Leveling a field camera

[Dan Fromm]

The rule is: your camera is horizontal if an existing line which is vertical in nature coincides with a vertical line in the centre of your viewfinder. Edges must be disregarded completely. Trying to align anything with the edges of the viewfinder is a sure way to have a tilted horizon.

Aligning lines to the frame edge is the typical mistake people make as it is in many introductory books to photography. That doesn't work. In the first ENI building example in my previous post it would work, but in the second and third example it would not. The centre line always works.

Aren't the vertical edges of the frame parallel to each other and to a vertical line anywhere on the GG? I have a couple of ruled ground glasses (from the factory) and they have this property.

 

[Leigh B]

Aren't the vertical edges of the frame parallel to each other and to a vertical line anywhere on the GG?

To each other and to the GG... yes.

To vertical lines on the subject... maybe, maybe not, depending on where the feature is and whether the camera is level.

A vertical line through the center of the viewfinder will always align with a centered subject feature
if the camera is level, regardless of camera tilt.

If you tilt the camera up, building edges away from the image center will converge toward the top.

Also, some lenses may introduce distortion near the edges.

- Leigh

 

[markbarendt]

PS It goes without saying that while using small format the vertical line to be taken as measure is to be on the exact centre of the frame.

Actually this shouldn't have been said.

An off-center vertical is actually better because it allows you to judge both tilt (right or left) and pitch (forward or back). A vertical dead center can only help with tilt.

Get one off-center vertical right and all the rest of the verticals will fall in line too.

Going farther, if you get one off-center vertical and one off-center horizontal composed "squarely" in the frame all the rest will follow. At that point only the lines leading away from the camera will converge.

While it is possible to use these concepts on any camera, fixed lenses, that can't be moved left, right, up, or down, force significant compromises in composition and camera placement, generally leading to serious cropping.

Also lens distortion, barrel or pincushion, complicates things too.

 

[cliveh]

Actually this shouldn't have been said.

An off-center vertical is actually better because it allows you to judge both tilt (right or left) and pitch (forward or back). A vertical dead center can only help with tilt.

Get one off-center vertical right and all the rest of the verticals will fall in line too.

Going farther, if you get one off-center vertical and one off-center horizontal composed "squarely" in the frame all the rest will follow. At that point only the lines leading away from the camera will converge.

While it is possible to use these concepts on any camera, fixed lenses, that can't be moved left, right, up, or down, force significant compromises in composition and camera placement, generally leading to serious cropping.

Also lens distortion, barrel or pincushion, complicates things too.

So it's not that simple then?

 

[markbarendt]

So it's not that simple then?

It is simple.

If you want all the verticals and all the horizontals square, the film plane needs to be oriented absolutely parallel to the subject plane.

It makes no difference if the subject is to the right, left, up, down, or centered.

 

[Diapositivo]

Mark, I said in my post #20 "while using small format".

What you say, if I get it right, is a way to obtain the entire façade, or set of lines, "vertical" on the frame by using "movements".

What I am talking about is a simple rule to judge when the camera is perfectly levelled in presence of a subject which is not necessarily parallel to the focal plane with a camera pointed upward and which has no movements.

The property acutely observed and illustrated by Dan in post #26, and which is in fact common to every ground glass with some sort of checkers pattern , is - in those conditions - not shared by the lines on the resulting image, because of the perspective convergence of lines (absent "movements").

PS I perfectly know that one can use shift lenses also on 135 format. I bought a camera on purpose to use it with my two shift lenses. What I am talking about is how to deal with perspective, not how to cheat with it

 

[Leigh B]

If you want all the verticals and all the horizontals square, the film plane needs to be oriented absolutely parallel to the subject plane.

That's correct.

Accurately assessing that condition may or may not be simple.

- Leigh

 

[markbarendt]

Fabrizio,

The physics remains the same regardless of the lens or format.

The only thing that matters is the relationship of the film plane and the subject plane, they must be parallel.

