DOF with 8x10....

[cluster]

Ok...here is a newbie question to all the experienced LF shooters

out there: If I am shooting a straight on "Head and Torso" portrait

of someone with a 360mm f6.7 lens on 8x10, how can I determine the

minimum f-stop that is going to render the subject sharp front to

back? As a guide, I figure I need at least a 24" section to be

sharp between the subject's front belly and their backside and I

don't want to resort to using f64 just to be on the safe side. I

vaguely remember the term "circle of confusion" which I once heard

about in the past but have no clue as to what it means. Does anyone

out there know of a simple formula that I can use? Many thanks.

[lbi115l]

I would have replied about 15 minutes ago, but I couldn't find the book.

 

This is straight from View Camera Technique by Leslie D. Stroebel. It's THE must-have book for any LF photog, and either it or some photocopied pages stay with my camera at all times.

 

From page 158:

 

First, you have to calculate the hyperfocal distance:

H= f^2 / (N*C)

where H is the hyperfocal distance, f is the focal length, N is the F-Stop number, and C is the acceptable circle of confusion.

 

To find the near distance in focus (Dn):

Dn = (H*U)/(H+U)

where U is the object distance (distance focused on).

 

To find the far distance in focus (Df):

Df = (H*U)/(H-U)

 

And finally, to find the Depth of Field (DOF):

DOF = Df - Dn

 

A few notes. First, make sure all measurements are in the same units. This applies to focal length too. It's a smiple point that can be easily forgotten if you're in a rush.

 

On www.lfphoto.info under "Misc" there is a page with links to photographic programs. I have Bob Wheeler's Vade Mecum on a TI-89, which is a great program to do many photographic calculations. I highly recommend it, it can amso be had for Windows CE, Palm, and some other OS's. The only problem is remembering what all of the variables are. Storing a text file on the device will help this.

 

There is also a program out there called F-Calc which does this type of stuff.

 

If anyone wants something that these programs won't do, drop me an email and I'll write one for you. I've been working on one off and on for about a year, though I always get caught up in adding more options.

[tito sobrinho]

Joseph: Four things.

 

1-Dark cloth.

 

2-Ground glass.

 

3-Loupe.

 

4- Your fingers to close and or open the diaphragm.

[tedharris]

Not to mention that all the formulas and rules in the world are no substitute for your

own style and imagination .... use Polaroids and check that the image is what you

want.

[lbi115l]

Yeah, checking it out in person is the easiest way. But the formulas can also come in handy, especially when you have a certain shot in mind and are renting a lens. It's a lot easier to see if it'll do what you want before you have it mounted on the camera and paid for.

[bruce_barlow]

Go to the Schneider web site and print off the depth of field chart for a 360mm lens. You'll be a little frustrated because they don't have all the apertures, but it will be a good start and a valuable resource.

[leonard_evens]

As others have indicated, it depends on a variety of factors, and the answer can be expressed in different ways, depending on the application.

 

Since the 8 x 10 format is so large, it is probably wise to use the exact formulas using the magnification M, which can be found in Jacobson's lens tutorial. The simpler approximate formulas referred to in a previous response apply for objects sufficiently far away, usually interpreted to mean at least 10 times the focal length. If you use a 360 mm lens, this would be 3.6 meters, which is a bit far away for a head and shoulder's portrait. I estimate that a head and shoulder's shot on an 8 x 10 negative would involve a magnification of M of between 1:2 and 1:3 (with the subject between about 0.72 and 1.44 meters from the lens). Let's take M = 2.5. From Jacobson, and some algebra, we get the following formula for the TOTAL DOF from front to back

 

2(Nc)(1+M)/[M^2 (1 - (Nc/f)^2)]

 

where N is the f-number, c is the acceptable circle of confusion, M is the magnification, and f is the focal length.

 

A common value used for 8 x 10 is c = 0.2 mm, but many people would go lower, possibly as low as 0.1. The term (Nc/f)^2 can probably be ignored because N will be smaller than 64 for sure, and

(0.2 x 64/360)^2 ~ .0013. So 1 - ((Nc/f)^2) is very close to 1, and dividing by it has virtually no effect.

 

That yields the slightly simpler formula for the total depth of field

 

2 (Nc)/(1 + M)/M^2

 

If you take M = 1/2.5 = .4 and c = 0.2 mm, this becomes

 

2 (N x 0.2)(1.4)/(0.4)^2 = 3.5 x N

 

Front to back as you describe it will probably be at least 200 mm, which means N will have to be at least about 200/3.5 ~ 57, and that is being generous. In other words, you will probably have to stop down to f/64, if you can do it at all.

 

If you actually try this, the image will probably be so dim at small apertures that you won't be able to see how much is in focus. But you should be able to estimate visually how much is in focus, say at f/8. The formula tells you that the total distance in focus is directly proportional to the f-number. So if for example you find at a certain distance from the camera, you have about 2 inch in focus at f/8 (optimistic at any plausible distance), then to get 10 inches in focus you have to multiply the f-number by the ratio of the two, 10 in/2 in = 5 and you would need f/40.

 

Look again at the formula

 

2 Nc(1+M)/M^2.

 

M tends to be quite large for an 8 x 10 photograph with the subject at normal portrait distance. For a smaller format, M (and hence even more its square) would be smaller, so dividing by it yields a larger number. That means that if you want a lot of depth of field in portraiture, you are better off using smaller formats.

[michael_briggs2]

Leonard has given a good estimate of the aperture required for the situation described in Joesph's question. My question to Joseph: are you sure that you need that much depth of field? Why do you want the back of the portrait subject in focus? Leonard's answer is the background that explains why the typical head-and-torso portrait made with a LF camera usually has shallow depth-of-field. The traditional LF portrait approach is to focus on the near eye of the subject and stop down from there, perhaps getting the far eye in focus. In this approach one accepts that farther objects, such as hair or shirt collar far from the plane of best focus will be out of focus. Besides technical factors of shallow dof leading to this approach, the shallow dof can serve to emphasis the important portions of the face that the photographer selected to be in focus. I am not suggesting that anyone should slavishly follow any "rules", but an understanding of traditional approaches is a good starting point.

[allen_whittier]

Joseph,

 

First, Get yourself a Rodenstock depth of field calculator. They only cost $15 and will give you exact f stop to work with. It's a little revolving slide ruler that fits in your pocket, on batteries needed. Until then use this guide for 8x10;

 

Focus on the near point you want in focus make a mark on a piece of tape or post-it that corresponds to some moving part on the focusing rail. Then focus on the far focus point and mark it. Note how far you moved between the two points by measuring between the marks. Take that movement distance and apply it to the following.

 

For a 1/1500 the film diagonal circle of confusion (COF);

 

Movement = f stop;

2mm = f5.6,

3mm = f8,

4mm = f11,

6mm = f16,

8mm = f22,

12mm = f32,

17mm = f45,

24mm = f64,

34mm = f90.

 

Once you have your f stop determined, without looking at the ground glass, move your focus to the point half way between the "near" and "far" marks you just made on your movement marker. You're now at hyper focus with the near and far points in focus to the indicated COF.

 

If I have the luxury, I always move to the next f stop to make sure everything's in focus.

 

To change the diameter of the COF; change the numbers in the Movement column the same percentage you change the COF diameter. Hence cutting the Movement numbers in half will cut the COF in half

 

Don't forget to recalculate apertures based on your bellows extension. The adjusted aperture numbers are the numbers you need to be using.

 

This works with any and all lens used on an 8x10. For 5x7 move the Movement numbers up one notch. 3mm = f5.6 4mm = f8...... For 4x5 move them up two notches. 4mm = f5.6 6mm = f8....

 

Happy shooting.

