These photos were taken with a Digital Camera, that has an auto-exposure function with the same goal as the human vision system: to produce images that are neither too bright nor too dark. In mathematical terms the average of all the image points (or pixels) RGB values must be in the middle of their range. This algorithm is a decent approximation of the human eye-adaptation function. But, in a case like this, where half of the scene is really bright and half really dark, neither half will get the right exposure. As a matter of fact the calculated exposure, mathematically correct for the all image, is too long for the right side (or over-exposed = too bright) and too short for the left (or under-exposed = too dark).
Photos 2 and 3 show what happens as we block (in two steps) the view of the window (the light source) from the camera:
- the luminosity of the scene (as seen by the camera) decreases, and, as a consequence,
- the exposure increases to keep the total brightness of the resulting image about the same. And this benefits the left side which gets brighter and brighter,
This is another example of the MFV-1 effect used in MFV TM eye wear and not noticeable with common sunglasses.