View Full Version : Lenses and tone reproduction

Michel Hardy-Vallée
03-04-2011, 03:22 PM
I use the term "tone reproduction" in the Kodak sense, hope that it will forestall any hairy misconstructs of language such as "tonality" or "creaminess" ...

I remember reading in the Kodak Encyclopedia of Practical Photography the article on Tone Reproduction. It was an explanation of the quadrant approach for understanding how tones evolve over the course of photographic reproduction between the original and the final print.

One of its major conclusion was that "perfect" tone reproduction is not satisfactory to the viewers. In other words, two things in a scene that reflect different amounts of light might look better on paper if they are in fact of the same shade of gray. That's what we call "compressing" tone. The opposite is true: two things in a scene may be very close in the amount of light they reflect, but on a print, this difference needs to be exaggerated to be satisfactory.

Eventually, the Kodak folks figured out which ranges of tones, on average, needed compression, and which ones needed expansion.

So far, if you have ever played with the Zone system, adapted your development for a specific scene, if you know which film/paper combination works better for a given subject, or just changed paper grades to fit a negative, you have a sense of what I'm talking about.

But I'm curious about the role of lenses in this process. We know that flare matters, that it can lower overall contrast, but do we know something about the actual tonal reproduction curve of our lenses?

Reading on painting, today, it made me think how much a lens was important in imaging tones on a negative. It is, after all, a brush of some sorts, light or heavy.

In other words, if your lens was the equivalent of a Photoshop curve, what would it look like? Would it compress or expand highlights? Would it have an upswept, straight, S-shaped, or other curve? How can we reasonably measure or witness it without adding too many variables to control?

Michel Hardy-Vallée
04-18-2011, 09:37 AM
Bump, anybody?

04-18-2011, 09:48 AM
I am not an expert on lens design but my first thought is to wonder how a lens would know which range of brightness was to be which range of tones for me. The brightness of light is very relative to I guess a comparative perception, a phrase I just made up. The lens would have to have a computer chip in it to calculate brightness range and contrasts. As far I my non expert knowledge goes I understand that the lens designs are balancing contrast and resolution. The highest resolution lens is too low contrast. A contrasty lens appears sharp but actually doesn't have as much resolution. My other non expert idea is that the European lens designs like Zeiss Schneider and Leitz go for a bit higher resolution and a bit less contrast compared to the Japanese designs like Mamiya and Pentax are contrasty. If you want to tell me I am full of barn yard manure, I won't argue.

04-18-2011, 11:50 AM
Flare compresses the shadows. It usually has little effect on highlights because the ratio of flare (which is a uniform, non image forming light striking the negative) to highlight is so small. The ratio of flare intensity to shadow is pretty large, in fact it could be larger than unity in the sub zone I area. So most of the effect is in the shadows.

There was an extensive thread on the effects of flare on exposure index this recently. The thread is difficult to read but summary as follows;
View 1: Flare causes pre-exposure in the shadows and can augment film sensitivity (speed) and makes the system faster.
View 2: Flare compresses shadow tones, thus decreasing separation. Film speed is determined by separation of tones (Speed point is 0.3 x G which is an instantaneous gradient) and thus flare makes the system slower.
My Conclusion: One can choose a film-speed definition to satisfy View 1 or View2.

Here is a typical tone reproduction curve, showing a typical flare curve.

(Image from an S. Beskin post)


Michel Hardy-Vallée
04-18-2011, 01:39 PM
Ah, that's the kind of diagram I was looking for.

The only thing that seems to impact tonality on a lens is flare.

Is it possible that other things affect the tone curve of the lens?

04-18-2011, 02:01 PM

Post Steven Beskin directly. He is the absolute expert on this subject. Search for his posts, or email him!

04-18-2011, 04:18 PM
For the strength of electromagnetic waves we discuss here we are in the region of linear optics. Linear means if i double the input the output doubles as well. This means the transfer curve of a lens is a perfect straight line. The glass in the lens does not saturate and limit the light nor does it add light.
Flare makes this complicated. A gray card will have a certain tone in a negative. This tone will change when flare comes into play. But the change comes not from the object itself (the gray card), it comes from another light source on a different position. The transfer curve becomes more complex and the output for every single point on the negative (i.e. its gray value) depends on all points in the object plane. Nevertheless this is still a linear function.