A Beginner's Guide to Stereoscopic Photography on a Smartphone: Basics

by Ian Hellström (9 October 2018)

3D and VR may sound like recent inventions, but the former has been around since the mid-1800s. Stereoscopic cards were extremely popular in Paris in the 1860s (e.g. Les Diableries). In a series of posts I’ll focus on how stereoscopic imaging works, how to create stereoscopic images with any camera or smartphone, and what mistakes beginners tend to make and how to avoid them.

If you look for vintage stereoscopic cards, you’ll find plenty of antiques and reproductions of 3D images of cities, portraits, and of course pin-ups. It would not have been popular without some nudity, now would it? The recent resurgence in stereoscopic imaging has mostly come because of virtual and augmented reality.

How does it work?

Humans see three dimensions because we have two eyes that are a distance apart; the interpupillary distance lies between 50 mm and 75 mm. Each eye sees the world slightly differently, which the brain uses to build a three-dimensional representation of the world. That’s why if you close one eye, your perception of depth is impaired. Movement and shadows can be used to figure out whether one object is in front or behind another though, but our binocular vision is designed for seeing perspective.

If you shoot a photograph in the same way (i.e. two images slightly apart) you can recreate that effect with two side-by-side two-dimensional pictures. Each eye is presented with a single picture of the same scene, and the brain does the magic of fusing the two images into one. Ever since Brewster’s lenticular stereoscope, most viewers have used lenses to focus each image.

If you have ever looked at stereoscopic images with a viewer you may have noticed the effect: it may take a few moments before your brain manages to meld the images into one, and until you ‘get it’, the images are blurry, flat, and partially overlaid.

What about movies and televisions?

Films use a slightly different technique: the scenes are captured with two lenses side by side, and then displayed at the same time on a single screen. To create the effect of two side-by-side images, either colour or polarization filters are used.

With colour filters, the scenes are recorded with filters on the lenses while shooting the movie. The projectors’ images are then superimposed on the cinema screen. The cinema goers have to wear red/blue (anaglyph) glasses, so that the images from the ‘red’ lens are perceived best by one eye, and the images from the ‘blue’ lens are seen best by the other eye. Both eyes see roughly the same contours, but red looks white and blue black through the red glass, and vice versa for the blue one. The disadvantage is obvious: the usage of colour is limited.

The alternative is to record with two lenses but later project the reels with different polarization filters that match the goggles everyone wears: the left eye sees only images with the matching polarization and likewise for the right eye. Again, the brain does the rest.

In recent years there have been 3D TVs and of course the Nintendo 3DS, some of which do not require special glasses. These technologies rely on autostereoscopy, which means the display projecting the images includes optics to split the images into the viewer’s eyes. The Nintendo 3DS (and the Fujifilm FinePix REAL 3D’s LCD display) uses the parallax barrier for that effect: an arrangement of slits, the so-called parallax barrier, sits on top of the display and it ensures that each eye sees a different set of pixels. The illusion of depth is due to what is known as parallax; parallax is the effect that objects in the distance appear to move less than objects in the foreground when you change your position. A similar autostereoscopic display can be achieved with lenticular lenses: an array of lenses that magnify different images when viewed from slightly different angles.

Because the slits and lenses are fixed, the perception of depth is only possible when viewed from a certain distance and angle, typically only when looking straight at the display from a short distance. For hand-held devices that’s acceptable, but not so much for televisions.

Note that it’s also possible to have active shutter glasses: the display sends the left and right frames seqentially and the glasses are synchronized to only let one eye see the current image. This is done at a very high refresh rate, so that it looks as if there is a continuous stream of pictures coming to each eye, even though the eyes’ images alternate. This is what some 3D TVs use.

What do you need?

To view stereoscopic images you need a stereoscope or viewer. You can of course build your own but there is no shame in buying one. Probably the most famous viewer is the View-Master. I personally like the OWL VR Kit from the London Stereoscopic Company, but you can also opt for Google’s Cardboard viewer, or something along those lines. Note that you do not need an expensive VR set to view side-by-side stereoscopic images.

If you want to take pictures, you also need a camera. Any smartphone camera will do, and I’ll explain below how it works and what apps you can use for it. You do not need an app, but especially for beginners an app can assist and make the experience a bit more pleasurable.

