In this chapter non linear scale factors will be described. The
motivation for using non linear scale factors is to be found in
Weber's law. This law derived in the last century states that the
ratio of the brightness discrimination threshold 
and
the corresponding brightness L, 
is constant over a wide
range of luminances. Logarithmic mapping represents this
law. According to some more recent experiments, exponential mapping
would be more appropriate. These mappings are expensive to compute and
always have some free parameters that should be set for each raw image
separately. Schlick in [Schl94] introduced uniform rational
quantization achieving comparable or better results than exponential
or logarithmic mapping with much less computational effort.