Optics of macro: do you gain anything with LF ?

[Sorin Varzaru]

I asked the same question on photo.net then I figured it belongs here.

I recently got my first 8x10 and I was thinking about trying some high magnifica tion macro work with it. And then I read the optics article on photo.net and I m ade some calculations and ended up with a surprising result. It seems that when shooting 3d objects (that require some dof) there is virtually no reason to shoo t larger format since it yelds pretty much the same result.

First, here is the assumption : We are planning to photograph an object that is 24x24 mm in size and we want it to fill the frame. So with a 35mm camera we woul d have 1:1 magnification. With an 8x10 camera we would need ~ 8:1 (I would use a 65mm lens so I'd have enough bellows). The required DOF for this application is 2mm.

Using the formulas in the optics page a got the following numbers : 1. The optim um aperture for 35mm work came to f19 (optimum meaning the smallest circle of co nfusion due to dof and diffraction). For this f number the circle of confusion i s 0.026. So the number of "pixels" of the image would be 24/.026 = 923

2. For the 8x10 camera the optimum aperture is ~ f64. When the circle of confusi on is 0.093 So the number of "pixels" of the image would be 2.56 * 8/.093 = 2202 .

Important, the purpose of the exercise was to see if shooting 8x10 instead of 35 mm in the above conditions would dramatically increase the detail captured on fi lm (like I thought considering the huge difference in the film size). It looks l ike shooting 8x10 only about doubles the detail that a 35mm can capture. I am aw are of the advantage of eliminating another optic system by making contact print s and presumably the wider tonal range the LF can provide (I'm not yet conviced about that though).

So, the question is, is my conclusion anywhere near the results obtained in prac tice? Again, I am ONLY talking about macro work where dof is required, not flat work.

[Steve Baggett]

There is one more consideration that gives some weight to using LF. If you plan to "reproduce" the image other than by digital scanning, i.e., by traditional enlargement, there is always some loss of "quality" at each step. The best way to preserve the most "quality" is to get the image as large as possible on the film before beginning any reproduction.

[Ellis Vener]

If this was only a mathmatical exercise than my take on it is that this sort of exercise really doesn't have much to do with actual photography. Important factors are left out: the resolution of the films used in both formats, film flatness, real world resolving power of a specific lens at this degree of magnification, subject vibration, camera vibration, both relative to the other, enlargement (and subsequent optical or scanning techniques and materials) of the image made on 35mm to the same size as the image made on 8x10, etc.

In short, math only takes you so far in photography, after that seeing is believing, or as they say in cooking classes: "the proof is in the pudding." (Editors at "Gourmet" also say that about food photography but only if something goes horribly wrong!)

[John O'Connell]

I don't understand your "pixel" theory. You're comparing the data in a 24mm image with the data in a 203mm image. Note the 8x difference in size here. Your circle of confusion values are based on less than a 4x difference. So you're willing to accept a 35mm original which is half as sharp as an 8x10 once you enlarge it.

And you can't claim to be relying on circles of confusion as your criteria for evaluating sharpness and then magically create some new unit called "pixels" to evaluate sharpness. Your "pixel," created by calculation, is expressed in units of measurement. As far as I can tell, "pixels" are mm^2. Since you seem to be discussing a linear measurement, I'm not sure what you prove with this.

BTW Stroebel, on p. 127 of VCT, says: "A diameter of 1/100-inch is sometimes considered to be the largest circle that will appear as a point in a print viewed at a distance on 10 inches. If the print is a contact print, the same criterion applied to the negative, but if the print represents a two-times enlargement from a smaller nagative, the acceptable circle of confusion in the negative is one-half as large as it is in the print, or 1/200-inch. The degree of enlargement of the negative has no effect on the depth of field provided the image is not cropped."

[Sorin Varzaru]

>If this was only a mathmatical exercise than my take on it is that this sort of exercise really doesn't have much to do with actual photography.

I was looking for the best way to photograph very small 3d objects.

>Important factors are left out: *the resolution of the films used in both formats *real world resolving power of a specific lens at this degree of magnification

Since the maximum resolution that can be achived is ~ 39 lp/mm I considered that film resolutuon and lens resolution not to be a factor.

*film flatness

Since is more likely this is an issue with 8x10 rather then 35mm is one more reson to stick with smaller formats.

*subject vibration, camera vibration

This was a comparison between two formats. It was assumed that the shootin gconditions were the same.

*enlargement (and subsequent optical or scanning techniques and materials) of the image made on 35mm to the same size as the image made on 8x10

This would be the only reason I would consider using the 8x10 for this kind of work.

To conclude, since I'm fascinated of the posibilities offerd by macro photography (it's amazing how a snowflake or a tiny bug look like under 8-30x magnification) I was looking for a better way to capture this images. Before I made those calculations, I was under the impression that the detail captured by a camera is roughly proportional with the film size (assuming the shooting conditions and optics are similar). And that is true for all the situations where dof is not critical (like focused at infinity or flat work). But it looks like for this specific application, shooting 8x10 or any larger format is not really justified. The result was so surprising I wanted to make sure I didn't miss anything. Hence the question.

[Pete Andrews]

John: I'm curious about your quote from Stroebel (who hell him?).There's a distinct contradiction in saying that the CoC is magnified by any enlargement, and yet simultaneously maintaining that DoF is not affected by enlargement scale. Since all calculations for DoF have to be based on an acceptable CoC, it follows that these two statements can't both be true. It's also very clear, from simply comparing a large print with a small one, that the apparent DoF is affected by scale.In real life, viewers don't automatically view a print from the mathematically 'correct' distance.

[Steve Clark]

Pete, I don`t know what he`s doing these days, but I had Dr. Stroebel as a prof. at RIT in NY about twenty years ago. He taught LF and other things...

[John O'Connell]

Leslie Stroebel is the author of "View Camera Technique," a book I've found handy using LF.

DOF is a value that's only valid for a certain minimum viewing distance. If someone is going to get close enough to a print, only the plane of exact focus is going to look sharp. DOF tables assume that as the image is enlarged, the typical viewing distance increases as well.

Yes, DOF and COC are related, but they aren't the same thing. COC simply refers to the smallest circle which can be acceptably resolved as a point; DOF refers to the distance around the plane of exact focus at which all subject details will be resolved as circles equal to or less than the COC.

Personally I agree with you, Pete, and these formulae are pretty useless. People get their noses all over prints sometimes. I focus on the nearest important part (or most important part if there is one) of my subjects. DOF is useful mainly in telling you where you have to hang a print to keep people from seeing it fall apart, not telling you how sharp something is.

[Chris Partti]

Sorin,

Instead of playing with equations that may or may not be pertinent, why not take some photographs using both systems and compare the results? Please let us know what you find.

[QT Luong]

Sorin: when you want to people to comment on your math, please quote the formulas that you are using, instead of numerical values. This way we can check your reasoning instead of guessing what you intended to compute and how.