PBRT_V2 总结记录 <93> Transforming Between Distributions In Multiple Dimensions

概述：

In the general n-dimensional case, a similar derivation gives the analogous relationship
between different densities. We will not show the derivation here; it follows the same
form as the one-dimensional case. Suppose we have an n-dimensional random variable
X with density function px(x). Now let Y = T (X), where T is a bijection. In this case,
the densities are related by

where |JT | is the absolute value of the determinant(行列式) of T ’s Jacobian matrix, which is

where Ti are defined by T (x) = (T1(x), . . . , Tn(x)).

 

Example：POLAR COORDINATES

The polar transformation is given by

Suppose we draw samples from some density p(r , θ).What is the corresponding density
p(x, y)? The Jacobian of this transformation is

and the determinant is r

cos θ ^ 2 + sin θ ^ 2= r.

So p(x, y) = p(r , θ)/r.

Of course, this is backwards from what we usually want—typically we start with a sampling strategy in
Cartesian coordinates and want to transform it to one in polar coordinates. In that case,
we would have

p(r , θ) = r p(x, y).

 

Example ： SPHERICAL COORDINATES

（这个例子就演示了

p(r , θ , φ) = r^2 sin θ p(x, y, z).

p(θ , φ) = sin θ p(ω). 的关系

）

Given the spherical coordinate representation of directions,

the Jacobian of this transformation has determinant |JT | = r2 sin θ, so the corresponding
density function is

This transformation is important since it helps us represent directions as points (x, y, z)
on the unit sphere. Remember that solid angle is defined as the area of a set of points on
the unit sphere. In spherical coordinates, we previously derived

So if we have a density function defined over a solid angle , this means that

The density with respect to θ and φ can therefore be derived: