Focussing (again!!!!!)

[paul owen]

Help, I am a new (lonely) LF photographer, (lonely because LF is highly unfashio nable here in U.K.). Question - In landscapes where the requirement is for depth of field from 3-4 feet to infinity I am finding that by focussing at infinity a nd then watching the GG as I tilt the lens, I require about 3-5 degrees of tilt to get the foreground in focus ( I found this out through your web site !). BUT on using an aperture of f22 the distant background is not in sharp focus. What a m I doing wrong - should I adjust my plane of focus to nearer the camera or use a smaller f-stop. I am using a Polaroid back but the image produced does not acc urately show depth of field. I would welcome any help/tips as Polaroid film costs a fortune here !! I like landscapes that have plenty of foreground detail as well as a sharp backg round. In MF I use a Silvestri with a 6x9 roll film back and with my 47mm Schnei der or 100mm Schneider have no problems despite a lack of tilt on the lens. My L F camera is an Ebony SW45 and I am using a Schneider 110mm XL. Am I expecting to o much from my LF ?

[fw]

On near/far type shots I can usually get both foreground and background in sharp focus on the groundglass, even at full aperture, which is consistent with Scheimpflug. Stopping down in itself only increases the sharpness of items that fall outside the chosen plane of sharp focus. Hence I think you may need to experiment by varying both lense focus and tilt angle together until you get what you want in sharp focus on the gg. Then you can probably see what remains out of focus, and by stopping down the lense (and drawing your darkcloth tight) you should be able to see the depth of field in the image changing. On complicated images, it can take me several "goes" to get it right, which means I have to bring the lense back to neutral tilt and median focus, and start again! Sinar has a system which enables the optimum tilt angle and the optimum aperture to be selected, and in an earlier thread on this board, I asked if the Sinar methodology could be replicated in a series of systematic focus/tilt steps. It probably can, but most people probably rely on a more intuitive approach by examining the gg. Good luck.

[Ron Shaw]

You will need to re-focus as you apply tilt. Apply tilt until your near/far points of interrest are equally out of focus, then re-focus, and make more adjustments if needed.

[Tony Brent]

A general rule is" "Focus on the Far; Tilt toward the Near." You do need to adjust focus after every movement. Eventually you will find the setting where the whole image (or at least the portion you want) will come into focus at once.

[Patrick Chase]

Hi Paul;

As a general rule, only things which lie along the tilt axis of your camera will stay in focus when you tilt. Thus, ANY rule of thumb, such as "focus far, tilt near", is specific to whatever camera the speaker happens to be using. Even then, it's applicability depends on where in the scene the "far" and "near" object happen to lie.

Fortunately, all cameras fall into one of 3 broad classes:

1. Cameras with center tilts: In these cameras, the tilt axis passes through the center of the lens. If you have one of these, then you should focus on whatever happens to lie at the center of the lens' FOV (which may or may not be the center of the film, depending on what other movements you've employed), and then tilt to get the rest in focus. If the horizon happens to be positioned at the center of the lens FOV, which it probably will be if the body has been levelled, then this is equivalent to "focus far, tilt near".

2. Cameras with base tilts: You MUST refocus every time you adjust tilt if your camera usess bae tilts. This really isn't as onerous as it sounds; My camera only has base tilts, and it barely causes me to break stride. In fact, I often find it preferable to axis-tilt designs (which may have something to do with the fact that it's yaw-free...).

3. Other: Sinar makes cameras with "asymmetric tilts". These are basically axis-tilt designs in which the tilt axis is laterally offset from the lens' optical axis. Onb these cameras, you focus wherever the tilt axis happens to be, and then tilt to get the rest in focus.

[Alan Gibson]

Don't worry, Paul, there are quite a few of us in the UK, honest.

The advice above is good. From your post, it sounds as if you are focusing, and tilting, and that's all. As a general rule, you have to keep repeating focus/tilt/focus/tilt until your chosen object plane is in focus. Then, if you want more than just a plane to be in focus, you decide how much to stop down.

As commented above, some cameras tilt in a precise manner. If you know exactly how the camera works, you should in theory be able get a specific part of the GGS in focus, then tilt, and that's that. In practice, you should always check the whole GGS, and be prepared to re-focus, re-tilt, etc.

[paul owen]

Thanks to everyone who replied !! I,m still struggling !! I exposed a few sheets today and I think I,m getting there - slowly !! Silly Question Number 2 !! My camera uses centre tilts on the front standard, therefore if I focus on the lens, central field of view and then tilt to get the foreground in focus, do I need to re-focus on the background/central field of view - I have tried re-focussing and it appears to have little effect on the GG image. Are there any LF (landscape-type preferably) enthusiasts in the U.K. (maybe Wales)who fancy a get together to discuss LF on a regular basis???? Thanks again for all the tips ! Paul.

