What am I missing in your point?
HLS (hue, lightness, saturation) is one of several color measurement systems; RGB/CYM is another and all are valid representations. Consider HLS in which S (saturation) is the percentage of grey in the image with zero being pure color. It is evident that denser negatives, viz. ones that contain greater amounts of CYM dye components, have greater percentage of grey and therefore less color saturation. Sometimes thinking digitally is worthwhile even in the analog world.
I too normally slightly overexpose color negative films, for example rating Portra 160NC or Fuji 160S at 125. I find highly saturated colors too garish and prefer less of them; in fact, black & white is often the most pleasing representation.
N exposure and N+x exposure for example;
N+x: (dMax - dMin) < N: (dMax - dMin)
N+x has a denser negative, but a thinner image (on the negative), the image density range is maximum minus minimum density, ie: black and white are closer together on the negative in terms of optical density.
Less gradiations on the negative (though still plenty unless it becomes severe)
Displaying min and max density at the same print reflectivity or RGB values requires the contrast to be stretched out, more than the N negative, grain is no longer noise once moving to a print or scan, as it is now part of the signal being used, and is treated the same - the grain contrast is increased too.
The thinner description still isn't making any sense. :confused:
What it seems like you are trying to say is that each stop of extra exposure adds less density than the one prior, essentially the separation of tones gets smaller and smaller as exposure increases.
If true, this would mean less detail is available in the highlight areas. If true it would also seem that Kodak's curves are wrong, i.e. The straight line should actually be curved.
Across a single negative in a shot of say a wedding party the bride's dress could never show as much detail as the groom's tux. The bride's chocolate Lab would have better color separation than the Bridesmaid's light blue dresses.
I did a test a while back. Several 5-shot/stop bracketed sets of 35mm 160nc film.
Contact printed them.
Across a single scanned contact print the color balance remained constant.
I printed several versions varying only enlarger exposure to place one common neutral tone at the same point on paper and it was pretty darn easy to make the variously exposed frames match nicely.
Admittedly this wasn't a rigorous lab study but I was carefull to do things carefully and used PS to measure the resulting print scans.
I found no practical difference in detail or color at the main subject.
Interesting discussion indeed.
Athiril, by saying exposure "thinner" in reference to overexposure, do you mean a thinner range?
As for increasing graininess, I just don't know about that. Larger grains are inherently faster, thus with underexposure you are getting a high percentage of these large grains. It makes sense that overexposure would lead to more of the smaller grains being exposed. Not to mention, this is something you notice in low-light photos, the shadows are often very grainy/noisy. Granted, I've never printed these optically...
Athiril's math indeed is just a description of the characteristic curve (exposure vs. density).
mts' HLS/CMY math is a mistake. This extra complication makes it difficult to see if it is in error or not. But in fact, the actual error in his deduction is that the printing stage, which is simply a subtraction operation from the CMY values, is ignored. When it is added, we end up with exactly the same CMY and HLS values given that the H-D curve is straight. In the end, it is just very simple; the curve shape tells us everything. If we have a linear system, the bias point does not matter; we can make additions and subtractions as long as we are on the linear region. Exactly the same works in analog electronics.
If the curve bows or shoulders, the contrast indeed is lower. Today's films, however, have very long linear portion.
The old wisdom says that increasing exposure reduces contrast and thus perceived saturation. This has been a well-known fact. Somehow it got swapped around and became an internet legend in a completely opposite form.
However, with today's films, in most cases, there should be no difference in contrast and saturation with slight overexposures (+1 to +2 stops) due to shouldering.
I think it depends on what you are shooting. Athiril's speaking about sunrise and sunset, where +15 stops is well possible when metered from the sun and nearby clouds compared to the landscape. This can really land on the shoulder of the film, unless metered like it was a chrome film - from the highlights, then adding a few stops. However, with neg film, you can of course place them higher than with chrome film (e.g. +5 instead of +2), but there still is a limit. If you measure a sunset scenery "from the shadows" as usually instructed for neg film, you go over this limit. "Expose for shadows" is just a rule of thumb for most subjects.
OTOH, even if there is no shouldering, there still is the toe. With subjects with important shadows with colorful objects in the shadows, by overexposing a stop or two, you make sure that the shadows are not on the toe - and increase their saturation and contrast, without altering the saturation and contrast in midtones and highlights - given that the highlights are not difficultly high. This part is true in the internet legend, but the legend is wrong in making this a rule of thumb which it is not. You need an understanding of shadow contrast vs. midtone contrast vs. highlight contrast and understand how this relates to a particular subject.
markbarendt, you are looking at a print, not the negative, there isn't less detail, but less contrast. But on top of lower contrast, minimum opposing density rises, so the colour separation is also smaller.
The best saturation/contrast is around mid tones. If it was such a straight line, you wouldn't need to raise exposure for things below mid tones. And I wouldnt see reduced contrast and saturation above mid tones. There is not much saturation in white to lose.
With a thinner range, grain contrast increases if they are displayed at the same contrast as a broader range negative.
Blue dresses are a more pure (saturated) colour than a chocolate Lab to begin with, and such light dresses would be around or just above mid tones, and a chocolate Lab below it by the sounds. There is no saturation in white to lose, contrast isn't detail, you can have more detail in low contrast object then you can in high contrast, and also vice versa.
"Yes this can be done like the zone system or a push."
You mean a pull.
Perceptual saturation is a different animal. Perceptual saturation is driven by composition, lighting, exposure placement of subjects, the development regime (N, N+, N-), paper choice......
The reason I normally reference the print is because the negative is just an intermediate media. I don't hang my negs on the wall.
I only care about what the negative produces.
I think I see where we are not talking the same language.
Chocolate lab brown and robins egg blue can both be fully saturated colors.
Saturation for me is simply a measure of how close to "normal" or "expected" the color gets.