There are certain tricks that can be used to get darn close with a fixed lens camera, for example if I'm on level ground and my camera is head high then placing someone's head in the view finder half way between top and bottom (anywhere from right to left) essentially levels the camera for pitch and the verticals will be darn close to right with just that and leveling to the horizon or a building edge anywhere in the frame for right left tilt. That only deals with the verticals though. You then have to pan to get the horizontals.

 

[tih]

Mark,

I think you're reading too much into what Fabrizio is saying.

There are certain tricks that can be used to get darn close with a fixed lens camera, for example if I'm on level ground and my camera is head high then placing someone's head in the view finder half way between top and bottom (anywhere from right to left) essentially levels the camera for pitch and the verticals will be darn close to right with just that and leveling to the horizon or a building edge anywhere in the frame for right left tilt. That only deals with the verticals though. You then have to pan to get the horizontals.

Correct, of course - but not the situation he was talking about.

All Fabrizio really claimed was that if you're deliberately not leveling the camera for pitch, and you have real verticals in the frame, such as building edges, then you can level the camera horizontally by making a real vertical in the center of the frame appear vertical on the ground glass - but it has to be in the center of the frame for it to work.

In other words: if the camera is on a tripod, and the tripod head is perfectly leveled, and then pitched up, panning it across the building will show each vertical on the building vertical on the ground glass when (and only when) it is exactly in the middle of the frame.

-tih

 

[Diapositivo]

Exactly.

But what I also meant to say more specifically is that "leveling" the camera or the tripod is a tricky exercise as there often remains a 0.2° mistake or so. The proof being ultimately in the pudding, what you do is to take the picture with your "levelled" camera and than check the "verticality" of a line in the centre of the frame, if any, to check that the image is actually horizontal.

Mark insists that I should only take pictures with the façade parallel to the focal plane! . I don't contest the correctness of his explanation of camera movements, I just gave a simple rule for knowing when a photo is correctly levelled "in the field", which - very importantly - can be applied also after the fact.

What one does after capture is rotate the image so as to eliminate that small 0.2° mistake which is easily there and which you really notice only after you redressed the problem. When the frame is perfectly levelled there is something snapping into place and you can almost hear the "click" ...

Rotating the image: those using a hybrid method will "rotate the canvas", those using an enlarger will I suppose either use some sort of film holder with lines to align the paper under the image or reframe the image with an enlarging easel.

And lest somebody says that art should be free and that my concept of art is limited I would also like to say that I do have a preference for correctly levelled images especially when the horizon or the ground is visible, but I might disregard horizontality sometimes, especially when the horizon or the ground are not visible (my "avatar" image is not levelled, as an example).

 

[markbarendt]

Mark insists that I should only take pictures with the façade parallel to the focal plane! .

Actually I'm not insisting or even suggesting that anyone shoot that way.

Mark is simply suggesting that math, including geometry, is a science.

 

[Diapositivo]

Mark I agree geometry is a science. Is there anything wrong in what I wrote?

I still maintain that your quotation of mine in post #28, in which you said I was wrong, it is instead right.

If a vertical line, which is vertical in reality, is also vertical in the centre of the viewfinder, the camera is horizontal regardless of the orientation of the subject and regardless of the relation between object plane and camera plane.

I still maintain that until proven wrong on the field of science.

 

[markbarendt]

Yes, you really are wrong in that case.

Format makes no difference, and leaving aside lens distortions/imperfections for clarity, any camera that is properly leveled so that the film Plane is vertical will render all verticals in the composition correctly, not just the one in the middle.

 

[Diapositivo]

Yes, you really are wrong in that case.

Format makes no difference, and leaving aside lens distortions/imperfections for clarity, any camera that is properly leveled so that the film Plane is vertical will render all verticals in the composition correctly, not just the one in the middle.

Aaarrgghhh!

I have already specified that I cited 135 as it has no movements of his own :devil:

A camera can be levelled along three axis but when we say "levelled" we mean "horizontal" parallel to the ground! A camera can be and is horizontal (on that plane) also if the film plane is not vertical!

