Physical coding v. Perceptual coding

We optimize the design of video systems and digital photography systems for visual perception. We place the intensities of the codes - say from 0 to 255 - where they will do the most good perceptually. Human vision responds to intensity in a logarithmic manner, so we space the intensities according to a fixed ratio, the Weber fraction, of 1.01 or 1.02. We do not space the notes according to an absolute intensity increment, as Timo suggests is optimum.

The assignment of code values to intensities in digital imagery is comparable to the way we optimize a piano for aural perception - for hearing. The keys of a piano have fundamental frequencies related by ratios: each semitone has a frequency that lies the twelfth root of two - about 1.06 - above the next lower semitone. (There are twelve semitones in an octave. If you raise 1.06 to the twelfth power, you get 2.) We place the frequencies of the notes of a piano where they will do the most good perceptually. Human hearing responds to frequencies in a logarithmic manner, so we space the notes according to a fixed ratio.

Lest you think I am making up this stuff about pianos, read a definitive source on the topic of Relative tunings.