[Les Moore]

When trying to get the desired tilt and focus I first tilt the lens a bit, focus on my foreground, and then look at where the sharpest focus is in the background on the ground glass. If the sharpest focus in the background is too high, I INCREASE the tilt a little, refocus on the foreground, and check the background focus again. if the background focus is too low, I DECREASE my tilt a little, refocus on the foreground, and then check the background focus again. A few quick iterations of this technique will allow you to get the desired tilt/focus. I usually try to get my sharpest background focus in the middle of the background vertically and then stop down enough to get the top of my background in acceptable focus on the ground glass and then usually stop down another stop to be safe.

[Robert A. Zeichner]

There are probably as many techniques for focusing when employing a Scheimpflug relationship as there are photographers doing so. Here's a technique I learned from a master in the field. I think it might help. With the camera fully neutral, establish your basic composition. Stand off to the side of your camera set up and visualize the Schempflug plane (adjusted subject focus plane). A yardstick (or meterstick for those on the east side of the pond) can be used to help visualize where all the planes converge. Make an estimated adjustment of either the back or front standard and refocus. This will help you get close enough to be able to make minor adjustments needed to optimize the Scheimpflug relationship. Now you focus far and near, observing the range, in millimeters that your focusing bed moves. Center the lens in the middle of that range and use the spread to determine aperture needed to achieve desired depth of field. Here are the numbers: 0.7mm = f16, 1.3mm = f22, 2.7mm = f32, 5.4mm = f45 With 4x5, when you get beyond f45 you run the risk of noticeable diffraction artifacts and so it's best to reconsider the lens, or refine your adjustments. You can tape or cement a plastic millimeter scale to the staionary portion of your cameras focusing bed and paint or scratch a witness mark on the moving part to make these measurements. One final thought: When you establish a Scheimpflug relationship, the subject plane (plane of focus) is one where depth of focus behind that plane remains double the distance of depth of focus in front. This is just as though the subject plane is parallel to the film! Knowing this means that you can establish points near and far through which the plane will pass, allowing depth of field to bring the remaining extremes into acceptably sharp focus. This is a little hard to describe, but easy to diagram. Imagine a scene in which you have a 1 meter high shrub in the foreground and a 300 meter mountain in the background. By establishing a pair of points, one about two thirds of the way up the shrub, the other two thirds the way up the mountain, then making the subject plane intersect these two points, you then only need to focus on near and far to figure the required aperture and you're set. The farthest focus point will be at the base of the mountain! The nearest will be the tops of the shrub! Let's say bringing these two points into sharp focus requires your moving the lens in and out over a range of 2.7mm. Park the lens exactly in the middle of that range, set the aperture to f32, and make your exposure. I hope all this helps you in some way. It may sound a little involved, but it does work and practice will make it all second nature. Good Luck

[Patrick Chase]

I wanted to comment a bit on Robert Zeicher's method for evaluating DOF. Measuring the range of focussing bed positions at which the image is in-focus, putting the standard in the middle, and then using the total excursion to determine the required aperture is a very sound approach (in fact, it's the approach I almost invariably use).

The concern I have is with the specific relationship of focussing range to aperture that Robert advocates: It's a little loose for critical work, IMO. For example, a range of 0.7 mm at f/16 corresponds to a maximum blur diameter on film of ~0.05 mm. That's OK if you don't plan on enlarging by much more than a factor of 2, but beyond that you will be sacrificing detail that the human eye could otherwise perceive under critical, close-up inspection. I prefer to use 0.03 mm instead as a starting poin for permissible blur diameter. To get the ranges which yield this maximum blur diameter, simply multiply Robert's numbers for each aperture by 0.6.

There is one exception: As Robert notes, you start to lose more sharpness to diffraction than to defocus at very small apertures. I consequently "de-rate" the maximum permissible defocus-induced blur diameter as I stop down: Up to f/32, I use 0.03. At f/45, I use 0.05. At f/64 and beyond I try to find another way to get the shot (failing that, I sometimes allow as much as 0.1 mm, though the resulting images do not hold up under high enlargement). Note also that these are extension-compensated apertures: If you're focussed at 1:1 (i.e. total extension = double focal length) then you need to double the f/# before you do any focus calculations.

Regards,

Patrick

[Michael Phifer]

I have been using the method and values that Robert Zeichner describes, since I read about it a few years ago in Photo Techniques. I thought the focus spread, a perture values (0.7mm=f16; 1.3mm=f22; 2.7mm=f32; 5.4mm=f45) were for the optimu m sharpest possible. However, because of reading Patrick's comments, I was not s ure. This made me hunt for the original article describing the View Camera Focus ing Method. (From my findings below it appears that Patrick is off by about a fa ctor of two in his defocus circle of confusion. You should get about a 10% sharp er photo using the above table as is, than if you divided by 0.6 as Patrick sugg ests)

In the March/April 1996 issue Paul Hansma's (excellent) article does describe an "Optimum f-stop Method" - that is, the f-stop (N) that makes the resolution (R= 2/C) as high as possible for a given focus spread (dv). Mr. Hansma shows that th ere is an "optimum" f-stop based on the balancing of the diffraction circle of c onfusion (Cdiffraction =N/750 ) , with the defocus circle of confusion ( Cdefocu s = dv/(2N) ). The resultant combined (diffraction & defocus) circle of confusi on (C) is the square root of the sum of the squares of the diffraction and defoc us circle of confusion { C=SQRT((N/750)^2+(dv/2N)^2) }. If one cranks some numbe rs, it can be seen that there is an optimum f-stop which occurs, when the diffra ction and defocus circle of confusions are equal. Mr. Hansma (fortunately) did the calculus for us and the Optimum f-stop, (N)= SQ RT(375*dv).

