Hi,

   To learn more, go to Paul Debevec's site. He has HDRShop online for
download.
There are several options, techniques, etc. available. The program, as the
name
might imply, is a versatile tool for converting images to/from HDR; creating
HDRI, creating light probes and mirror-balls (while they appear similar,
their
roles are different) and converting probes and balls to images. One of the
techniques is converting a series of standard photos to a software light
probe.
Sources of 3" ball bearings are mentioned, use ,and drawbacks of, lawn
ornament
mirror balls.
   Spend a couple of days studying the site and the material, work with
HDRShop and you'll have a better idea of what you need and what options are
open to you.
   My question? Does MLPOV give us access to all the features of HDRI
and HDRShop? Reflections, environmental lighting, motion blur, full
HDRI file format compatibility?

   Regarding the use of Povray spherical uv-mapping, the idea is to
map an environment on a sphere, capturing what could be considered,
both a vertical and horizontal panoramic map, combined. Just think
of an image of the earth globe with all the lines of latitude and longitude
and you will get an idea of the relation between panoramic images
and spherical surfaces.
HDRShop can consruct a light probe using a photo-cross. The cross-shape
defines the shape of a cube when opened and flattened. If you develop the
math, you can transform an image mapped on a cube to a sphere and back.
This is true for
any regular polyhedron (3d polygon).
   As the number of sides of a polygon increases, the figure approximates
a circle. A dodecagon more closely resembles a circle than a square.
The same is true for polyhedra. The more surfaces, the closer the
approximation of a sphere. It may be intuitively obvious that the more
closely a figure approaches a sphere, the less distortion involved.
Knowing this, you can deduce that a cube is almost, but not quite, the
minimum figure needed to map an image to a sphere and almost the greatest
distortion. This distortion is taken into account in the software.
The choice of the cube over, e.g. a dodecahedron, is that
cameras take quadrangle shaped (4-sides, usually not square) photographs.
   Anyway, study the material on the site and you will begin to understand
what you need to know.
   There is a mention of and a link to, a website with a
beginner/non-science based explanation. I've seen it and it does a good job
as an introduction to HDR.

                                      Jim Julian

Post a reply to this message