This function sets the location of the viewer.

observer3d(x, y = NULL, z = NULL, auto = FALSE)


x, y, z

The location as a 3 vector, using the usual xyz.coords conventions for specification. If x is missing or any coordinate is NA, no change will be made to the location.


If TRUE, the location will be set automatically by RGL to make the whole bounding box visible.


This function sets the location of the viewer relative to the scene, after the model transformations (scaling, rotation) have been done, but before lighting or projection have been applied. (See par3d for details on the rendering pipeline.)

The coordinate system is a slightly strange one: the X coordinate moves the observer location from left to right, and the Y coordinate moves up and down. The Z coordinate changes the depth from the viewer. All are measured relative to the center of the bounding box (par("bbox")) of the subscene. The observer always looks in the positive Z direction after the model rotation have been done. The coordinates are in post-scaling units.


This function is likely to change in future versions of RGL, to allow more flexibility in the specification of the observer's location and orientation.


Invisibly returns the previous value.


Duncan Murdoch


example(surface3d)  # The volcano data
#> srfc3d> #
#> srfc3d> # volcano example taken from "persp"
#> srfc3d> #
#> srfc3d> 
#> srfc3d> z <- 2 * volcano        # Exaggerate the relief
#> srfc3d> x <- 10 * (1:nrow(z))   # 10 meter spacing (S to N)
#> srfc3d> y <- 10 * (1:ncol(z))   # 10 meter spacing (E to W)
#> srfc3d> zlim <- range(z)
#> srfc3d> zlen <- zlim[2] - zlim[1] + 1
#> srfc3d> colorlut <- terrain.colors(zlen) # height color lookup table
#> srfc3d> col <- colorlut[ z - zlim[1] + 1 ] # assign colors to heights for each point
#> srfc3d> open3d()
#> null 
#>   64 
#> srfc3d> surface3d(x, y, z, color = col, back = "lines")
observer3d(0, 0, 440) # Viewed from very close up