next up previous contents
Next: Main Goal of the Up: Search for the Optimum Previous: How is it done

Mean Light Method in Photography

In photography light measurement techniques work with the assumption that the scene contrast is 32. The value two times greater than the average measured lightness will be the white on the picture. The logarithmic middle gray, which is very near to the perceptual middle gray is tex2html_wrap_inline5289 times the lightness of white or tex2html_wrap_inline5291 times the measured average value. All values larger than the white value (two times average) will be clipped to white, and all values below white/32=average/16=black will be clipped to black. In other words, the clipping window on the logarithmic scale is logarithmic middle gray tex2html_wrap_inline5293. Let us consider an example from computer graphics with classical mean value linear mapping. We assume a gamma corrected display device with linear response and input range [0,1]. 0 corresponds to the minimum displayable luminance level (in our case let it be 1/32) and 1 corresponds to the maximum displayable luminance value (let it be 1). The values larger than two times the average value will be clipped to 1, as in photography, but there is no clipping for low values. 0 absolute luminance will be displayed as 1/32 in our example. Figure 6.1 illustrates the two approaches. The differences in the final images are not significant because the main difference occurs in dark image parts where our perception in bright surroundings can not perceive them.

  figure907
Figure 6.1: Photographic and CG mean value mapping


next up previous contents
Next: Main Goal of the Up: Search for the Optimum Previous: How is it done

matkovic@cg.tuwien.ac.at