Chrome Ball Photography and Shooting for Normal Maps

Hello everyone,

I have recently taken to photography as a means to enhance my 3D work in Autodesk Maya and I'm interested in specifically two things: Utilizing HDR lighting during the rendering process and generating normal maps with xNormal software.

For shooting skies in for HDR, my understanding is that an inexpensive chrome ball can be used. However, I have read that a full frame camera is the preferred option over a cropped frame DSLR due to the "flattening" effect produced by cameras with smaller than 35 equivalent sensors. I'm not quite sure what this is referring to but it seems to imply that my cropped frame camera is insufficient for the job. Can anyone confirm if a full frame is needed to produce "usable" images with chrome balls?

With regards to normal map generation, the Xnormal software is said to be able to create highly accurate normal maps by plugging in four photos that have been lit from the top, bottom, left and right directions using studio lighting. Would it be possible to achieve the same effect with just a reflector and built-in flash or would I need a more sophisticated setup? This would only be used for small objects with a camera mounted on a tripod, of course.

Thanks.

What you are describing does not sound like HDR (high dynamic range) imaging. Where you take several exposures at different shutter speeds (or ISO) values with the same light source to capture more information in the shadows and highlights that might otherwise be recorded as pure black or pure white.

What you are talking about with a "chrome ball" and moving the light source around sounds like creating Polynomial Texture Maps (PTMs) using Reflectance Transform Imaging (RTI).

I'm not exactly certain which "flattening" effect they are refering to, but I'd guess it's one of two things:

1. Smaller-than-full-frame sensors tend to have more depth of field (technically it's a lens thing, but still...) causing less separation and the shallow DOF provided by a full frame might be useful for these purposes
2. The same focal length lens on a smaller-than-full-frame camera will be more telephoto, which flattens the perspective.

I really doubt full frame is required, and "usable" means different things to different people. The only way that it would really cause problems is if the program was written to read the exif data for things like the focal length of the lens and does calculations based on full-frame sensor size. But if that's not the issue I don't see why you couldn't get a 28 or 35mm prime lens with a wider aperture that would work on an APS-C SLR.

I know people who are working on RTMs using 9 or 16 images. Some people are using the chrome ball (or in some cases billiard balls) others are using arrays of lights set up in a fixed pattern.

Here's some useful information:
http://www.c-h-i.org/technology/ptm/ptm.html

--
~K