton to the binding pose.
The skinBindingCoords field specifies an array of vertex locations that define the mesh in the binding pose. For each set of values corresponding to a given pose, the same vertex indexing system that is used for the skinCoord field also applies to this field.
The skinBindingNormals field specifies an array of normal vectors that define the mesh normal vectors in the binding pose. For each set of values corresponding to a given pose, the same normal indexing system used for the skinNormal field also applies to this field.
These five fields specify the necessary information so that both the skin and the skeleton can be in the same binding pose. Thus, implementations can calculate the location and normal of each vertex in the local coordinate system of each Joint object, a step required in the skin deformation process.
Support for additional fields used for Motion Data Animation is specified in ISO/IEC 19774-2.
6.3 JointThe Joint object is the fundamental building block that represents articulated body parts.
This section describes the interface of the Joint object and the fields that allow the functionality of the object to be defined and accessed.
interface Joint {
float[3] bboxCenter 0 0 0 (-∞,∞)
float[3] bboxSize -1 -1 -1 [0,∞) or -1
float[3] center 0 0 0 (-∞,∞)
sequence<Object> children [] [Joint, Segment]
string description ""
sequence<Object> displacers [] [Displacer]
float[3] llimit 0 0 0 (-∞,∞)
float[4] limitOrientation 0 0 1 0 (-∞,∞)
string name required name
float[4] rotation 0 0 1 0 (-∞,∞)
float[3] scale 1 1 1 (0,∞)
float[4] scaleOrientation 0 0 1 0 (-∞,∞)
sequence<integer> skinCoordIndex [] [0,∞)
sequence<float> skinCoordWeight [] [0,1]
float[3] stiffness 0 0 0 [0,1]
float[3] translation 0 0 0 (-∞,∞)
float[3] ulimit 0 0 0 (-∞,∞)
}
The bboxCenter and bboxSize fields specify a bounding box that encloses the children of the Joint object. These fields are hints and if defined may be used for optimization purposes. A default bboxSize value, (-1, -1, -1), implies that the bounding box is not specified and, if needed, shall be calculated by the browser.
The center field specifies a translation offset from the root of the overall humanoid body. Since all of the Site and Joint objects have their center values defined in the same coordinate frame, the length of any segment can be determined by calculating the distance between the parent Joint object's center field and the child Joint object's center field. The exception is length of segments at the ends of the fingers and toes, for which the distance to the end effector Site objects within the Segment object shall be used.
The children field specifies the child objects of a Joint object in a humanoid hierarchy. It is dependent on the levels of articulation and their hierarchies. Any number, including zero, of Joint objects, may be used. The children field may also contain a single Segment object.
The rotation field specifies a rotation of the coordinate system of the Joint object. The scale field specifies a non-uniform scale of the Joint object coordinate system and the scale values shall be greater than zero. The scaleOrientation specifies a rotation of the coordinate system of the Joint object before the scale. The scaleOrientation applies only to the scale operation. The translation field specifies a translation to the coordinate system of the Joint object.
The skinCoordWeight field contains a list of floating point values between 0.0 and 1.0 that describe an amount of weighting that shall be used to affect a particular vertex from the skinCoord field of the Humanoid object. Each item in this list has a corresponding index value in the skinCoordIndex field of the Joint object indicating exactly which coordinate is to be influenced.
The ulimit and llimit fields of the Joint object specify the upper and lower joint rotation limits. Both fields are three-element float sequences containing separate values for the X, Y and Z rotation limits. The ulimit field stores the upper (i.e. maximum) values for rotation around the X, Y and Z axes. The llimit field stores the lower (i.e., minimum) values for rotation around those axes. A sequence containing zero elements indicates that the joint is unconstrained. A non-empty sequence can only consist of a single three-value element. If ulimit and llimit values are provided, both must be defined together. If llimit and ulimit are equal, no constraints are applied to Joint object rotations.
The limitOrientation field gives the orientation of the coordinate frame in which the ulimit and llimit values are to be interpreted. The limitOrientation describes the orientation of ulimit and llimit fields in the local coordinate frame, relative to the Joint object centre position described by the center field.
The stiffness field, if present, contains values that specify to an inverse kinematics (IK) system how much each degree of freedom (DOF) shall scale the calculated joint rotation at each step of the IK solver.
EXAMPLE 1 A Joint object's stiffness can be used in an arm joint chain to give preference to moving the left wrist and left elbow over moving the left shoulder, or it can be used within a single Joint object with multiple degrees of freedom to give preference to individual degrees of freedom.
A scale factor of (1 - stiffness) is applied around the corresponding axis (X, Y, or Z for entries 0, 1 and 2 of the stiffness field). Thus a stiffness value of zero means that full rotation without scaling about the corresponding access can be computed. Similarly a stiffness value of one means that no rotation occurs about the corresponding axis, regardless of any computed value.
EXAMPLE 2 If a joint has a stiffness value of 0.5 for the X-axis DOF, then the joint moves half as much as it would have if the scaling value were not present (see [BADLER]).
