"Depth of Field" & Focus Problems withLF. QUESTION

[ardea]

I'm a newbie to LF so please bear with me.

 

I have noticed lately and in the archives, problems some people are

having with unsharp negatives when the image appeared sharp on the

GG. It seems that most folks feel that the problem can be traced to a

difference in film plane to ground glass plane. Theres a whole lot of

measuring going on.

 

Out of curiosity I measured my new Shen-Hoa 4X5 GG distance,Toyo &

Fidelity film holders and found maximum deviation of three and a

half thousandths (.0035") or .09mm.

 

Now the question/s. Assuming a circle of confusion (CoF) of two and

a half thousandths (.0025") or .06 mm. A 150 mm lens focused at inf.

and stopped down to f32. How far forward and behind the film plane

will the above CoC lie? I just wonder if using the above values does

the CoC lie beyond most mfg. tolerances and slight film bowing.

 

Thanks..Richard Martel, Florida Keys

[george_jiri_loun]

Richard, assuming you take your pictures more or less at infinity (i.e. up to small magnification) the depth of focus t= 2CN which gives for C =0.06mm and N = 32 the "nice" value of 3.84 mm. All other depends on where you have placed the ground glass in this depth...

[ardea]

Hi George

 

Sorry to be so dense but I don't understand how the formula was derived. The number "3.84 mm" is well over 1/8 of an inch and most likely well beyond the most gross mfg. error in filmholders or GG location. I believe the number 32 representing f 32 is just the ratio of the focal length to aperture diameter or A= FL/fstop #. Where 150/32=4.69=A for an aperture of 4.69mm. I'm not that versed in mathamatics but I would have thought that one would have to use actual values of A to determine how far the CoC lies, forward or aft of the rear focal point. If my assumptions are incorrect please don't hesitate to enlighten me.

 

Thanks, Richard

[brian_ellis3]

I don't understand what you mean when you ask how far a circle of confusion of a given size lies forward and behind the "film plane." I understand the concept of a circle of confusion of a given size being a certain distance in front of or behind the plane of focus but I have trouble relating it to being formward or behind the film plane. Did you mean to say the plane of focus rather than the film plane or am I missing something here (probably)?

[ardea]

Your right Brian I should have said plane of focus.

 

Thanks, Richard

[george_jiri_loun]

Sorry Richard, we are being lost here... The circle of confusion lies on the film plane for a focused gglass. For a wrongly focused gglass, well, it lies where you focus it... I thought you wanted to know the depth of focus to see what is the focusing tolerance. Anything else I don't get any sens in the question, I'm afraid. Sorry.

[bill_proud]

Richard,

 

There are two depths in focussing, depth of field for the subject, and depth of focus, which George is trying to relate, for the film plane.

The depth of focus is the same, in front of and behind the film plane. My formula is the same as George's but doesn't have a two factor so using your numbers of .06 and 32 I come up with just over 1/16".

In reality, I'd look more seriously at operator error before film plane mis-alignment but I've been wrong before.

 

good luck

[leonard_evens]

Let me try to explain how it works. You have the film plane somewhere. For the moment don't worry that it is not quite a plane or not exactly where it should be. Any image point produced by the lens will subtend a cone with base the aperture of the lens. If that image point is not in the film plane, the cone will intersect the film plane is a disc. You want that disk to be smaller than your acceptable circle of confusion, in this case .06 mm. The reason you want that is so that disc, after enlargement in a print, won't be distinuishable from a point. If you stop down to f/32, then any image point which is within the distance N x c = 32 x .06 = 1.92 mm on EITHER side of the film plane will produce a sufficiently small disc according to your standards. This is the formula that applies if the principal plane of focus is distant, as you specified. If the principal plane of focus is close up, you have to multiply by 1 + M where M is the scale of reproduction. (Even that is not quite right, but let's not go into that here.)

 

When you project the range we have been discussing on the film side of the lens out into the subject side of the lens (using the lens equation), you get what is usually called the depth of field.