[cluster]

Thank you all so much for the help. I will definitely use the formula given above and experiment with it myself. As far as selective focus on the subject is concerned, there are times I want to see the entire subject area in sharp focus. It's just an aesthetic thing for me...I just want to see sharp skin tones from the tip of the nose to the back of the neck. One of the reasons I moved up to 8x10 was to get brutally sharp detail in my portraits right down to counting the pores on the skin. I know this may sound gruesome and some of my fellow LF shooters out there may cringe at this idea, but I personally find that that level of clarity and sharpness pleasing. BTW: I once saw a photography show at PS1 in NYC in 97 which depicted huge (I mean huge!), mural/wall sized black and white close up nude photographs (self portraits) of a very old man (probably in his late 70�s) shot against a white wall. I forgot the name of the photographer but the images where so brutally sharp and clear it was as if I were really experiencing him right in front of me. His skin, old and wrinkled like leather was just hanging off his bones. Maybe too close for comfort for some people but for me it was awesome. Thanks everyone again.

 

Joseph

[leonard_evens]

If you are going to use the formulas I gave, you have to know how to estimate the magnification. You can of course do this by measuring some absolute distance in the subject and the corresponding distance on the gg and dividing the latter by the former. A simpler method is to find the distance of the subject to the lens, divide that by the focal length, subtract one, and take the reciprocal. Both subject distance and focal length should be in the same units. If you want to convert inches to mm, multiply by 25.4. So, if the subject is 5 ft = 60 in from the lens, it is 1524 mm from the lens. If the focal length is 360 mm, the magnification is

 

1/(1524/360 - 1) ~ 1/3.233 ~ 0.3093

 

With this value for M, the total depth of field would be about

 

5.48 x N.

 

If the aperture is between f/32 and f/45, you should get 200 mm depth of field, if a coc of 0.2 mm is acceptable to you. If you are more demanding and want a coc of 0.1 mm, you would have to double the f-number.

 

I wouldn't describe the resulting image as 'head and shoulders" since an average head would only take up about a third of the 8 x 10 negative, but it may be what you want. It would certainly give you better perspective than the subject distances I described previously.

 

It might also be added that to get all that depth of field, you would have to focus roughly halfway between the near and far limits of dof. That means that the face, the nose in particular, would be right at the limits of the dof and any slight departure from ideal optics would tend to put that part of the image slightly out of focus. In a portrait, the face should be well within the region of adequate focus rather than at the far limit, preferably very close to the plane of exact focus.

 

The full figure you describe was probably at some distance from the lens, and its focal length may have been less that 360 mm. That would tend to increase the depth of field and might very well result in an image like what you describe. But a 'head and shoulders' portrait is another matter entirely.

 

Finally, it is possible to produce extremely sharp images with medium format equipment when the image is at portrait distances.

[howard b. schwartz]

this is from the large format technical page:

 

 

"To summarize, if D is the focus spread expressed in millimeters, then the optimal f-stop which yields the sharpest possible image at the depth of field limits is N = sqrt(375 D). This works regardless of focal lengths, formats, and movements. The resulting resolution at the limits of depth of field (ie for your far and near points) cannot be improved in anyway and determine the maximum possible enlargment."

[ralph_barker]

Aside from the theoretical objectives, and the pre-calculated assumptions, Joseph, don't forget the practical aspect of the amount of light needed to shoot at small apertures. You'll need a pretty hefty set of stobes to get beyond the moderate apertures at which most portraits are done. Additionally, the more light you pump into the subject's face, the less comfortable the session becomes for them.

[leonard_evens]

Dr. Schwartz raises a very important point, which we have not addressed here. At very small apertures, diffraction becomes an issue, though less so for an 8 x 10 image which is not enlarged too much than for a smaller format image which has to be enlarged more. The formula he gives is that of Paul Hansma and it tries to balance diffraction against defocus, which is what the previous formulas are concerned with. The total focus spread is given approximately by 2Nc(1 + M). That formula ignores diffraction, but when viewing the 8 x 10 image on the gg, diffraction will probably not be that much of a problem in determining the near and far point, so the value N = 45 is probably realistic under the given circumstances. That results in a focus spread of about 23 mm. Thus Hansma's formula would suggest choosing something like f/90 as the aperture. That would allow enlarging the negative something like 12-14 times and then viewing it at close quarters. I'm not sure how important that is in practice.

