[Leica] RMS granularity

[Erwin Puts <imxputs@knoware.nl>]

The new Provia 100F has a published RMS granularity value of 9 and 
is, according to Fuji the lowest value of any 100 speed slide film. I 
do not wish to dispute this claim. Film manufacturers are 
conservative in their claims. But what does it mean? It is being 
cited as a premium characteristic, that distinguishes this film.
First someRMS  figures.
Provia (old); 	10
Velvia		9
Astia 		10
Provia F (new)	9
Fujicolor 100	4
Fujicolor 400	5
Fujicolor 800	5


So we see that even an 800ISO color neg film is much finer grained 
than a very fine grained slide film.
The RMS granularity is a figure derived as follows.
A microdensitometer with a circular aperture of 48 micron scans the 
uniformly exposed area of an emulsion. Because of grain distribution 
the density of the emulsion is not equal (as it should be because of 
uniform exposure). The density thus fluctuates and these fluctuations 
indicate the presence of grain. The variability of the fluctuations 
around the mean is well represented by the deviation which is the RMS 
value. Remark that it is a fluctuation around the mean and because of 
the reading aperture of 48 micron IS ONLY VALID at a magnification of 
12 times. At any other magnification (usual for slide film) the 
correlation between graininess and RMS value may not hold.
The other point of this measure is worth noticing: the viewer of the 
grain pattern looks at such a distance that the apparent grain 
pattern seems to  be not visible or blends. We may wish to ask 
whether this threshold value  is of overriding importance. If the 
viewer still sees a slight grain pattern (disregarding here all kinds 
of psychological and physiological factors), would he/she detect a 
significantly larger grain? And if so what is more important: light 
scatter because of finer grain which will reduce the impression of 
sharpness quite a lot or the fine grain impression. Furthermore: the 
chracteristics of B&~W grain and of dye couplers are very different. 
Because of the uniformly spread dye couplers  the random fluctutaion 
around the mean is reduced. So the value of 5 or 9 for dye coupler 
films does not indicate the same grain impression a sa value of 11 
(Kchrome 25) or a 13 for B&W negative film.
As so often: any single value taken out of context and without 
reference to the measurement parameters may be quite misleading.

Erwin