 

Now let's get back to the error you are talking about. Again let's ignore the fact that the film may not be flat in a plane. Because of a possible error in the position of the gg and because of variation in the thickness or position of the film, the film plane will be slightly off from where it should be. You've estimated that this error may be as large as .09 mm. All that does is shift the whole range we have been discussing by .09 mm either towards the lens or away from the lens. Say it moves it away from the lens. That will mean that more will be in focus in the foreground and less in the background. (And vice versa if it moves it towards the lens.) But suppose when you chose f/32 as your f-stop, you already overcompensated beyond what would have been satisfactory for your desired depth of field. Say f/22 would have been adequate had everything been perfect. That means that you had (32 - 22) x .06 = .6 mm to waste. In other words, at f/22, every image point within 22 x .06 = 1.32 mm on either side of the film plane would have yielded a disc in the film plane not distinguishable from a point. 1.92 - 1.32 =.6 is your safety factor. Since .09 is within this safety factor, you should be okay.

 

You can work the arithmetic backwards the other way also. Suppose you decided to get just barely enough depth of field you theoretically would be okay with f/22. Suppose also there may be a shift in the actual position of the film plane of .09 mm. Divide the shift by the coc to get .09/.06 = 1.5. If you add this to 22, you get 23.5, and so if you stop down to f/23.5. you should be able to accomodate the shift in focus within the additional depth of field and get everything you wanted (and a bit more at the other end of the range) in focus.

 

Using the formulas for depth of focus, 2 Nc, is not really relevant with N, the f-number, equal to 32. You would be crazy to try to focus at f/32. You probably are focusing at f/5.6. Then the formula does apply,. There is a range of 2 x 5.6 x C where you won't be able to distinguish one position from another as being more in focus. Just how large this is will depend on C. That in turn will depend on the quality of your ground glass, the magnification of your loupe, and how well you see. It won't necessarily be simply related to the .06 mm you quoted. That is a figure chosen in relation to what you expect in a final print. I estimate that for me, with a 4 x loupe and my current rather coarse gg, that c is about .06 mm, but that may be something of a coincidence. But suppose we take that value. Then the depth of focus is 2 x 5.6 x .06 = 0.676 mm. That means that if I focus several times with that loupe, my different focusing positions may differ as much as that. That is much bigger than .09 mm, so in this case the error due to the gg position is overwhelmed by the much larger error in focusing. On the other hand, one may improve the situation by focusing several times and taking the average position as the one to use. That can reduce the possible range significantly. In that case, it is conceivable, but not likely that you could detect this slight error as a systematic bias in your negatives.

 

P.S. Everything said above applies when the film plane and lens plane are parallel. If the lens plane is tilted with respect to the film plane, many of the same principles apply, but it gets more complicated.

[philip_sweeney]

I thought some of the responses were a little too "techy" so hope this helps. The difference should be negligible once you stop down. The value you state for the circle of confusion (COC) is a little tight. I personally decided a few years ago to use a COC of 0.004 inches. This is 1/1500 times the diagonal of the film. Actually most small format fixed length lenses are calibrated using 1/1000 for depth of field scales. So first one must decide which spec to use. I am sure your camera is fine, do some practical tests with some polaroids! Once you decide on a COC use the formula for depth of field and do some calculations for depth of field to familiarize yourself with what the lenses can do (see appendix "the Camera" or elsewhere, its hard to type the formulas on screen - you need hyperfocal distance and then you calculate the near and a far focus points). For a 4 x 5 camera using 6 inch or longer lenses I rely on tilt and swings for depth of field control and rarely use any stop larger than f22 (I mean area not the number!). Now lets talk about a 90mm lens for a 4 x 5. With my calculations: focus at 11 feet, stop down to f22 and from 5.5 to inf everything will be in focus. Some tables may indicate otherwise, but from 5.5 to way out there, will be in focus. I have pencil marks on my camera for pre-focus at 11 feet. Now if one were to focus at 11 feet, use a tilt and stop down to f22, a tall vertical object say out 15 feet would not be in focus at the top, because we shifted the plane of focus and it is hard to see in low light at f6.8. I never use tilt with my 90mm lens. I always test a new camera or lense with a couple of polaroids. Remember: format goes up, depth of field goes down.