Please don't let the math scare you away. All you need to know is the method tha t Robert Zeicher described so very well above. It has helped my photographs imme nsely. (Say, "good by to Hyperfocal", and "Thank You" Mr. Hansma !!!!)

Mike Phifer - Just an Amateur

[Glenn Kroeger]

I agree with Robert Zeichner's answer in principle, and use that technique, although I have found myself modifying it due in part to dissatisfaction with the results and in part to a different perspective on depth of field offered by Harold Merklinger in his books and articles.

Harolds writings can be found at:

http://fox.nstn.ca/~hmmerk/HMbook14.html

and represent a tour-de-force in mathematical analysis of geometric lens behavior. My problem, and Harolds original problem, was the impression that infinity was not sharp enough even when DOF numbers indicated it should have a small enough circle of confusion. Harold's math points out that even though mathematical depth of field extends further behind the plane of best focus, perceptual depth of field often favors nearer objects. He suggest always focusing at infinity (for no movements). My take on this is a bit different. I think the perception is due to smaller detail in the distance. In a typical landscape shot, there is not much small detail in a nearby blade of grass, it's close so it's pretty big, but the eye wants to see all of the tree branches on the distant slopes. What I have found works well, is to find the near and far focus points, but set the focus 2/3 of the way toward the FAR point, then using a smaller f-stop based on 4/3 the distance the standard moved to catch the near. This results in the DOF extending somewhat beyond infinity which some would find a waste of good DOF. I find it makes distant images conform to a higher standard of focus that my eye seems to demand.

[fw]

Glen - interesting response. Could you please elaborate on how you compute your f-stop - i.e. how you convert 4/3 of the distance into an f-stop equivalent?

[Glenn Kroeger]

fw:

It's pretty simple, and I don't use a calculator! Take the numbers from Robert. Lets say the travel from far to near is 0.7mm. With Roberts table that suggests f/16 and move the focus 0.35mm back toward the far. I would move the focus 0.5mm back toward the far, about 2/3 of the way, then set the aperture as if the total travel had been twice that (0.5 * 2 = 1.0mm). So I would end up using an aperture about halfway between f/16 and f/22. So it's like using a tighter circle of confusion but biasing the focus toward the far.

Glenn

[fw]

Glenn ; Thanks, I understand what you're doing now. In about 50% of my photographs, I can get the movements set in such a way that almost everything of importance is in focus anyway, even at full aperture, and then I would normally operate at f/16 or f/22, depending on wind/shutter speed required etc. What you have described probably applies in about 30% of my photos, where I cannot get both near and far fully in focus, and I'll try this technique out. There is a third category where I can get far and near in focus, but the middle ground of the image stubbornly refuses to get sharp, as it is in a different plane to the plane of focus. So far, I have simply relied on small apertures, f/45 or f/64, to guesstimate my way through, with mixed results. I'm now wondering if it would be better to get near and middle in focus, and then stop down to f/22 or f/32 to get the far in focus.

[Robert A. Zeichner]

In response to that last post, the situation you described where near and far are in focus, but middleground is not is one of those instances where the apparent subject distance and the true optical distance are different. Think of it this way: By tilting you've established a plane of focus that is no longer parallel to the film. As such, objects in front of or behind that plane of focus may in fact be above or below it! Take the example of a scene where you have a small bush in the foreground and a mountain in the background. The ground between is for purposes of this illustration, flat and level. The nearest and furthest objects in the scene that you wish to have sharp are the bush and the mountain. So, you tilt the rear standard backward to establish a focus plane that intersects these two objects. But where along the vertical rise of the bush and mountain do you place this plane? If you were to pick a point about one third of the way down from the tops of each and use those to points to anchor your new focus plane, you will also be establishing a one third forward (think above) and two thirds behind (think below) relationship that an appropriate amount of depth of field will work for you to bring everything else above and below in acceptable focus. So, where is the optically furthest point from the film? If you guessed the base of the mountain, you've got it. It may seem like the middleground, but it's really the new "background". I hope all this makes sense. It's so difficult to draw this diagram with words alone.

[Glenn Kroeger]

Robert just took the words out of my mouth (or keyboard). The "far" isn't always the farthest thing from the camera location. By the way, I think the numbers in Roberts table are cm not mm, since 0.7mm is a pretty small focus shift!