Tilt-Shift...destined to be a classic or just another fad?

[DiskoJoe]

I think maybe the question is whether the "tilt-shift effect" which is usually done in post is going to be a fad or not. I think's already been a fad and is done, mostly. It's a big internet, and there's always someone who has just discovered a thing, but I think this one has had it's peak a long time ago.

Real tilt-shift with equipment that actually wiggles is pretty expensive and specialized, so you're typically NOT seeing actual T/S kit being used for cheap effects. Usually the people who bother with that kind of kit are solving actual problems of managing the plane of focus for a purpose, rather than for an effect.

Exactly. I know lots of landscape photographers who do amazing work with tilt shift lenses and not the cheesy tilt shift effect that most people think about. When used properly the image wont have that cheesy blur because you would stitching multiple images together like a pano or sorts.

My buddy here is a good example for what Im referring to....

Flickr: Kevin Dickert's Photostream

 

[Murray Bloom]

Here are a couple of Tilt/Shift/Swing examples. In the first, I laid the plane of focus back and turned it so that the hand, watch and pin, the grenade and model's face were all in sharp focus. The vertical depth of field was limited, as you can see (look how soft the necklace is), but that didn't matter much to me for this shot:





In this case, I tilted and turned the plane of focus to get the entire camera and lens in focus. This was my first zoom lens, and also the first zoom that Nikon ever made, 85-250mm. From front to back, the distance was about sixteen inches, with the lens hood only a few inches from the taking lens.





Both images were shot with a Linhof Super Technika 4x5 camera. I lost the neg from the Nikon shot over the years, so what you see is the result of scanning an 11x14 print in two sections and then stitching them together.

I agree that you often see images in which these sorts of optical acrobatics are used without ever noticing. As for whether T/S is a fad, who knows. It's probably a lot like bad tonemapping or Lensbabies for a lot of people, just something they think will make their images look distinctive, where in reality it just makes them look like others using it for a gimmick. Realistically, I can now make geometric adjustments better in Photoshop than I could with a view camera or with a shifty architectural lens. But if you want to manipulate depth of field and know how to use them, flexible lenses are about the only way.

 

[Ilovemycam]

I think maybe the question is whether the "tilt-shift effect" which is usually done in post is going to be a fad or not. I think's already been a fad and is done, mostly. It's a big internet, and there's always someone who has just discovered a thing, but I think this one has had it's peak a long time ago.

Real tilt-shift with equipment that actually wiggles is pretty expensive and specialized, so you're typically NOT seeing actual T/S kit being used for cheap effects. Usually the people who bother with that kind of kit are solving actual problems of managing the plane of focus for a purpose, rather than for an effect.

Exactly. I know lots of landscape photographers who do amazing work with tilt shift lenses and not the cheesy tilt shift effect that most people think about. When used properly the image wont have that cheesy blur because you would stitching multiple images together like a pano or sorts.

My buddy here is a good example for what Im referring to....

Flickr: Kevin Dickert's Photostream

Yes, he has some nice work. But my question was along the lines of the 'cheesy blur' aspects of TS.

 

[Buckster]

But if you want to manipulate depth of field and know how to use them, flexible lenses are about the only way.

Two words: Focus stacking.

 

[amolitor]

Also three words: Light Field Camera.

all three options do somewhat different things, but all allow a good degree of control over what is and is not in focus.

 

[Murray Bloom]

Yes, focus stacking is a possibility, but I've seen very few that looked natural enough for me. If I can see the process, I'll usually pass on the image.

As for light field cameras, I'll admit to not staying up to the minute on that technology. Last I saw, it was an either/or technology where you had to select one focus point, not enough of them to mimic extended depth of field. Also, image resolution was a problem. Has the technology improved?

 

[Gavjenks]

The OP is confusing "tilt shift" with "one specific thing that people do with tilt shift sometimes to make scenes look like toys."

That is not the main reason why tilt shift lenses were designed or are manufactured. Their main purposes are to correct perspective (without losing resolution by stretching pixels in post processing); create easy seamless panoramas without special software; to selectively focus on one thing and not another, even if they are equal distances from the camera; and to give yourself vastly deeper depth of field in landscapes, cityscapes, whatever, by tilting the plane of focus to line up with the dominant line of objects in the world, thus giving yourself an f/32 amount of depth of field even though you're shooting at f/8, etc.

The toy people fad is obviously short lived. The other abilities, however, which simply add technical flexibility to your basic camera settings will not ever go out of favor, no more than large apertures will go out of favor.

 

[Gavjenks]

Software substitutes for any of these things will always be significantly limited versus the optical solution. For examples:

Focus stacking does not work well if your subject is moving (leaves rustling or plants entirely bending with significant wind, or a bug on the move will cause your layers to not line up). Focus stacking also does not work at all if you are shooting VIDEO (tons of great video effects are possible with tilt shift focal plane manipulation).

Light field cameras are in their infancy and are not yet commercially available in anything larger than a couple of megapixels. Even later on, they will suffer from maximum limits on resolution and data transfer speed that will always put them a step behind traditional imaging devices along those parameters (an equivalent megapixel image will always require dozens of times more data to transfer, and there are physical limits in how closely you can pack redundant light field sensors compared to single plane sensor pixels). These issues are also magnified by video.

And blurring things in photoshop will in many cases look vastly and noticeably different than actual optical blur (mostly when there are lots of tiny point sources of light and interesting bokeh, in particular), at least with current technology. It also simply takes a really long time to edit an image to have blurriness that corresponds to distance from the focal plane in the real world, since you have to hand-annotate each object with its real world distance. For example, if you take a photo from on top of a parking garage, a street lamp that is half a block away does not have uniform distance from you. Optically, the top of the lamp would probably be in sharper focus than the bottom of the lamp, as well as in sharper focus than the stuff surrounding the top of the lamp in the image. Simply blurring a whole half of the image in photoshop won't look as good; instead, if you want realism, you'd have to selectively blur the top of the lamp less than the bottom or its surroundings if you're simulating a narrow DOF or very selective tilt. And so on for every object in the scene. There aren't any good software solutions to give you a depth map of the world, either (it's a matter of mathematical ambiguity, not software shortcomings), so no easy shortcuts here. And once again, this becomes almost impossible to do with video in any cost effective manner.

If you look closely, I find that about 1/4 to 1/3 of all documentary movies or shows I watch tend to use tilt shift lenses to some degree for their classical purposes. Something that would be vastly more expensive or impossible to simulate with software with anything less than a full length feature film blockbuster budget.

For example, I saw a show just yesterday where they were filming the host in front of a cathedral in a major European city. There were retail businesses across the street that detracted from the feel of it, and that they didn't want to show the names of on film. Instead of compromising with a less effective shot angle or using gaudy blur bubbles, they instead opted to use a tilt shift lens to selectively blur out that side of the street (which was at a diagonal and required a lot of swing tilt), while keeping the church and the host in good focus. Very tasteful and non distracting, and as a bonus, lent a very professional feel to the shot, since people often to associate creamily blurred backgrounds with expensive cameras and high production quality.