So the late afternoon sun was beaming in the window and passing through our 175 gallon fish tank a few days ago, and the prism effect was casting a rainbow spectrum on the side of a matte steel finish computer case sitting in the dark under the desk. The illuminated area was about 1 square foot so it was really easy to see a lot of detail. I though, "I should take a picture to see how the colors will be rendered against this natural standard". Once I took the first shot I decided to then take a series with various exposure compensation settings to see if it caused a shift in the color fidelity. This strip is -3, -2, -1, 0, +1, +2, +3 Subscribe to see EXIF info for this image (if available) E-P5 with Zuiko Digital 70-300 mm and EC-14 What really surprised me was that the yellow part of the rainbow was really hidden until some over-exposure was dialed in, but then it really took over the image. The other noticeable change was the blue shifting to cyan, but the deep violet starting to show up at the same time (top right corner). The camera's sensor is set up in groups of one Red, one Blue, and two Green cells, so I can understand the RGB dominance at normal exposure. For Yellow, Cyan, and Purple to come out like they did, the over-exposure must have been pushing the Red and Blue signals relatively higher in preference to the less sensitive Green.