Each Joint object shall have a name field,
which shall not have the empty string value, that is used for identifying
the object. Within the local scope of a Humanoid
object, each Joint object can be referenced by
its name alone (e.g., r_shoulder, l_hip, or
skullbase). However, when referring to a
Joint object within a larger or global scope,
the name of the Humanoid object shall be added
as a distinguishing prefix.
EXAMPLE 3 If a Humanoid
object's name field contains the value "joe",
the l_hip Joint object of that
Humanoid object is referred to globally as
"joe_l_hip".
The name field is the only field which shall be defined, all the other fields are optional. However, it shall be noted that those fields are still required to be part of the interface, even if none of the actual Joint objects define values for those fields.
The displacers field contains a list of Displacer objects that are used to morph the deformable mesh using the morph targets defined in the Displacer objects. See the Displacer object reference for more information.
The person creating a humanoid may wish to add additional joints to the body. The body remains humanoid in form, and is still generally expected to have the basic joints described in 4.9 Structure of a humanoid. The basic joints are known as the basic Joint object set to which additional Joint objects may be added (such as a prehensile tail). See 4.9.8 Additional Joint and Segment objects for more information.
If necessary, some of the joints (such as the many vertebrae) may be omitted. See Annex A Nominal body dimensions and levels of articulation for examples.
6.4 SegmentThe Segment object is a container for information about a segment of the body.
interface Segment {
float[3] bboxCenter 0 0 0 (-∞,∞)
float[3] bboxSize -1 -1 -1 [0,∞) or -1
float[3] centerOfMass 0 0 0 (-∞,∞)
sequence<Object> children [] [Site, indexed mesh objects as defined by the representation]
sequence<float[3]> coord [] (-∞,∞)
string description ""
sequence<Object> displacers [] [Displacer]
float mass 0 [0,∞) or -1
float[9] momentsOfInertia 0 0 0 0 0 0 0 0 0 (-∞,∞)
string name required name
}
The bboxCenter and bboxSize fields specify a bounding box that encloses the children of the Segment object. This is a hint that may be used for optimization purposes. A default bboxSize value of (-1, -1, -1) indicates that the bounding box is not specified and, if needed, shall be calculated by the implementing environment.
Each Segment object shall have a name
field, which shall not have the empty string value, that is used for
identifying the object. Within the local scope of a
Humanoid object, each
Segment object can be referenced by its name
alone (e.g., r_upperarm, l_thigh, or
skull). However, when referring to a
Segment object within a larger or global scope,
the name of the Humanoid object shall be added
as a distinguishing prefix.
EXAMPLE If a Humanoid
object's name field contains the value "joe",
the l_thigh Segment object of that
Humanoid is referred to globally as
"joe_l_thigh".
The name field is the only field which shall be defined, all the other fields are optional. However, it shall be noted that those fields are still required to be part of the interface, even if none of the actual Segment objects define values for those fields.
The mass field is the total mass of the segment. A value of -1 for mass specifies that no mass value is available.
The centerOfMass field is the location within the segment of its centre of mass.
The momentsOfInertia field contains the moment of inertia matrix. The first three elements are the first row of the 3×3 matrix, the next three elements are the second row, and the final three elements are the third row.
The coord field is used for Segment
objects that have deformable meshes and shall contain coordinates referenced
from the indexed mesh for the Segment object.
The coordinates are given the same name as the
Segment object, but with a
"_coords" appended (e.g.,
"skull_coords").
The displacers field stores the Displacer objects for a particular Segment object.
Any Segment object can also contain zero or more Site objects in the children field, which allow landmarks to be established with respect to the specific Segment. These Site objects are moved with the parent Segment but are not affected by skin deformation or displacer animations.
In addition to any Site objects the children field also specifies any indexed mesh geometry that is to be used in accordance with the skeletal body geometry specification detailed at 4.3.2 Skeletal body geometry specification.
6.5 SiteThe Site object identifies locations that serve as landmarks for attaching objects with semantics.
interface Site {
float[3] bboxCenter 0 0 0 (-∞,∞)
float[3] bboxSize -1 -1 -1 [0,∞) or -1
float[3] center 0 0 0 (-∞,∞)
sequence<Object> children [] [any object defined by the representation]
string description ""
string name required name
float[4] rotation 0 0 1 0 (-∞,∞)
float[3] scale 1 1 1 (0,∞)
float[4] scaleOrientation 0 0 1 0 (-∞,∞)
float[3] translation 0 0 0 (-∞,∞)
}
The name field is the only field which shall be defined, and which shall not have the empty string value. All the other fields are optional. However, it should be noted that those fields are still required to be part of the interface, even if none of the actual Site objects define values for those fields.
Each Site object shall have a name field
that is used for identifying the object, typically based on names in
B.2 Feature points.
Within the local scope of a
Humanoid object, each
Site object can be referenced by its name alone
(e.g., r_neck_base, l_femoral_lateral_epicondyles,
sellion, etc.). However, when referring to a
Site object within a larger or global scope,
the name of the Humanoid object shall be added
as a distinguishing prefix.