[cluster]

What is this elusive �diffraction� affect that everyone speaks of? At what effective f-stop does diffraction begin to deteriorate sharpness? There are lenses out there (MF to LF lenses) that have minimum apertures of f16 to f90. Why does closing down a lens with a min aperture of f16 all the way (MF: example Hassey 80mm CF lens) create diffraction where as closing down a 360mm Fujinon to f16 doesn�t? Isn�t f16 in both cases the same diameter? I don�t understand this inconsistency�

[edward_kimball]

I cannot strongly enough recommend downloadng the program fcalc off the net. If you provide subject distance, focal length, f-stop and film format it will calculate depth of field and near/far in focus points. If you are in the studio the PC version would be adequate but in the field I belive there is a version for use on PDAs.

[michael_briggs2]

<p><i>What is this elusive �diffraction� affect that everyone speaks of?</i> A simplified explanation: Because of the wave-nature of light, light bends as it passes the aperture of a lens. The smearing from this bending limits the resolution of the lens. A smaller diameter aperture causes greater diffraction and lesser resolution.</p>

 

<p><i>Why does closing down a lens with a min aperture of f16 all the way (MF: example Hassey 80mm CF lens) create diffraction where as closing down a 360mm Fujinon to f16 doesn�t?</i> Diffraction is created in both cases. Imagining perfect lenses (aka diffraction limited), the resolution on the film will be the same for both the MF and LF lens used at f16. Assuming the same size print (which is what the viewer cares about!), the resolution of the print from the MF lens will be worse because the MF film must be enlarged more.</p>

 

<p><i>Isn�t f16 in both cases the same diameter?</i> F-number is focal length divided by diameter, so f16 means a different diameter for lenses of different focal lengths. Again, for resolution on the film from a diffraction limited lens, f-number is the parameter that directly gives the resolution.<p>

 

<p>You might want to study David Jacobson's <a href="http://www.photo.net/learn/optics/lensTutorial">Lens Tutorial</a>, available on photo.net. An excellent book that will teach you some optics is <b>A History of the Photographic Lens</b> by Rudolf Kingslake. The book is more readable (fewer equations) than the Lens Tutorial.</p>

[leonard_evens]

Joseph,

 

The effect of diffraction depends only on the aperture (f-number) and the wavelength. Usually, one middle wavelength is chosen to specify it. Since it is independent of focal length, it does depend on the degree to which the final image is enlarged. If you enlarge more, you need to compensate by using a lower f-number. Since smaller formats have to be enlarged more for the same size final image, usually those formats are restricted to larger apertures. Of course, it all depends on the circumstances, including what you expect to do with the final image. But usually, f/22 is considered the smallest acceptable aperture for 35 mm lenses, f/45 for medium format lenses, f/64 for 4 x 5 lenses and f/128 for 8 x 10 lenses.

 

I think for most large format photography, diffraction is usually a minor issue. The usual DOF formulas and calculators ignore diffraction and give reasonably accurate answers. However, if you expect to be making large prints which 'grain snifters' will be examining with their noses right up against the print, then you have to consider diffraction. See www.largeformatphotography.info for further discussion and for a description of Paul Hansma's method of combining defocus and diffraction.

[cluster]

How then does �Diffraction� compare to the so called �Sweet Spot� (sharpest aperture setting on a lens--some say f5.6 or f8)? Why does this middle �Sweet Spot� aperture result in the best sharpness? Further, why are there examples out there (such as: Hasselblad 100mm CF) that are tack sharp at all f-stops? Also, does multicoating have any effect on diffraction? This is probably beyond the scope of this forum but it does beg the question?