First shift (rise) lens, could use some tips

[loccdor]

I've purchased an Olympus OM 35mm f2.8 Zuiko Shift Lens to adapt to both a full frame 35mm camera and a half frame (1.44x factor). I'm interested in taking pictures with parallel verticals. Having some trouble visualizing the geometry/trigonometry. I've never used a shift lens but I have tried a little freelensing.

Questions:

1) What's the calculation to how much rise I need (in mm) for parallel verticals, based on how far I am away from the building and how tall it is? Or maybe better expressed in the vertical angle.

2) How will the shift qualities of the lens change when full frame and half frame are compared?

3) Does a shift lens need a wider lens hood compared a non-shift lens of the same focal length?

4) Are there differences in a how a polarizer or other filters would work when shifted?

 

[1957ExaktaVarexIIa]

Interesting questions!
I must admit that I have not yet thought about these aspects when I use the Canon FD TS 35mm f/2.8 .

As for your first question: I suppose it might be an appropriate method - for example if you take a picture of a house in front of you:

• first to make sure when setting up your camera that your film and the house's walls are exactly parallel. You will notice that the distortion disappears;
• then move up the lens's shift mechanism until you get the whole house on your picture (or the house as you like it).

But: I suppose it's not always wise to eliminate distortion completely. If you look at such a picture you might get the impression that the building starts to fall over you. I've often made that mistake.
So it might be a good idea to leave the building a very little bit distorted...

 

[ic-racer]

You shouldn’t have to do any math .
Shift lets you frame the image the way you want so adjust the shift up down or side to side until the images framed the way you want

 

[ic-racer]

Likewise, tilt lets you adjust the plane of focus to match the plane of your subject so again looking through the camera finder, you would adjust the tilt until the plane of focus matches the plane of your subject. Again, no math is needed.

 

[reddesert]

You can think of a shift lens as a somewhat wider-angle lens that gives you some freedom to control where to place the image, within the wider field of view. (What it actually is, is a 35mm focal length lens that covers more than the 24x36mm image area; it's sort of like a wide lens for 6x4.5 but you have to choose which 24x36mm area of the image to put on film.) This may help to find answers to some of your questions, for ex the hood and polarizer.

In order to get verticals parallel, you need to level the camera back, so you need a bubble level and a tripod will be very helpful.

1. You could calculate the rise needed based on angle of view, but I'm not sure it will better than just leveling the camera and looking through the viewfinder.

2. With half frame obviously you don't get as wide a field, and you can use as much shift as you want without getting to the edge of sharpness (I'm not familiar with what Olympus says about the edge of the image circle for this lens).

3. You can anticipate needing a wider hood, if you used a hood that was exactly right for 35mm, it would vignette when shifted. Or do like a LF photographer and shade the lens by holding a darkslide.

4. It's like a very wide angle lens. So the way that polarizers tend to darken the corners in wide-angle shots will be exaggerated. Most other filters probably not affected.

 

[loccdor]

Thanks! I had a feeling I was overcomplicating this for myself. Just a medium format ultrawide where you decide the crop. Easy enough when thought about that way.

 

[Taylor Nankervis]

No mathematical gymnastics is needed for implementing shift in a shift lens. However, if one is using a tilt and shift lens (combined) like Canon's bespoke TS-E lenses, there is some initial math (optionally) involved (sine angle converted to degrees [of tilt] ) if (tilt) alignment is not executed visually.

Generally a shift-only lens has a very limited use other than the points outlined in the foregoing posts. Run-out of shift when photographing e.g. buildings is common; then the user tilts the camera back to try and correct it, only to introduce convergence and thus effectively negates any advantage. There are small yet undeniable applications for shift lenses including off-setting things like annoying street 'furniture' when physically moving the camera is not desired, or for shifting one's reflection out of a mirror for example. Yet more advanced tricks can be carried out with multiple exposures for those cameras that allow it.

Since a shift (or tilt-shift) lens 'bends' the path of light hitting the camera's meter, metering errors will occur unless you use AE-lock e.g. compose, apply shift then zero. The procedure is meter, lock in the exposure and re-apply the shift desired and trip the shutter.

 

[wiltw]

I've purchased an Olympus OM 35mm f2.8 Zuiko Shift Lens to adapt to both a full frame 35mm camera and a half frame (1.44x factor). I'm interested in taking pictures with parallel verticals. Having some trouble visualizing the geometry/trigonometry. I've never used a shift lens but I have tried a little freelensing.

Questions:

1) What's the calculation to how much rise I need (in mm) for parallel verticals, based on how far I am away from the building and how tall it is? Or maybe better expressed in the vertical angle.

2) How will the shift qualities of the lens change when full frame and half frame are compared?

3) Does a shift lens need a wider lens hood compared a non-shift lens of the same focal length?

4) Are there differences in a how a polarizer or other filters would work when shifted?

Responses covered your quesitons, and I willl raise a different issue...
Generally speaking with using a camera with TTL metering with shift lenses, you meter the scene UNshifted, set the exposure, then shift the lens and take the photo.

I cannot tell you if that technique is avoided with any particular design of TTL meter, but I speculate that it would be unnecessary with a mirrorless SLR that meters via the imaging sensor itself rather than some photosensor elsewhere in the body.

 

[Besk]

Use the same hood. The field of view of the lens does not change when the lens is shifted. A 35mm lens is still a 35mm lens when shifted.

 

[MattKing]

Generally speaking with using a camera with TTL metering with shift lenses, you meter the scene UNshifted, set the exposure, then shift the lens and take the photo.

