4.3.4 ND-Transform Objects

Where – in the context of ND-viewing – a single (N+1)x(N+1) matrix is expected – as in the INST ntransform field, or the ND-xform* (see GCL) commands – use an ntransform object.

ntransform are NRows x NCols transformation matrix where usually NRows = N+1 in the context of N-dimensional objects and viewing. The homogeneous component of an ntransform sits at column zero (in contrast to ordinary transform objects where it is located at column three). ntransform objects operate on points of any dimension: if a point is to be transformed by an ntransform object and the dimension of the point does not match the number of rows of the ntransform object, then either the point is implicitly padded with zeros to match NRows or the matrix is implicitly padded with ones down its diagonal (and zeros everywhere else) such that it will operate as identity on the excess dimensions of the input point.

Syntax for an ntransform object is

     <ntransform> ::=
       [ "{" ]             (curly brace, generally needed to make
                            the end of the object unambiguous.)
     
        [ "ntransform" ]    (optional keyword; unnecessary if the type
                            is determined by the context, which it
                            usually is.)
     
        [ "define" <name> ]
                            (defines a transform named <name>, setting
                            its value from the stuff which follows)
     
          NRows NCols
                            (number of rows and columns of the matrix,
                            typically N+1 N+1, but any dimensions
                            are possible)
          <NRows x NCols floating-point numbers>
                            (interpreted as a NRows x NCols
                            homogeneous transform given row by row, intended
                            to apply to a row vector multiplied on its LEFT,
                            so that e.g. Euclidean translations appear in the
                            top row -- in contrast to the ordinary
                            transform objects where the translations appear
                            in the bottom row)
        |
           "<" <filename>  (meaning: read transform from that file)
        |
           ":" <name>      (meaning: use variable <name>,
                           defined elsewhere; if undefined the initial
                           value is the identity transform)
     
      [ "}" ]             (matching curly brace)

The whole should be enclosed in { braces }. Braces are not necessarily essential, so e.g. two integers – NRows NCols – followed by a NRows x NCols array of floats standing alone may but needn't have braces.

Some examples, in contexts where they might be used:

     # Example 1: A GCL command to define a 6x6 transform called
     # "fred", a mere translation by the vector -3 0 1 1 0. This
     # transform is meant for a five dimensional space, with an homogeneous
     # component a index zero.
     
     (read ntransform { ntransform  define fred
              6 6
              1 -3 0 1 1 0
              0  1 0 0 0 0
              0  0 1 0 0 0
              0  0 0 1 0 0
              0  0 0 0 1 0
              0  0 0 0 0 1
         }
     )

     # Example 2: Set the ND-xform of an object -- a geometry or a camera
     # cluster. Given the definition above, this puts the object at (-3 0 1 1
     # 0) in the five dimensional space.
     
     (ND-xform-set focus : fred)
     
     # or
     
     (ND-xform-set g1 : fred)