Depth-of-field correction

As a new 4x5 photographer, I have recently made my first large

(32"x40") print, and I've noticed a problem: My photo, taken with a

150mm lens, has foreground at about 20 feet, with background at about

200 yards. The background is in the upper third of the photo, and

the foreground is in the lower third of the photo.

 

Now, when I set up my view camera, I tilted the lens so that I

simultaneously put both the background and the foreground into sharp

focus, and then I stopped the lens down to F22, and took the photo.

 

My problem is this: I love the sharpness of the 32x40 print in both

the foreground and the background--but the middle-ground is just ever

so slighly out of focus.

 

In my day job, I am an engineer, and it occurs to me that the tilt of

a view camera provides "first-order" correction to focus errors.

But, my photo clearly needs second-order correction to accomplish

superior sharpness.

 

Here is my question: has anyone out there done any experimentation

with second-order correction to focusing? One way to do this would

be to put a cylidrical glass plate in front of the film. Ideally, of

course, the radius-of-curvature of the glass plate should be variable-

-a difficult trick. Any experimentation out there?

In your example photograph the lens tilt hasn't brought all objects to equal focus because they objects don't lie on a plane. The unstated assumption in your ideas for "second-order" focus correction is that all of the objects fall on a cylinder. If the middle-ground had a tree for which the upper-branches fell on the plane of best focus, somhow making a cylindrical surface of focus would improve some of the middle ground, but would make the focus of the tree worse.

 

Probably an easier way of making a cylindrical question than using a cylindrical lens in front of the film would be to curve the film in some sort of special filmholder. You could either have to be able to curve the ground glass to judge the effect, or calculate the required curvature from the focus positions of the ground glass.

 

Considering the difficulties of implementing the cylindrical focus, and the few cases in which the subjects follow a smooth second order curve much better than a plane, the idea doesn't seem worth the work. What most photograhers do is to stop down further. You could have probably stopped down to f32 to improve the sharpness of the out-of-focus area, with little or no decrease in sharpness of the focused areas from diffraction. If the subject was sufficiently non-planar, you would need to stop down even further and accept some loss of sharpness from diffraction, or if you prefered to maintain optimum sharpness in the in-focus areas, you would have to accept some areas being out-of-focus.

 

Most people view large prints from a distanct of approximately the diagonal of the print. I am guessing from your description "just ever so slightly out of focus" that the slightly out-of-focus nature of the middle ground is seen by you only from close examination. If so, most viewers won't notice.

Michael,

Thanks for the discussion. Regarding your point at the end:

 

"Most people view large prints from a distanct of approximately the diagonal of the print. I am guessing from your description "just ever so slightly out of focus" that the slightly out-of-focus nature of the middle ground is seen by you only from close examination. If so, most viewers won't notice"

 

The only thing that has drawn me to LF photography is the (to me) compelling notion of seeing a large photo across the room, and then, being drawn into it by the ever-increasing complexity and detail that is revealed as one approaches the photo. For me, this is the entire reason for LF as an art medium.

 

At the moment, I have little experience, and will be experimenting with stopping down to f32 or f45, but I have been shooting some resolution charts in the back yard with 150 Sironar N and 110 SuperSymmar XL--am learning that these lenses are incredible at f11 and f16, but f32 starts to take a toll.

 

I have thought about the curved-film idea, but rejected it because of the concommitant need for a similarly-curved ground-glass. It seems to me that, if one needs to introduce curvature, it would be best to introduce one curve to serve both for the focusing and for the exposing.

 

Anyone out there ever experimented with this kind of thing?

 

Thanks again, Michael.

This is the way it works:

you have focused on the foreground and then tilted either the lens plane or the film

plane to get the background in focus (top & bottom third. But now, even at f/22 your

middle ground is slightly out of focus. This is quite common. Essentially what has

happened is that, from the camera's optical perspective the top and bottom of your

scene, where you have focused have become the closest plane of focus and the

middle ground is geometrically further away from this plane. What you need to do,

without changing your tilt or resorting to a secondary lens, is find an intermediate

plane of focus between those apparent near and far points and then stop down

enough to pull both the near and far points into focus. Some cameras -- I am

thinking of the Sinar F,C, P & X cameras -- have built in depth of field & tilt/swing

calculators to help you find the correct focus position. Most cameras do not. I use a

$35.00 tool from Rodenstock that is essentially a circular slide rule to help me find

my correct focusing distance , f-stop, bellows extention factor (if any), and tilt or

swing angle. it takes into account format size (this plays a big role in determining the

'circle of confusion" which is responsible for "sharp" resolution, and varies from

format to format) reproduction ratio (size of object being photographed to the size it

is reproduced on film), camera base tilt angle, It is very handy and slips into a shirt

pocket. if your are interested, post back and I'll dig up the product number.

The Rodenstock analogue calculator is a very useful tool - but it should be noted that your camera has to have metric with scales (with 1mm/1degree precision) for focus and tilt/swing angles in order to apply the calculated results.

<br>The calculator works O.K. with my Arca F-line camera (though applying 6.5 degree tilt with a scale where only 5 degrees are marked can be fiddly), but is essentially of no use with my Toho FC 45X (which has only inches marked for horizontal shifts).

It would also help to know which 150mm lens you are using...

<I>"The Rodenstock analogue calculator is a very useful tool - but it should be noted

that your camera has to have metric with scales (with 1mm/1degree precision) for

focus and tilt/swing angles in order to apply the calculated results. "</I><P>

 

No it doesn't although that is handy; The calculator has a 9cm scale along one edge.

Hi steve,

before stopping down to 45 remember, that at those small apertures diffractions become a reason for unsharpness. 150mm is a fairly short lens for 4x5 film. Therefore the method to find the middle between all points to be in focus should be more efficient then stopping down.

After all, you might shoot all three alternatives (first you got already, stopped down and intermediate focus) an compare results.

Martin

While Martin's statement is true, most of the time you're much better off stopping down enough to gain the needed depth of field and accepting a slight loss of "sharpness" due to diffraction than you are opening up the aperture to eliminate diffraction but then losing the needed depth of field. In other words, inadequate depth of field is usually more noticeable than the effects of diffraction.

Response to Bob Fowler:

 

Bob, I'm curious about why you think the lens may be important. The lens in may particular case was a 150mm Rodenstock Sironar N.

Steve

Ellis and Stefan:

Thanks for the Rodenstock calculator suggestion. This seems like the most effective next-approach for improving my work.

Steve

I agree with the others. Your problem with the middle distance is that it is outside of the DOF region. This region is an open "wedge", which starts on a line below the lens---called the hinge line--- and opens outward without an upper limit. The plane of exact focus is roughly speaking in the center of that wedge. As you rack the standard back and forth, the plane of exact focus and the wedge surrounding it rotate on the hinge line. The smaller the aperture, the larger the wedge. It is in fact possible to calculate in advance roughly how much you have to stop down to have the wedge include a desired part of the scene, once the plane of exact focus has been determined, but most people prefer to do it visually.

 

The region of exact focus is quite narrow close to the camera and widens as you get further away. That is why it is not usually possible to use tilts productively when there is large vertical extent near the lens which must be in focus.

 

Since you have an engineering background, you might be interested in looking at my essay

www.math.northwestern.edu/~len/photos/pages/dof_essay.pdf

It discusses all these matters in some detail. You can also study the discussions of Bob Wheeler and Merklinger, each of which also go into these matters in some detail. You can find links to them and a lot of other useful information at

www.largeformatphotogray.info

 

There is in fact a second order correction, but in most practical situations, it is too small to matter. It is caused by the fact that the boundaries of the wedge are not planes but in fact curved surfaces. This is discussed in Bob Wheeler's notes to a limited degree and also in my essay. The complete analysis is quite complex. I've done it but I've never written it up. Suffice to say that the second order effect is neglible in practical photography except possibly when using extremely wide angle lenses.

Your focusing priorities are incorrect. Most of us do exactly what

you have done, place the plane of sharp focus on foreground

and the extreme background. The proper way to focus is to

place the plane of sharp focus on the foreground and

middle-ground and manage the extreme background with DOF.

This is because for a given DOF, the DOF for both the foreground

and middle-ground is small compared to the extreme

background which can be many 1000s of feet.

 

On another note my priorities for producing a photograph are

mood and composition are number one. Technical perfection is

secondary. I will sacrifice technical perfection if slows me down

or gets in my way. One of my best selling photographs has the

very same technical imperfection you are talking about. The

middle-ground is out of focus, but no ones sees that because

the mood of the photograph is so striking. I have attached a

scan of this photograph to this response. The scan will not

show the middle-ground out of focus, but believe me it is way out

of focus. I setup my 4x5 camera, focused, metered, and took this

shoot in less then 3 minutes.

It does not look like my attachment worked. Let me try again.

im not a math expert or engeneer but i have very good luck setting the camera level, keepong lensboard square and using shifts.. going to wider lense to keep all waanted subjects on the film. then go to lenses chart for depth of field to cover all that is needed to keep in focus in the photo.. i never push the chart i give at least one step of factor(sorry for thepoor explanation) for safty.. this in its simplicity and speed of set up on a crown graphic gives me very good results with a 135 rodenstock sirinar s and super angulon 90 f8.. i feel tilts are for strait rectagular shapes like buidings.. i do agree that probably most mistakes are made by focusing on top and bottom and not checking the middle.. like was said earlier.. front and middle focus with depth of field correction would maybe give the best results for tilts..but your getting into the millions of things problem, which tends for a photgrapher to make silly mistakes while tending to the millions... good luck dave..

 

im still upset the buffalo moved !!

<p>Ellis has a good point. When you found the tilt that simultaneously placed the near and far in focus, you probably had found close to the best tilt for the situation. But when you left the focus so that the near and far were in focus wide open, but the middle wasn't, you didn't have the optimum focus position. At that point you want to note where the back is position for the focus on the near and far, then change position to put the middle in perfect focus and note the position. The best position for the focus will be halfway in between those two positions. This is explained more extensively at <a href="http://www.largeformatphotography.info/how-to-focus.html">http://www.largeformatphotography.info/how-to-focus.html</a>. The last part that I described for your particular case is II-2 in that explanation.</p>

 

<p>As a bonus, you will have measured the difference in focus positions and can use this to help you select the f-stop to use: see <a href="http://www.largeformatphotography.info/fstop.html">http://www.largeformatphotography.info/fstop.html</a>.</p>

 

<p>One suggestion for your lens testing: be sure to check the sharpness across the field, not just at the center. You might want to use the maximum front rise that you are likely to use (or an equivalent shift, if more convenient). Rather than using lens charts, I just use a brick wall and qualitatively judge the performance of the lens. You need to be sure to have the optical axis perpendicular to the wall. I think you will find that for 4x5 with some movements, f11 won't give you incredible sharpness over the entire negative. With several modern lenses that I tested, I found that f32 wasn't greatly off in sharpness, considering the entire field, to the best the lens could do.</p>

 

<p>I agree with Brian that if the only way to get the desired objects all in focus is to stop down past the point that the lens gives its best sharpness, than that is the best thing to do. The viewer is more likely to notice an out-of-focus fuzzy object in comparison to in-focus sharp objects. If the entire print is slightly off in sharpness, they won't notice because they won't have a reference. That said, knowing the technial capabilities of our cameras can lead to an artistic trap of <i>always</i> trying to get everything in focus. Sometimes a photo is better with some items out-of-focus.</p>

 

<p>For landscape work I don't find the Rodenstock calculator that useful -- visual methods and using the scale on my camera are at least as quick. Others will have different opinions and it is worth trying. I find it more useful for studio work, where tilts or swings tend to be larger.</p>

 

<p>Two past threads that you might find interesting: <i>Beginner, Optimum Aperture</i> at <a href="http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=004NpY">http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=004NpY</a> and <i>Clarification on f11 sweet-spot</i> at <a href="http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=004Olk">http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=004Olk</a>.</p>

We all go through this. Focus on the near subject then the middle ground and trust

that the distant part of your image will snap into focus by depth of field. I usually add

another stop to whatever f-stop I want to use just for safety. You have to think of

depth of field as being "vertical" as well as distance (flat). Depth of field becomes

"vertically" taller (longer) with every higher f stop chosen. Just as depth of field

increases into the distance, it also increases vertically (up and down) from the plane

of focus by using higher f stop numbers. I suggest that you need to make your near

and middle areas sharp on the groundglass and forget about the distance. Also you

need to use higher f stops than you think you need. This technique helped me.