ConstFind the cage face and local coordinates of a surface coordinate on one sampled tier.
A patch's grid is even in (s, t), so the face holding a coordinate is
column floor(s * columns) and row floor(t * rows), the last column or
row taking s or t equal to 1, and the local coordinates are the
fractional parts. On a patch whose faces are no grid, a transition band,
the face is the first whose bilinear map of its corner coordinates reaches
the coordinate, inverted by Newton's method, and the local coordinates are
that map's parameters. They follow IAutoMovieSubdivisionParameter's quad
convention: (0, 0) at corner 0, u toward corner 1 along s, v toward
corner 3 along t, so the result addresses subdivideAutoMovieSurface's
output of this cage directly.
Borrowed sampled tier whose patch grids and face-corner charts define the lookup. Cage geometry is not projected or read; callers must preserve the chart topology.
Patch index plus dimensionless s/t in [0,1]. The locator rejects a missing patch or out-of-range coordinates; a non-grid patch throws if no face accepts the inverted bilinear coordinates.
The face's parameter, with corner null because every face is a quad.
Find the cage face and local coordinates of a surface coordinate on one sampled tier.
A patch's grid is even in
(s, t), so the face holding a coordinate is columnfloor(s * columns)and rowfloor(t * rows), the last column or row takingsortequal to 1, and the local coordinates are the fractional parts. On a patch whose faces are no grid, a transition band, the face is the first whose bilinear map of its corner coordinates reaches the coordinate, inverted by Newton's method, and the local coordinates are that map's parameters. They followIAutoMovieSubdivisionParameter's quad convention:(0, 0)at corner 0,utoward corner 1 alongs,vtoward corner 3 alongt, so the result addressessubdivideAutoMovieSurface's output of this cage directly.