Anyone Feel Like Explaining Tilt/Shift To a Dummy?

[scott_fleming1]

OK, I read Luminous Landscape's ity bity explanation of

movements ... and I did not get it.

 

http://luminous-landscape.com/tutorials/movements.shtml

 

I'm not asking for a detailed treatise just a geometric base

explanation. Mr. Reichmann has a diagram but I can't figure if it

is describing the verticle, horizontal both or either? He shows a

picture taken while he was laying in the middle of a highway but I

just don't get what angle he was tilting his lens. Was the front

lens tilting down, as I would be if I was leaning into a stiff wind,

or up as in the reverse of a man leaning into a wind.

 

How might this apply to photographing a cascading sequence of

cataracts ... like a stairway? Are we simply talking about tilting

the lense and or film plane so that it is more parallel to the angle

of the cataract, or stair ... thus creating infinite depth of field

while staying near f8? (that's not it, ... is it .... ?)

 

 

And how would this apply to his highway shot?

 

Then ... what are sideways anglings of the lense for?

 

Thanks for any help y'all can provide.

 

Scott Fleming

[todd peach seattle, washi]

<i>

He shows a picture taken while he was laying in the middle of a highway but I just don't get what angle he

was tilting his lens. Was the front lens tilting down</i>

<p>

Yes, the lens was tilting down in the highway shot. His 'scheimflug' diagram that accompanys it should be interpreted as viewed from the side; lens tilts down.

<p>

Let me try putting scheimflug (maximizing DOF through lens tilt) another way: if your subject plane and film plane aren't parallel, if you 'split the difference' between them with your lens plane, you will maximize DOF in the subject plane. So yeah, we are trying to approach infinite DOF at f/8 (or at least enough DOF to solve some tricky problems).

<p>

<i>

what are sideways anglings of the lense for?

</i><p>

Same thing only sideways. Imagine you're standing in the street, and across the sidewalk from you is a white picket fence. If you shoot 'down the sidewalk', DOF will limit what portion of the fence is in focus. By tilting your lens towards the fence (laterally, sideways), you can increase DOF in the plane of the fence.

<p>

Your specific questions were about tilt, not shift (though you mentioned shift in your title). Shift, rise, and fall are all movements of the lens in its own plane. The most basic example of a rise is photographing a tall building. Without rise, you would point the camera up to take in the building. This would make the camera not parallel to the building, and results in keystoning. With rise, the film plane remains parallel to the building, but the lens alone is 'raised' up a few mm. This allows the top of the building to be seen without keystoning.

[NetR]

Hello SCott

 

I'll have a go.

 

Firstly; shift. Your lens aperture is round and delivers a round circle of light towards the film plane (except that some is cut off on the way inside the camera) and then onto a square or rectangular negative or slide. The image circle is usually not much bigger than the maximum diagonal distance across the film. Roughly in symbols: (X), if you imagine the brackets as being a complete circle and the x as being the diagonals of the film. Now,if you want to photograph a tall building (say) or tree or whatever, if you stand back far enough to get it all into the shot while the film back is straight up and down, then the shot's lower half has a whole lot of foreground in it. If you tilt the camera back to get rid of the foreground from the composition, then the building or whatever will appear to be tilting backwards and won't look natural. It will appear like / \ instead of like | |. With a shift lens, the lens is physically larger and produces a much larger image circle. There is room in this circle for the square or rectangular film size to be in many different places within the circle. For a building, you start off with the camera back vertical, so that the image has a bunch of foreground in it, but then shift the lens upward, so that the foreground is no longer in the shot. It is as if you were standing several yards up in the air. The result is a picture that has no tilt and no un-necessary foreground. You could have taken the shot with all the foreground in it, then cropped it back so that you only had the building in it. But this would use only a quarter or so of the negative and the quality would not have been very good when you enlarged the image. With shift, you get the exact same details and proportions of the shot, but full size on the negative or slide. So you are not giving up quality when you enlarge it.

 

So shift is used to keep parallel lines parallel for tall structures. You can also use shift sideways to get the impression of (say) standing between the tracks of a railway line as a train comes, when you are actually standing to the side.