The rule, which is absolutely right unless you have some decent argument to propose against it, is that if a line, which is vertical in reality, is vertical in the CENTRE of the screen, then the camera is properly HORIZONTAL (levelled). That is true even if the film plane is not vertical, and even if the subject "plane" is not parallel to the film plane. It's a rule that always works.

I NEVER said that if a line, which is vertical in reality, is vertical in the CENTRE of the screen then the film plane is vertical!

The film plane is in my examples NOT vertical because as I stated too many times the camera is pointing upward which creates convergence of all lines toward some vanishing points somewhere!

And I am BLOODY RIGHT on that :devil::devil::devil: Please read carefully before expressing scientific truths in the wrong context.

How tiring!

 

[markbarendt]

Aaarrgghhh!

I have already specified that I cited 135 as it has no movements of his own :devil:

A camera can be levelled along three axis but when we say "levelled" we mean "horizontal" parallel to the ground! A camera can be and is horizontal (on that plane) also if the film plane is not vertical!

The rule, which is absolutely right unless you have some decent argument to propose against it, is that if a line, which is vertical in reality, is vertical in the CENTRE of the screen, then the camera is properly HORIZONTAL (levelled). That is true even if the film plane is not vertical, and even if the subject "plane" is not parallel to the film plane. It's a rule that always works.

I NEVER said that if a line, which is vertical in reality, is vertical in the CENTRE of the screen then the film plane is vertical!

The film plane is in my examples NOT vertical because as I stated too many times the camera is pointing upward which creates convergence of all lines toward some vanishing points somewhere!

And I am BLOODY RIGHT on that :devil::devil::devil: Please read carefully before expressing scientific truths in the wrong context.

How tiring!

Ok, let's try a little context. You led the thread off topic. This thread was started to discuss leveling a field camera, one with movements.

If you want to discuss leveling cameras without movements and the compromises you might want make, it might be courteous to start a thread along those lines instead of muddying up the discussion here.

Also, it is possible that I missed it, but the first reference that I saw that you acknowledge tilting the camera Up is in your response to tih in #35.

 

[Leigh B]

This thread was started to discuss leveling a field camera, one with movements.

Whether or not a camera has movements is absolutely irrelevant as regards leveling in the general case.

The only case in which it does matter is using front and back tilts to restore those planes to vertical after the bed has been tilted.

- Leigh

 

[Diapositivo]

Mark, in post #1 the user basically wondered about the reliability of spirit level when on the field. He posed a question about levelling and about levels accuracy.

In post #17 the question is raised again about discrepancies between what the picture says and what the instrument (level, camera) says.

In my post #20 I said that I do observe discrepancies between what the level says and what the image says. My remedy is to adjust the image after the fact, using the "vertical line in the centre" rule. This was perfectly on topic.

In post #21 Clive asked an explanation about that rule. A legitimate question!

In post #22 I answered Clive's question! That's certainly not off-topic. It's all about levelling an image. Post #23, 24 and 25 dealt again with the simple rule.

In post #28 you said that the rule was wrong. Which is wrong because the rule is right. And the matter is in topic.

The rest were boring posts in which you applied geometric principles to things I had not said.

Fabrizio

 

[markbarendt]

Whether or not a camera has movements is absolutely irrelevant as regards leveling in the general case.

The only case in which it does matter is using front and back tilts to restore those planes to vertical after the bed has been tilted.

- Leigh

Absolutely

 

[markbarendt]

Mark, in post #1 the user basically wondered about the reliability of spirit level when on the field. He posed a question about levelling and about levels accuracy.

In post #17 the question is raised again about discrepancies between what the picture says and what the instrument (level, camera) says.

In my post #20 I said that I do observe discrepancies between what the level says and what the image says. My remedy is to adjust the image after the fact, using the "vertical line in the centre" rule. This was perfectly on topic.

In post #21 Clive asked an explanation about that rule. A legitimate question!

In post #22 I answered Clive's question! That's certainly not off-topic. It's all about levelling an image. Post #23, 24 and 25 dealt again with the simple rule.

