By default, Presenter uses lights with unitless, or empirical intensities. These are physically meaningless, and are just chosen to give a visually pleasing result. Presenter can also use physically accurate intensities. These are defined in real-world units such as Candela, Lumen, or Lux. However, once you start using lights with real-world intensities, you start to produce images with a real-world variation in luminance values.
By default, Presenter uses lights whose intensity remains constant as you move further from the light. In the real-world intensity is reduced proportional to the inverse square of the distance from the light. Changing a light’s Fall Off parameter to Inverse Square Law will more accurately model a light’s fall off in intensity. However, once you start using lights with real-world fall off, you start to produce images with a real-world variation in luminance values.
In the real world, the human eye is capable of resolving images in extremely varied lighting conditions, ranging from bright sunshine reflecting off snow to a room lit only by a single candle. In computer graphics, however, you need to produce an image on a display device which has a very limited range of luminance values. Therefore, it is necessary to compress the range of luminance values found in a real-world scene into the displayable range in such a way as to produce a realistic looking image.
Photography, of course, has exactly the same problem. If a photographer (or camera) does not take into account the light levels in a scene before calculating the exposure of the shot, the likely result will be an image which is either over-exposed (everything is too bright) or under-exposed (everything is too dark). A professional photographer will also use different speeds of film for different lighting conditions. The aim is to produce an image on film that is representative of how that scene would have looked to a human observer.
Presenter includes the Auto Exposure option (see Auto Exposure). When enabled, Presenter will render the image twice. Once to sample the range of luminance values in the output image, then a second time to render the actual image with the luminance values adjusted to match the behavior of the human eye.
In general, when using physically accurate lights, the Auto Exposure option should be on.