 

Secondly, tilt is used to increase depth of field. The concept is named after Scheimpflug (spelling?) although I'm not sure he invented the idea. The idea is that you get maximum depth of field in a shot of something that is at an angle to the camera axis if the lens axis is tilted to an angle halfway between the angle of the film plane and the angle of whatever the subject of the photo is. This enables you to get a lot of depth of field from something that is fairly close to you. That is, if the film back is upright | and the subject is level __ then the maximum depth of field is obtained when you tilt the lens to /. You tilt the lens separately from tilting the film plane.

 

You can combine tilt and shift to get increased depth of field.

 

People doing architectural photograohy have most use for shift. Macro and nature photographers have most use for tilt.

 

Shift and tilt are actually most easily obtained with large format cameras, which also provide swing and other fancy stuff. Hasselblad did things differently and had camera bodies that provided the tilt instead of using a lens.

 

I hope this helps, regards, Ross.

[bob_salomon3]

Tilts and swings control the plane of focus Aperture controls DOF.

[andre_oldani]

Scott, shift is parallel to the film plane (positive or negative). I made some test

shots with my Alpa 12 SWA to demonstrate this. The image circle needed to fit

a 6x6 shot (56x56mm) for example is some 80mm (diagonal). The Schneider

Super Angulon XL 5.6/58mm used has an image circle of 166mm and fits

therefore the original image circle plus the possible 25mm shift on that

camera easily. You can find the sample pictures at http:/www.alpavison.ch

(goto "alpaca" and there "shift examples"). Please note all those pictures are

made from the same point. Only the geometry of the camera changed with the

shift :-).

[todd peach seattle, washi]

<i>Tilts and swings control the plane of focus Aperture controls DOF.</i>

<p>

Bob is of course correct. The beauty of Scheimflug is that you can tilt the plane of focus so that it extends to the horizon and just happens to look like 'infinite DOF' (ala that highway shot).

[david_goldfarb]

As Bob says, tilt and swing can't increase depth of field--only stopping down will do that.

 

Tilt and swing ("swing" refers to horizontal tilt) can be used, however, to get more of the image in focus by tilting the plane of focus.

 

The Scheimpflug rule says that the lens plane, film plane, and the plane of focus all intersect in a line, except when they are parallel (the normal situation on a camera without movements). If you want to get a long flat surface, or perhaps a surface that is sloping away from you, all in focus you tilt the lens forward, and lay the plane of focus right on the surface you want to image.

 

Swings come in handy for things that slope horizontally, like an angled shoreline.

 

Now say that this surface is an uneven surface with little hills and small things sticking up from it. Then you stop down to increase DOF "thicken" the plane of focus into a kind of wedge of acceptible focus.

 

What if you have lots of intersecting planes in all directions? Tilt and swing won't help much then, because you don't really have a single image plane that could coincide with the plane of focus. In that situation, keep the camera square and stop down as necessary, just like you normally would.

 

Rise and shift? Think of those as a means for built in cropping. Say you want to photograph a tall building without the converging verticals that would occur if you were to point the camera up. You could set up the camera level, use a wide lens, and crop out the excess foreground, but the image would suffer, because you wouldn't be using the whole area of the film. A TS lens can project a larger image circle than a normal lens that covers the format exactly. By raising the lens, you can use the edge of the image circle instead of the center, get the tall building in the frame, and you shouldn't need any additional cropping. You are already selecting a portion of the total image projected by the lens.

 

Another common use for horizontal shift is to photograph a reflective object without the camera appearing in the image. To do this, set the camera to the right or left of the object, then shift to re-frame the image.

 

For a full explanation with illustrations, you might read Steve Simmons' _Using the View Camera_, which discusses lens movements for the view camera in a way that applies equally to TS lenses.

[squareframe]

Scheimpflug should be ignored. he meant well, and those that attempt to describe his principle mean well. it isn't any more difficult than saying that if I point my lens at a scene, the foreground elements will require a lens adjustment as compared to focusing on elements in the distance. I can turn the focusing ring on the lens to change the distance between my lens and the film plane, *or* I can tilt the relationship between lens and film, such that I can essentially have foreground and background in focus by judicious selection of the angle between lens and film. close objects will focus at the long end and distant objects at the short end. if we want to define a very narrow region to be in focus, and have everything surrounding it blurred as in Keith Carter's work, we invert the principle such that the plane works against focusing and everything but a narrow band is out of focus.

