Karsten Bruun Qvist
Introduction
Shadows are important, perhaps especially in black and white photography where they often provide as much, or even more impact than the physical subject. Here, we will explore a way to make striking photographs of shadows by giving them 3D properties. Doing so, we will be photographing shadows that have never existed, nor could ever exist, except perhaps in some surrealist parallel universe. Although we will use tools from stereo photography this is not about rendering the known world in 3D, it is about giving volume and life to shadows, while the rest of the image remains flat.
.. we will be photographing shadows that have never existed, nor could ever exist
Let us start by looking at an example, first as a normal photograph, a wire mesh basket casting a shadow on the surface it is placed on:
Now, let us take look at the 3D shadow version, here presented in the so-called universal view format with an anaglyph on top (requiring red-cyan glasses for viewing). The bottom row consists of a left eye image, a right eye image, and finally one more left eye image, providing opportunity for both parallel and cross-eye free viewing.
As always with anaglyphs, it may show up on your screen with ghosting that was not visible on the screen used to prepare it. So, if the anaglyph does not work for you, try the free-view versions below it. Also, be prepared to give yourself a bit more time than you normally need with a stereo image; after all you are asking your brain to allow you to see something that does not exist. Hopefully you will then begin to see a shadow coming out towards you, living freely in 3D space.
Obviously, this is not normal shadow behavior; something is seriously wrong here. In the real world, shadows are always bound to the surface on which they are projected. That surface may be planar, or curved, causing the shadow to be planar or curved, but the shadow never exists as a volume, free in space, unbound of any physical surface! Nevertheless, you can make photographs that makes it appear that way…
The Method
The method is in principle quite simple, and it is not my invention. I came across it in a presentation by George Themelis in which he credits photographer Hans Knuchel for the idea (in his book STEREO that I have not been able to access). Instead of moving the camera between two exposures, as you might do in stereo photography, you move the light source that casts the shadow. Then you assemble the 'left' and 'right' photos in a suitable piece of software, such as Masuji Suto's StereoPhoto Maker.
When starting to experiment, my first attempt was to use a Godox constant LED light (SL200 II) as light source, together with Masuji Suto's smartphone app 3DSteroid. The idea was that a constant light would make it easy to see where the shadow falls, and that the app, by giving you an overlay of the first shot, when you prepare to take the second, would help determine how much to move the light between the shots. The practical challenge with this approach is that a constant light is heavy, and therefore difficult to move in a precisely controlled way.
Instead of moving the camera, you move the light source
I therefore turned to a flash setup, with a speedlight (Godox TT685F) first mounted to a small ball head, then to a 25 cm slider with a built-in ruler, and finally to a standard light stand. For the distances and object sizes used below, simply moving the flash along the slider provided all the mobility needed. In most cases, six or seven exposures were made with the flash at different positions, and the pair providing the most suitable shadow separation was chosen for processing in StereoPhoto Maker.
While shooting any given object the camera (Fujifilm X-T3) remained in a fixed position on a tripod. A radio trigger (Godox XProF) was used to control the speedlight, and exposure settings were chosen such that ambient light only contributed marginally to the exposure.
Other examples
In the next example, it is hopefully possible to see that the shadow of the temples of the glasses makes an unexpected curve through space.
Be prepared to give yourself a bit more time than you normally need with a stereo image; after all you are asking your brain to allow you see something that does not exist
In the eggbeater example, the effect is a bit more subtle, but it should be possible to see that the various parts of the shadow are not all in one plane.
In the chair example, a more complex scene we see the shadow of the seat and back rest floating in midair.
Obviously, this is not normal shadow behavior; something is seriously wrong here
Reflections and suggestions
Although the basic principle may be simple, that does not guarantee good results with every attempt. Below are some observations and reflections on what seemed to work best under my conditions. Perhaps that can help you get started if you decide to experiment yourself, but please do not consider any of the following to be the final word on how to photograph impossible 3D shadows.
There are a number of things to consider when positioning light source and camera relative to the object to be photographed. You may want to de-clutter the scene as much as possible, so that object and shadow stands out. Having said that, it may not always be ideal to have the shadow fall on a completely uniform surface; a surface with a bit of texture may help sell the illusion that the shadow exists free in space.
You will also want to consider the angle from which the light hits the object, as it determines the length of the shadow, and thereby the amount of 'depth' you have in the scene. In my tests, shadows coming towards the camera usually created more interesting results than shadows pointing away. Shadows with little detail, that just appear as blocks of black do not give very interesting results.
If you need to increase the sharpness of the shadow, you may need to increase the distance between light source and object and therefore also the magnitude of the displacement of the flash between the shots. If the light source is close to the subject, one can inadvertently create a situation where the light is sufficiently closer to the subject in one shot than it is in the other, causing creation of shadows of different sizes. This will result in alignment problems that are hard to resolve in postprocessing.
When positioning the camera relative to object and light source, it is helpful to prevent the camera from 'seeing' any specular reflections from the surface the object is placed on. Such reflections are usually not pleasing, and when combining two photos where they are in slightly different positions it creates a disturbance in the '3D' image. To avoid specular reflections, move the camera off the axis defined by light source and object, and/or make sure that the camera is seeing the scene from an angle that is different from the angle of incidence of the light source with respect to the object. Combining both will provide the best insurance against specular, blown out reflections.
If you try this technique, you are most welcome to share your experience in the comments section
When processing image pairs in StereoPhoto Maker it definitely is worthwhile to check the effect of swapping the two images, as one arrangement is usually much more interesting that the other.
There is a lot one can experiment with, using this technique, such as objects, scenes, the angles the light source and camera are pointed at, distance between light source and object, how the light source is moved in a controlled way, etc. Moving both the light and the camera is certainly also a possibility, as is using the Sun as light source, and simply allowing a few minutes between the shots. If you try this technique, you are most welcome to share your experiences in the comments section below.
To be honest, the most successful use of the method I have seen still is the image that inspired me, a photo by Hans Knuchel, shown in George Themelis' presentation. Below it is shown in a reprocessed version, to match the format of the previous photos. The amount of depth is insane, so allow you brain a moment to take it in, and use the free-view versions if the anaglyph proves too ghostly for your viewing circumstances.
I am still trying to understand how Hans Knuchel got away with creating this image, so if you have some good ideas, please let me know...
References and helpful links
George Themelis: Flash in stereo. Presentation at NSA 2010. http://www.drt3d.com/W03-Flash.pdf
Masuji Suto: StereoPhoto Maker (software for stereo photography). https://stereo.jpn.org/eng/stphmkr/
Masuji Suto: 3DSteroid (app for Android, equivalent for iOS is i3Dsteroid)
Rebecca Sharpe: How to View Stereoscopic 3-D Images. https://stereoscopy.blog/2019/08/04/how-to-view-stereoscopic-3-d-images-basic-tutorial/