Why film is highly nonlinear ? Fluid dynamics , nonlinear optical characteristics , I cant think other ?
may be nonlinearity comes from layers of disoriented lattice and with its interaction with light .
this is about grain and color layers.
Its scene brightness to projected image brightness characteristic looks linear on a logarithmic scale, but since the slope on this logarithmic scale is larger than one, it turns into a power law on a linear scale. If you, for example look at the characteristic curve of Fuji Velvia 50 (
section 19 on page 8 here), and if you set projected image brightness to -D, you get a straight slope of
((-0.25) - (-3.25)) /
(0 - (-2)) = 1.5, which means linear projected image brightness P and scene brightness S are related through the following power law: P(S) = S
(3/2). As a result, P(S
1+S
2) is not equal to P(S
1) + P(S
2). Weak spectral contributions, or spectral contributions outside the wavelength sensitivity maximum, will yield sub proportional effect. This will show strong differences, if the spectral sensitization of two films is different, too.
Now look at the graph in section 20, also page 8 of this data sheet, here you see the spectral sensitization of each colour layer. This is the part where the whole spectral distribution of incoming light is compressed into a vector of three colour data components. Each film does this differently, and the same thing is done in digital sensors. The graph in this section e.g. shows that Velvia 50 is mostly insensitive to light between 480 and 500 nm, and also between 590 and 620 nm, i.e. light at these wavelength ranges will show up only weakly in your slides. Since real information is thrown away here, this information can not be recreated by a computer, and therefore I think that even computer algos can not recreate the true colour characteristics of Kodachrome regardless of which other source medium you start from.
As you develop colour slide film, Fuji creates intentional inter layer effects in order to boost their film's colour saturation. Development and dye formation in one colour layer will repress development and dye formation in adjacent colour layers, and obviously this will differ between slide film product lines. As I mentioned before, you can't change colour saturation with optical filters.
Now finally, once these silver grains have been developed and image dyes have been formed, and let's assume through a magic wand you overcame all the previous differences, then you will discover differences in output spectrum because the image forming dyes differ between films. The characteristics of Vevia 50's dyes are shown in section 22, again on page 8 of this data sheet. While Velvia 50 uses very pure colours, there is still some density in wavelength regions that were not supposed to have density, especially in the region between 400 and 450 nm. This leads to the following effect: the dye created by its supposed wavelength attenuates not only in its own wavelength range, but in wavelength ranges which "belong" to other colour layers. Both cyan and magenta will mess with yellow's wavelength range, and like interlayer effects, this will show up as a non-uniform change in saturation, not correctable by optical filters.
For a short, clear and less technical answer: see PhotoEngineer's post #4 in this thread.