[squareframe]

photography should be intuitive. if you were able to focus the scene on a smoked or ground glass plate at the back of your camera, you would find that close objects focus farther away than objects in the distance. if you ignored Scheimpflug and other abstractions and rules, you would probably discover that you could focus both ends, foreground and background, at the same time by tilting your viewing surface. glass plate or film ... you have embraced the laws of physics, made them your servant, and performed magic.

[scott_fleming1]

Even I get it now. Kindasorta.

 

Hey, somebody who knows Michael Reichmann ... e-mail this thread to him and ask him to include it on his site in place of that lame page he has on movements now. (just kidding about the lame part Mr. Reichmann. I LOVE your site. AND You.)

 

Whoever owns this site ought to print threads out like this with proper attribution and publish them. I've been trying to figure out movements for months now ... without buying a whole book .... and this thread certainly got me far enough that I would have the confidence to play around with a T/S lense and figure it out.

 

Thanks all. You're a wonderful bunch.

 

Scott Fleming

[louis_webb]

Have a look at Bob Manekshaw's 'site.Pictures included!Find it via PhotoSites.co.uk

[james phillips]

Here is a quick layman's answer to the 'scheimflug' principle.

 

Think of looking through a funnel (your eye is at the tiny end) that is tipped up at 45 degrees aproximately. The area of sharpest focus (that which you see through the funnel) is now straight out from you and continues up into the sky. Just like the growing angle of the funnel as it expands from your eye to the horizon.

 

"Tilt" the funnel down towards the pavement a tiny bit and now the area of sharpest focus starts to move down towards the ground. The more you tilt (to a sensible degree), the more the area of sharpest focus starts closer to the ground in front of you and moves up towards the horizon.

 

The trade-off is that the higher part in the sky now starts to become "less sharply" in focus. (poor english grammer.. I know)

 

Now you can use the same principle for swing. (left or right) Swing the lens a bit to the right while focusing on that long fence that starts at your right shoulder and goes off into the horizon.

 

As you swing more to the right (once again to a certain degree) the more the fence closest to you becomes sharper in focus. Trade-off is that eventually you will notice the fence in the far distance becoming less sharp.

 

That is a layman's explanation.

[q.g._de_bakker]

Let me try too.

 

There's no mystery in the Scheimpflug principle. Remember Newton's lens formula? 1/f = 1/object distance + 1/image distance?

 

The focal length of a lens is fixed (even with zoom lenses it's difficult to use all available focal lengths at once. However, if you do, Scheimpflug (like focus in general) will be difficult to achieve).

So what remain as linked (!) variables are the conjugate distances: for every distance from lens to subject, there is only one lens to film distance at which a sharp image is formed.

 

Now what Scheimpflug noticed was that there is no law saying that the conjugate distances should be the same across the entire image.

In real life, lens to subject distances vary a lot. So why not let lens to film distance vary too?

 

If the bottom part of a picture is filled with nearby objects (nearby = smaller lens to subject distance = larger lens to film distance), leading to increasingly more distant objects (more distant = larger lens to subject distance = smaller lens to film distance) going up, you can get all of them in focus if you make sure the increase in lens to subject distance is countered by an equal decrease in lens to film distance.

 

Of course it only works when the increase in subject distances is gradual, linear even. Unless we can bend the film 'plane' to any shape we want.

Now there's a thought...

[q.g._de_bakker]

Uhmmm...

 

When i wrote: "you can get all of them in focus if you make sure the increase in lens to subject distance is countered by an equal decrease in lens to film distance", the "equal" should be taken to mean according to the fixed, Newtonian relation between subject and image distances.

[dick roadnight cotswolds]

Theory is OK - but how do you set the camera?

 

We who have professional cameras appreciate scales, as they allow us to calculate settings.

 

Scheimplug can seem complicated, and a Sinar monorail is supposed to do it all for you without knowing about it.

 

But even with a Sinar, it helps to know where to start, and you can estimate without scales.

 

The angle between the standards is the angle whose tangent is the

extension divided by the offset,

 

where

the extension is the distance from the film to lens board or lens node

 

and the offset is the distance from the lens axis to the plane of sharpest focus.

 

But the angle (in degrees) is about 50 times this ratio.

So for 150mm extension and 1m offset you get (150/1000) x 50 = 7.5degrees.

(the right answer is about 8.6 degrees)

 

But you can make it simpler:

 

Take the extension (e.g. 150mm)

 

Divide by ten (15)

 

Divide by 2 (7.5 degrees)

 

Divide by the meters of offset

 

Job done.