Testing quality of photographic lenses

[holsetymoon]

Hi experts,

First of all, I did this topic thanks to Sciencebuddies.org
https://www.sciencebuddies.org/mentorin ... ?from=Home
It is PHOTOGRAPHY. My question for this topic is:
What focal length and f-stop will produce the best resolution?
My hypothesis is: The widest-angle (smallest focal length) lens with minimum f-stop (aperture) will produce the best resolution
Currently, I am thinking about changing the question to: What focal length and f-stop will produce the sharpest image? I know that sharpness consists of both resolution and acutance, but I don't know how to measure acutance with my experiment. My experiment is with the USAF lens test chart, I did it with four different lenses : two 35mm camera lenses and two 35mm camera zoom lenses. I load my camera with Fujicolor Pro 160C color film. I maintain the exposure so that it is not over- or under- exposed in any of my picture by changing the shutter speed. My independent variable is the f-stop (aperture).
My main question to ask you experts is:
How do I measure the line pairs per milimeter from the data I got in my experiment? Hope I made sense. Thanks for helping me.

[deleted-71588]

IMO: You need to do some more investigation before you pose a hypothosis.

Lets start with your choice of films - Color negative - bad choice for lens tests. Evaluating color negatives directly is almost a lost art. All indirect methods involve another lens. Try direct positive film typically referred to as "slide" or "reversal" or in the case of Fuji Films, Fujichrome series.

Now to the lenses and subject distance and flatness of field and contrast and depth of field / circle of confusion and diffraction and other factors besides focal length and f-stop. Read up on them.

I'll give you some hints at some techinical factors that are at odds:
If you use a very small physical aperture (e.g. pin hole camera fstop << f/32), you get more diffraction effects which reduces resolution. If you use a very large physical aperture > f/2.8, you get a very small depth of in focus field and any lack of flatness of field degrades your resolution.

There are lens to lens variations even within the same manufacture and model. Comparing lenses of different focal lengths with different cost factors and manufacturing tolerance quality factors involves a lot of uncontrolled variables. For 35mm film formats, don't expect an $800 24mm lens to outperform a $200 50mm lens because 50mm designs are much simpler and less costly to produce. One of the techincal challenges of wide angle lens design is producing rectilinear designs so that rectangles don't become concave or convex shapes.

If you are attempting to take a picture of a perfectly flat subject perfectly centered in the image circle with the film plane perfectly perpendicular to the subject plane with no relative motion, almost any single focal length macro lenses of any focal length at f/4 through f/8 will beat any other type of lens in this kind of a test. In this particular kind of test setup, the flatness of field of macro lens designs will meet or beat anything else.

[holsetymoon]

IMO: You need to do some more investigation before you pose a hypothosis.

Lets start with your choice of films - Color negative - bad choice for lens tests. Evaluating color negatives directly is almost a lost art. All indirect methods involve another lens. Try direct positive film typically referred to as "slide" or "reversal" or in the case of Fuji Films, Fujichrome series.

Now to the lenses and subject distance and flatness of field and contrast and depth of field / circle of confusion and diffraction and other factors besides focal length and f-stop. Read up on them.

I'll give you some hints at some techinical factors that are at odds:
If you use a very small physical aperture (e.g. pin hole camera fstop << f/32), you get more diffraction effects which reduces resolution. If you use a very large physical aperture > f/2.8, you get a very small depth of in focus field and any lack of flatness of field degrades your resolution.

There are lens to lens variations even within the same manufacture and model. Comparing lenses of different focal lengths with different cost factors and manufacturing tolerance quality factors involves a lot of uncontrolled variables. For 35mm film formats, don't expect an $800 24mm lens to outperform a $200 50mm lens because 50mm designs are much simpler and less costly to produce. One of the techincal challenges of wide angle lens design is producing rectilinear designs so that rectangles don't become concave or convex shapes.

If you are attempting to take a picture of a perfectly flat subject perfectly centered in the image circle with the film plane perfectly perpendicular to the subject plane with no relative motion, almost any single focal length macro lenses of any focal length at f/4 through f/8 will beat any other type of lens in this kind of a test. In this particular kind of test setup, the flatness of field of macro lens designs will meet or beat anything else.

I have a few questions for you. You said that using color film is not good. So why is black-and-white film better than color film? About the subject distance, I have considered it already. The longer the focal length, the longer the distance from the lens to the target according to this formula:
- For lenses up to 135mm focal length, distance = focal length * 50 (milimeters) - For lenses above 135mm focal length, distance = focal length * 25 (milimeter). I have reduced the number of controlled variables to the minimum. You said that "almost any single focal length macro lenses of any focal length at f/4 through f/8 will beat any other type of lens in this kind of a test". Why is it so? Thanks for your help.

[deleted-71588]

I didn’t say “color film is not good.� What I said was “Color negative� film was a bad choice. I actually recommended Fujichrome which is a color positive (aka slide, aka reversal) film.

If you were intending to get prints made from your negatives and evaluate the prints, then that is a second generation photographic process. To print a negative, you use an enlarging lens, perform another exposure onto a photographic paper and develop the image. This means that you are evaluating a second generation which is bad from a scientific standpoint. Film curl in the enlarger, quality of the enlarging lens, flatness of the photographic paper, and lots more second generation variables....

As far as why “almost any single focal length macro lens... will beat any other type of lens in this kind of a test", I guess you didn’t understand what I said next “the flatness of field of macro lens designs will meet or beat anything else�.

Macro lenses are specifically designed to have a very flat in focus “field�. I can’t use the word “plane� here because that has a geometric mathematical meaning of perfectly flat. Even the best lenses with the best manufacturing tolerances have some slight “warp� to the in focus “field�. Getting this very flat in focus field is a design trade off that comes at the expense of other factors (complexity, weight, cost, etc).

Non-macro lenses trade off the flatness of the in focus field for other design factors.

In other words, macro lenses are specifically design to perform very well under these specific test conditions. The market for macro lenses is much smaller than for other lenses and those who buy these want a high quality image so there aren’t any low quality low cost versions.

Given that:
1) You have to use the same test chart to avoid introducing another variable
2) You have to reproduce the test chart in the full 35mm image frame to avoid other variables
3) Different focal length lenses have different reproduction ratios
You aren’t going to be able to control subject distance. You will have to pick the subject distance that precisely fills the 35mm image frame for each focal length lens you test.

You need to consider camera motion and shake and the effect of the reproduction ratio on these. A ½ degree motion of the center axis of the lens will produce a larger image motion blur the longer the subject distance. In other words, a setup that is stable enough for a shorter focal length lens may not be stable enough for a longer focal length lens. You don’t want your results skewed by this.

As far as “How do I measure the line pairs per milimeter from the data I got in my experiment?�, you don’t. You observe and evaluate the image (using a magnifying glass or loupe) using the USAF test methodology criteria. You need to read up on it. I’m sure you can find some digital images of this if you do some searching. The USAF chart has groupings of different line pairs per millimeter in different areas of the chart.