If the camera is an OM model that employs OTF (off the film) metering in Auto mode, you may find that you can use that to meter as you expose the film. I would be likely to do both - follow @wiltw 's advice first, then switch it into Auto mode and take another shot to back up the first.

 

[loccdor]

I was having some trouble figuring out how the vertical angle of the photograph (portrait orientation) would change when comparing an optically centered shot to a shifted one where the optical center is on the bottom edge of the frame. From my calculations with a building about 20 feet away where you're trying to fill the frame with it, if the vertical angle was 37 degrees centered, it would be 34 degrees when heavily shifted. Requiring the user to step back slightly, but not a huge impact.

 

[Ian C]

I presume that you’re familiar with the purpose of a shift-axis lens. It’s really the same as using view camara with a lens that projects a sufficiently generous image circle so that you can keep the lens axis as close to horizontal as possible and still be able to record the top of a building, tall trees, and so forth while keeping the verticals parallel for a natural looking photo.

Compose and focus. Then shift the lens axis upward to capture the top.

Here is the manual for the lens.

https://lens-club.ru/public/files/pdfs/674118ca30939f775cf7b275c019bfe4.pdf

 

[loccdor]

Thank you for the manual. I understand what it's intended for... my main question was about the angle that could be captured when a non-shifted and heavily-shifted lens were compared. It turns out the heavily-shifted lens captures a smaller angle and requires stepping back slightly from the building, but it's a small enough effect that it probably doesn't have much practical significance.

It confused me because they say that "angle of view" does not change when you shift the lens, but it's only because of the way they're defining angle of view.

To get what I am saying, look at the picture I made. As the subject (the long vertical line) shifts up toward infinity, the angle made between the lens and the subject shrinks toward zero. In practice, you can't shift the lens enough for this to matter much. You can only shift it about as much as figure 2.

 

[wiltw]

Thank you for the manual. I understand what it's intended for... my main question was about the angle that could be captured when a non-shifted and heavily-shifted lens were compared. It turns out the heavily-shifted lens captures a smaller angle and requires stepping back slightly from the building, but it's a small enough effect that it probably doesn't have much practical significance.

It confused me because they say that "angle of view" does not change when you shift the lens, but it's only because of the way they're defining angle of view.

To get what I am saying, look at the picture I made. As the subject (the long vertical line) shifts up toward infinity, the angle made between the lens and the subject shrinks toward zero. In practice, you can't shift the lens enough for this to matter much. You can only shift it about as much as figure 2.

View attachment 407545

The Angle of View is determined by the Frame Dimensions, and so is never changed (when the lens is not changed)


You need not have a mathematical comprehension (and abtility to compute) the situaion, although an understanding can be beneficial in understanding which direction to shift or tilt, or the magnitude of the shift/tilt. But not necessary in general usage of the feature. 'Nice', but not 'necessary'

 

[loccdor]

The Angle of View is determined by the Frame Dimensions, and so is never changed (when the lens is not changed)

That makes sense, but the way I intuitively think about angle of view is an angle going from my eyes (or a lens) out into the world. The size of this angle does change when you shift the lens.

 

[wiltw]

That makes sense, but the way I intuitively think about angle of view is an angle going from my eyes (or a lens) out into the world. The size of this angle does change when you shift the lens.

In the case of a shift lens, you merely are altering the centering of the large image circle as you shift it, thereby causing the 24x36mm window at the focal plane to capture a different view (but fixed 'Angle of View') captured with that window. In the case of 24mm shift lens, its image circle might hypothetetically encompass a 100 degree area within its image circle, but it is always projecting a 24mm lens' 84 degree angle of view within the 24x36mm window at the focal plane.

 

[loccdor]

One other rule of thumb I'm coming across as I work through this:

Let's say your film height is 24mm. You walk toward the building with your unshifted lens and look through the viewfinder, camera parallel to the building. If the top of your image is at the halfway height mark of the building, then you will need to shift your lens up 12mm (read: 1/2 film height) for the top of your image to be at the top of your building.

(In practice you need to shift slightly less because your camera is already at some small height above the ground, but then slightly more because you want to include a bit of sky above the roof line)


I worked on some math with the help of our friendly computer so I could see how the vertical angle coming from the lens would change as the lens was shifted. This is for a 35mm lens on half-frame in the normal portrait orientation, or on full frame in the landscape orientation (since they're both 24mm). There's hardly any change in angle at all until you approach 1/2 film height.

Sensor Shift (s)	vFOV (°)
0 mm (centered)	37.00°
4 mm	36.63°
8 mm	35.41°
12 mm	33.69°

The effect on vFoV gets quite a bit worse when you're using an ultrawide shift lens though like the Laowa 15mm that recently came out. That one has vFOV of 77 degrees unshifted, and 58 degrees with a 12mm shift, losing 25% of its angle.


The formula:

 

[xkaes]

As far as I can tell, this lens has Shift and Rise/Fall -- about 1/2". I don't know why Tilt has been brought up in some responses.

The angle of view is for a 35mm lens about 63°, but the image circle is much larger for a Shift lens than a normal 35mm lens. That's one of the reasons it's so much more expensive. Any 35mm lens will have the same angle of view, but the image circle can be small or large.

As for a lens hood, I'm sure Olympus designed one specifically for it. If you want a lens hood -- good idea -- use that one and nothing else. For example, Minolta also made a 35mm shift lens -- but with a special hood that is much wider than the one for their normal 35mm lenses. When you get out to the edges of the image circle a lens hood can cut off the corners if not designed for the lens. Minollta took an unusual approach with its shift lens in that it has TWO "filter rings". The innermost is for standard filter use, while the outer ring is for the extra-wide lens shade.