Fake Miniature Photography

I'm curious about the optical/viewing principle behind those - I understand the technical part (high angle, tilt or PS, etc.), but I'm not sure what makes them look like models in our minds. When I look at something from the roof of a building (or, say, a model train track), I'm not seeing a narrow band of focus, but the same deep depth of field as always.
 
FOr me, When done well, it's a small area of focus with a small transition too heavy out of focus. Looking at the below shot though you can see how it's badly done and takes you out of the macro-like shot because the bottom of the tower goes out of focus while the tower itself is in focus.

90-tiltshift-church-moilogo_210907.jpg


I'm curious about the optical/viewing principle behind those - I understand the technical part (high angle, tilt or PS, etc.), but I'm not sure what makes them look like models in our minds. When I look at something from the roof of a building (or, say, a model train track), I'm not seeing a narrow band of focus, but the same deep depth of field as always.
 
When this was new (to me), I did took many shots like that. First with a selfmade lens-tilt-adapter, then with a better one for the Olympus Pen.

Its not only the tilted DOF, what makes the illusion, but also the color post processing. I learned, that oversaturated and "vintage" colors help much to make the illusion better.

It's a nice toy for some time, but I lost interest a little bit.
 
I'm curious about the optical/viewing principle behind those - I understand the technical part (high angle, tilt or PS, etc.), but I'm not sure what makes them look like models in our minds. When I look at something from the roof of a building (or, say, a model train track), I'm not seeing a narrow band of focus, but the same deep depth of field as always.

To see a real scene with such shallow depth of focus you need to be up close. Very close. So, the focus effect creates the optical illusion that you are looking at the scene from a few inches away. Hence your mind interprets the image as a close up view of tiny features.
 
I'm curious about the optical/viewing principle behind those - I understand the technical part (high angle, tilt or PS, etc.), but I'm not sure what makes them look like models in our minds. When I look at something from the roof of a building (or, say, a model train track), I'm not seeing a narrow band of focus, but the same deep depth of field as always.
It's not that you want to recreate what a model looks like but what a photo of a model looks like. The look you see in model railroad magazines and catalogs, for example.

Personally, I think it's a stupid fad. It annoys me much more than HDR does, in fact.
 
Not to be ignorant, but is this a photoshop thing, or optical?

How can I get such an incredibly shallow DOF with the constraints of the fastest shutter speed I can use? Need to use a filter so I can open the lens all the way?

This is not something I want to pursue, but I have seen this effect a number of times now (David Burnett has a nice example) and it IS striking.

Randy
 
Not to be ignorant, but is this a photoshop thing, or optical?

How can I get such an incredibly shallow DOF with the constraints of the fastest shutter speed I can use? Need to use a filter so I can open the lens all the way?

This is not something I want to pursue, but I have seen this effect a number of times now (David Burnett has a nice example) and it IS striking.

Originally it was optical - you use a camera or lens with tilt, and tilt against (rather than with) the convergent planes. A fast (or long) lens does help keeping the focus depth low, too. David Burnett (who first exploited it for press photography) utilized a Speed Graphic with Aero Ektar to combine tilt and a fast lens. When Canon and Nikon came up with shift/tilt lenses for small format DSLRs it turned very widespread - and now that there are photoshop plugins that do gradual blurs it is rather pointless to do it at all, as every Joe Average photo rag user is miniaturizing his holiday snapshots.
 
The technique doesn't have to be used for faking miniatures. I sometimes use it to fake the use of a wider aperture in photos. I modify the usual technique by masking off what I want in focus using quick mask and a brush. Then I use a minor amount of lens blur on everything else. Next I use quick mask and a brush to expand the masked area "farther" into the photo and then apply another slight amount of lens blur. I repeat these steps until I'm satisfied with the amount of focus falling off into the background.

Here is an example where a slow lens, short focal length and a distracting background didn't work for me. The hangers in the photo were fairly sharp. Note the focus falling off on the closest hanger, yet the left wing hasn't been altered. All of that hanger was pretty sharp in the original.

6163781800_856666c948_z.jpg
 
Last edited:
Well, as an average joe photog who recently miniaturized some of my holiday shots, I'll defend it as a fun, if admittedly lightweight, fad. I use it when I'm feeling playful and when I'm trying to emphasize the depth of a scene. I posted some of these on the Ricoh GXR M module thread just a few hours ago. I like those shots, I have no illusions that I am looking at a model nor that I am making great art, but they capture the slight surreality I was seeing in Tokyo.

The Ricoh GXR has an interesting implementation of this that is really much better than Olympus': you can control both the position and width of the in-focus area. Very fun!
 
Didn't know I needed to ask your permission first. Sorry, my bad. Please don't forget to go after sevo, too.

Thanks for the heads up.

Yes, sevo has pointed out an easy way to do this in photoshop. As everyone on the forum will agree, if anything can be done on a PC with Photoshop, then there's no point in doing it using "traditional" methods (optical, chemical, analog).
 
So compute a spatial gradient to determine the in-sharp points of focus, set a threshold, and then set a Gaussian blur proportional to the gradient for everything below that threshold?

I think my wife did that in FORTRAN in the 1980s.
 
Try to calculate that gradient in a picture with many vertical structures, or a fence or anything like that. Photoshop or any other programm fails on such scenes (yet).
 
Back
Top