[philip_sweeney]

found this site with the formula: http://www.students.tut.fi/~vahonen/computer/digiphoto/DOF/

[bill_proud]

Richard,

 

Are you now sorry that you asked this question?

 

(:>,)

[leonard_evens]

It should be reiterated that his question is not really primarily about depth of field but rather the effect of focusing error due to a difference between where the gg is and where the film is.

 

If you read my "techy" long discussion, you will see that his estimated error of .09 mm in the position of the gg relative to the film can be compensated for by stopping down slightly more. That is because the additional depth of field will accomodate any focus shift due to the aforementioned error. This same consideration would apply even if one used lens tilts or swings, although the analysis is somewhat more complicated.

[philip_sweeney]

Leonard:

 

Gee wiz wally I did not mean to offend anyone! the beav

 

I thought Richard's question required him to understand hyperfocal distance, in order to understand why the differences did not matter stopped down. It sounded to me he did not. Thought I could contribute.

 

regards

 

phil

[bob_salomon]

"A 150 mm lens focused at inf. and stopped down to f32"

 

Will put you into diffraction with any modern 150. They perform optimally at

f22 and lose performance at 32 and beyond due to diffration.

[leonard_evens]

To Richard:

 

Sorry that I sounded offended; I wasn't. I think your comments were quite appropriate, and from the way he originally posted the question, he may very well be confused about such matters and may need to go over the basics.

 

I just thought the whole discussion had departed somewhat from the original question. Since he was inspired in the first place by a discussion I started, it may be that I feel have have some "ownership rights". :-)

 

To Bob:

 

I've often seen comments like yours about diffraction, and I'm not sure I understand them. According to the usual estimates---see Jacobson's Lens Tutorial, for example---the effect of diffraction is something like 1500/f-number in lp/mm. So at f/32, that would amount to 1500/32 or about 47 lp/mm. That seems to me to be well above an observable effect for most large format photography, but maybe I'm missing something. It is about comparable to the resolution of most color film, and even the resolution of most lenses, so perhaps that is the point. Could anyone elaborate more?

[ardea]

What a fun site...Thanks so much for your informed comments and your valuable time.

I was not having any focusing problems...at least not yet. My basic question was answered. Stopping down to f22 or (with respect to Mr. Salomon) f32, probably covers most manufacturing tolerances.

 

Mr Salomon raised a point that modern 150mm lenses "perform optimally at f22..." and at smaller f stops diffraction starts to degrade the image. If one NEEDED the depth of field afforded by, say f32 or 45 would not the diffraction loses be less than the "out of focus" loss from using the optimum (f22) stop?

 

I will go back over your kind posts and try to glean more info. Thanks so much...Happy new year.

 

Regards, Richard

[dave schlick]

if the ground glass is too close compared to film plane you can get a peiceof plastic or cardboard to use as a spacer where you put the film holder to correct on critical shots.. what you must be carfull is that you dont focus further than infinity. thiswill cause all kinds of trouble. and then closeups are critical too... i had this problem with a 6x6 back..

what i suspect is wrong is someone took out your fresnell and plate glass and got them mixed up.. remeber the film plane is at the fresnal not the see thru glass. the distance is measured from the fresnal not the clear glass if there is one, maybe there shouldbe one if there is not... you can move the fresnal back with a shim if that is what is wrong or maybe the see thru glass is on the back not inside the camera side.. any good repair shop should be abble to fix it for a reasonable price, you can talk in the circle of confusion for a month and waste the amount of film testing for what it would cost to get a pro to do it right.. from expericne witht his problem you will make critical mistakes and ruin some good shots if it is not fixed and you try to compensate for it..there are too many things to do right in lf as it is.... good luck, dave..