Rendering Buglets

[DavidBrainard]

1) MakeRadianceTestFigure crashes. The reason is that it assumes that the width of the 8 images it analyzes is the same. But, the first of the images has width 160, while the rest have width 100. Since the code pulls the size from the first, it crashes on the second time through the loop where it plots slices, because the slice size doesn't match the number of pixels expected.

I don't know if the fix should be to the plots or to figure out why the sizes are different.

2) The RGBPromotion scene is writing temporary files into the toolbox (some spectra and some conditions.) I think these should go in the target directory, to keep with our principle of not modifying the toolbox itself.

[benjamin-heasly]

In f631654561ead560f63f066830b05ad8974a1459, MakeRGBPromotion no longer puts outputs in the current folder.

[benjamin-heasly]

In 8e71b83ba9e84e2907dec9568c43d0f8f1e259a8, MakeRadianceTestFigure tolerates images of different sizes.

I was not able to reproduce the problem where only the first image has size width 160. But I removed the assumption that all images have the same size. So now, when the images are different sizes, the script should still run, event though the analysis of radiance profiles might come out as nonsense.

The images should all be the same size. If this problem persists, we should figure out why some images have a different size. It probably has (or had) to do with incorrect propagation of image size specification through all stages of Collada conversion and rendering. 01e80c7dd0c617d4ecb93c636f3c7e8f7a2451b5 should have fixed a similar issue.