That’s all. You do not need a special stereo camera. The most famous one is the Stereo Realist, which is the most popular 35 mm camera of all time. The most recent semi-professional digital stereo camera is the FinePix REAL 3D W3, which is nearing its tenth anniversary. While there are newer digital stereo cameras on the market (e.g. TwoEyes VR, SID 3D or Jedeye), they typically lack a view finder, optical zoom, flash, or even a display with a live preview. In my opinion, that makes these cameras less suitable for stereoscopic imaging and they are also not exactly cheap, if you consider the basic functionality they lack. Note that it’s also possible to ‘hack’ a stereoscopic camera by placing two identical cameras next to each other. And of course you can also use mirrors to achieve a similar effect with a single camera lens. I’ll talk more about the last option in a future post.

What apps are there?

At this point you may wonder how you take stereo images with any smartphone, especially when considering that most phones only have one camera on each side. The obvious way to do that is by having two lenses, but few phones do. Those that do have two cameras have a very small baseline, that is, distance between the lenses, so that the depth perception of the human vision is not reproduced accurately. This is usually good enough for selfies (and close-ups) though.

The trick is in realizing that you can also achieve the same desired effect by taking pictures sequentially but from slightly different angles. That way, you only need a single camera and some software to place the images next to each other.

On my iPhone are several apps for stereos: Camera 3D Pro, i3DSteroid, and 3D Photo. I have tried others but I usually stick to Camera 3D Pro or i3DSteroid.

Nicola Vacca’s Camera 3D Pro is my go-to app: you choose the distance to the closest object in the image, tap on the camera shutter button on the screen and slide the camera across your hand until it reads ‘release’, which depends on how far away objects in the scene are: the closer, the smaller the distance to slide the phone.

The apps sports several export options and it has its own (independent) storage for stereoscopic images, which means they do not need to clog up your camera roll unless you decide to export side-by-side or anaglyph pictures. Most apps require saving or exporting stereo images before you can take another picture, which is inconvenient, especially when experimenting with stereo photography. Camera 3D Pro does not have that limitation, which is why I recommend it for beginners. The way it’s designed by requiring the user to move the phone also helps beginners avoid the common mistake of too small or too large a baseline. The app comes with the ability to crop and re-align the left/right images in case you disagree with the proposed alignment, which is based on the motion sensor’s data.

i3DSteroid and 3DHolic Camera work in the same way: you shoot the left image and then the right. When taking the right photograph, you are presented with a transparent layer of the left image to aid in the alignment. The easiest way to get a left/right view is to gently move your weight from one leg to the other.

These apps work well if you have normal lighting conditions and do not need to adjust focus, shutter speeds, or light sensitivity (ISO). If you need (or want) more control, you can use your favourite photography app or DSLR camera, and shoot sequential left and right images.

To generate a stereo image on your phone from individual images, you can then use either Camera 3D Pro or i3DSteroid, or use a separate app, such as 3D Photo, which loads the left and right image, and displays them side by side with an eye aid (i.e. a dot at the top to help focus). For desktops there are many more programs to help edit or make stereograms from two pictures, and a comprehensive list can be found here. StereoPhoto Maker is probably the best-known tool. i3DSteroid is from the same developer.

Where can I see images?

Below are a few of mine, taken in Gothenburg, Sweden:

Archway in Seminarieparken (4.5% deviation)

Sunflowers in Annedal (6.1% deviation)

Sculpture outside the Department of Journalism, Media, and Communication (2.3% deviation)

Thistles in Annedal (5.5% deviation)

A rock (2.9% deviation)

A building in Linné, near Skansen Kronan (3.0% deviation)

Roses in Annedal (6.8% deviation)

Sculpture of cobble stones near Olivedalsgatan (6.4% deviation)

Is it a fad?

Stereoscopic imaging has been around since 1838 and people have been interested in it ever since. It’s 3D movies and related technologies that have come and gone (e.g. House of Wax in 1953, IMAX, or Avatar in 2009). Whether 3D films and virtual reality are here to stay I dare not say. Whatever form they take in the future, stereoscopy will be an integral part of it as that is the principle behind human vision.

The images in the series have been taken with Camera 3D Pro on an iPhone 7 without any post-processing except occasional cropping. Note that none of the apps mentioned supports modifications to the default shutter speed, light sensitivity (ISO), or autofocus. This shows that even with a basic setup you can make pretty decent stereoscopic photographs.

In the next parts I’ll look at common pitfalls for beginners of stereo photography and an add-on for smartphones: Kúla Bebe.