EXAMPLE If a Humanoid
object's name field contains the value "joe",
the l_index_distal_tip Site object
of that Humanoid is referred to globally as
"joe_l_index_distal_tip"".
Depending on the intended purpose, the Site
object's name field shall be altered to appropriately describe its
function. If used as an end effector, the Site
object shall have a name consisting of the name of the
Segment to which it is attached with an
"_tip" suffix appended.
EXAMPLE The end effector
Site on the right index finger shall be named
"r_index_distal_tip", and the
Site object shall be a child of the
"r_index_distal" Segment.
Site objects that are used to define viewpoint
locations shall have a "_view" suffix appended.
Site objects that are not end effectors and not
camera locations shall have an "_pt" suffix.
The bboxCenter and bboxSize fields specify a bounding box that encloses the children of the Site object. These fields are hints and if defined may be used for optimization purposes. A default bboxSize value, (-1, -1, -1), implies that the bounding box is not specified and, if needed, shall be calculated by the browser.
The center field specifies a translation offset from the root of the overall humanoid body. Since all of the Site and Joint objects objects have their center values defined in the same coordinate frame, the length of any segment can be determined by calculating the distance between the parent Joint object's center field and the child Joint object's center field. The exception is length of segments at the ends of the fingers and toes, for which the distance to the end effector Site objects within the Segment object shall be used.
The rotation field specifies a rotation of the coordinate system of the Site object.
The scale field specifies a non-uniform scale of the Site object coordinate system and the scale values shall be greater than zero.
The scaleOrientation specifies a rotation of the coordinate system of the Site object before the scale. The scaleOrientation applies only to the scale operation.
The translation field specifies a translation to the coordinate system of the Site object.
The children field is an generic container for any appropriate object defined by the representation system.
6.6 DisplacerThe Displacer object is a container for information that specifies the additional displacements of portions of an HAnim figure.
interface Displacer {
sequence<integer> coordIndex [] [0,∞)
string description ""
sequence<float[3]> displacements [] (-∞,∞)
string name required name
float weight 0 (-∞,∞)
}
Each Displacer object shall have a name field required for identifying the object. This field shall not have the empty string value. All the other fields are optional.
Within the local scope of a Humanoid object, each
Displacer object can be referred to by its name
alone (e.g., "l_eyebrow_feature" or
"l_eyebrow_raiser_action". However, when referring to
a Displacer object within a larger or global
scope, the name of the Humanoid object shall be
added as a distinguishing prefix.
EXAMPLE If a Humanoid
object's name field contains the value "joe",
the "l_eyebrow_feature"
Displacer object of that humanoid is referred to
globally as "joe_l_eyebrow_feature".
Displacer objects that are used to modify a
feature point location are typically based on names in
B.2 Feature points
and have a "_feature" suffix appended.
Displacer objects that are used to move a
feature shall be given a name with an "_action"
suffix, usually with an additional pre-suffix to indicate the kind of motion
(such as "l_eyebrow_raiser_action").
Displacer objects that correspond to a
particular configuration of the vertices shall have a
"_config" suffix.
For the Displacer objects that are contained in the displacers field of the Segment objects, the displacements are defined and applied in the Segment coordinate system. The base mesh for the morphed Segment is the original mesh defined in the Segment. For the Displacer objects that are contained in the displacers field of the Joint objects, the displacements are applied to the deformable mesh specified in the skinCoord field of the Humanoid object. The displacements are defined in the local coordinate system of the Joint object that contains the Displacer. The weighted displacements shall be applied to the skinCoord array in the coordinate system of the Humanoid object. Therefore, the displacements shall be transformed from the Joint object space into the Humanoid object space. The base mesh for the morphed skin is the deformed mesh that results from the bone-based skin deformation.
The coordIndex field contains the indices into the coordinate array for the mesh of the vertices that are affected by the Displacer.
EXAMPLE
Displacer {
coordIndex [ 7, 12, 21, 18 ]
name "l_eyebrow_feature"
}
means that vertices 7, 12, 21, and 18 of the mesh are from the left eyebrow.
The displacements field, if present, provides a set of 3D values that are added to the neutral or resting position of each of the vertices referenced in the coordIndex field of the associated mesh. These values correspond one-to-one with the values in the coordIndex array of the mesh. The values shall be maximum displacements, and the application shall uniformly scale them using the value of the weight field before adding them to the neutral vertex positions.
EXAMPLE
Displacer {
coordIndex [ 7, 12, 21, 18 ]
displacements [ 0 0.0025 0, 0 0.005 0, 0 0.0025 0, 0 0.001 0 ]
name "l_eyebrow_raiser_action"
}
raise the four vertices of the left eyebrow in a vertical direction. Vertex 7 is displaced up to 2.5 millimetres in the vertical (Y) direction, vertex 12 is displaced up to 5 millimetres, vertex 21 is displaced up to 2.5 millimetres, and vertex 18 is displaced by just one millimetre.
By dynamically modifying the value of the weight fields of the Displacer objects, an application can morph the mesh object using a linear combination of the morph targets defined in the Displacer objects.