<--back--      home      --next-->
As we've seen, adding dimensions to the transfer function provides more flexibility and specificity in what parts of the volume contribute to the final rendering, which can help make a better visualization.

However, as you may know from last year's Transfer Function Bake-Off, setting transfer functions can be a little tricky, even with just 1D transfer functions, so adding dimensions can make this worse. Also, this relationship between values and gradients and second derivatives can be a little confusing, especially in the context of a complex combination of boundaries. Also, specifying any two or three dimensional function is very difficult task from a user interface standpoint.

[Gordon finished, Joe starts]