In post #28 you said that the rule was wrong. Which is wrong because the rule is right. And the matter is in topic.

The rest were boring posts in which you applied geometric principles to things I had not said.

Fabrizio

So yes I did miss the comment in 31 where you said not parallel to subject. In that case a vertical line dead center can provide a reference to square up to.

My apologies for not reading that close enough.

Still the "ps" statement in 20 is just wrong, the format has nothing to do with the problem. The lenses you describe in 31 prove that. It falls back to the orientation of the film to the subject.

 

[Diapositivo]

Still the "ps" statement in 20 is just wrong, the format has nothing to do with the problem. The lenses you describe in 31 prove that. It falls back to the orientation of the film to the subject.

I see what you mean, yes the film format has nothing to do with geometry. That "while using small format" meant "when there is no way to tame lines" (camera movements) but the principle holds true in any case. The "while using small format" was meant to avoid objections of the kind "while using movements all lines can be made vertical regardless of camera orientation".

 

[Alan Klein]

I shoot MF but was wondering if ball heads are easier to use then three way head for leveling and keeping the tripod level?

 

[Leigh B]

[personal opinion]
Ball heads are a total fraud as regards still photography. They're designed for video cameras when you don't care which end is up.

Pan/tilt heads are much easier to set up for any format that requires leveling because you set each axis individually and it stays put.
[/personal opinion]

I too shoot MF (Hasselblad both film and digital) and would never consider using a ball head for it.

- Leigh

 

[Diapositivo]

Video cameras typically use 2-way heads because they must never loose "horizontality" once properly set. You can move the camera while "filming" on two axes without undergoing the risk of a movement on the third axis.

Ball heads are certainly faster in use as all three movements are performed at the same time. I personally find that for urban landscape / architecture photography accurate levelling is important. Ball heads don't allow a precise levelling just because all movements are "moved" at the same time. If you reach "perfection" on two of the three axes, as soon as you try to adjust the third axis you loose the set of the first two.

Three-way heads are apparently slower as you spend a lot of time manipulating knobs, but if you need to work with a certain precision they end up being faster, because you do the work only once.

If I had to draw a line I would say that ball heads are probably better suited for all those works, such as naturalistic photography / birdwatching, where readiness is more important than correct levelling and precise framing.

 

[Kevin Kehler]

I can't stand ball heads, for the exact reason listed above - as soon as you want to make a minor adjustment to any one angle, all angles need to be re-adjusted. It is not so bad when the camera is <5-6 lbs. but put a 13-14 lbs. camera on a ball head with any sort of angle to it, it falls over as soon as you loosen it for a small adjustment (notice I didn't say falls over due to the weight of the camera as there as some very heavy duty ball heads but you have to loosen the head to adjust it and there is no easy way to do that while holding the camera and re-locking the knob). I have several 3-way heads and as they say "slow is smooth, smooth is fast". Yes, it is not instantly locked to the ideal position but when does it ever get to the ideal position the first time, every time? I prefer the ability to make small adjustments easily.

 

[Dan Fromm]

I shoot MF but was wondering if ball heads are easier to use then three way head for leveling and keeping the tripod level?

Absolutely not. Neither type will do the job. If the tripod's platform isn't level, the head's pan axis can't be made vertical. This is true for ball heads and for pan/tilt heads.

With a pan/tilt head (two- or three-axis) the camera's orientation on one axis can be adjusted without changing its orientation on any of the others. This is impossible with a ball head; when the camera is free to move on one axis, it is free to move on all. What can move, will move. That is the pan/tilt's head advantage over a ball head.

Real cinematographers (spelled "Hollywood") use tripods made for cinematography. These tripods incorporate a ball leveler (also called claw ball) that sits under a pan/tilt head. The leveler is used to make the head's pan axis vertical.

Still photographers rarely pan while capturing an image but cinematographers sometimes pan during shots. If the pan axis is vertical, the horizon won't move (rise or fall) during a pan. If the pan axis isn't vertical, the horizon will move during a pan. I think you can see why cinematographers regard the ball leveler as essential.