3D and Holograms

10 September 2013

SIG: 3D Imaging and Holography

Overview of how stereo 3D and holograms are made and displayed by Dr Alan Cooper LRPS.   To see further images, go to our Gallery.

1.   Capturing the world in 3D photography

A 3D (stereo) camera records two views as they would be seen by our two eyes.  The differences between the two are clearly noticeable for objects up to a few metres away, but diminish with distance; the same simple geometrical effect ("parallax") that is measured in an old fashioned rangefinder camera.  So 3D photography is mostly about objects a few metres away. However, with some modifications, macro/micro stereo photographs can be taken using a much smaller separation of the lenses, and conversely distant objects need a much longer baseline.  

Suppose the stereo pair is printed at such a size that the centres of the images are about 65mm apart, as is the case with our eyes.  When we look at them, head on, optical axes parallel, the task for the brain in achieving alignment is so similar to what our eyes are doing all day long that no special viewer is needed.   This is called freeviewing a stereo pair.   It is not quite instant, it does need a little practice, but once the eye/brain has discovered the trick, it works automatically.   Here are three examples, the first and second with a standard baseline (65mm), the third with a longer baseline.   The gap between the two images is not necessary, but it makes freeviewing a little easier.


Such small images are useless for the cinema.  For early 3D films, in black and white, the left view was coloured red and the right one blue (or cyan) and they overlapped, so alignment was built in and the pictures could be large.  Using filters of the same colours, each eye saw the appropriate image - it does not give a clean or complete separation, but it works.    It is called anaglyph, here's an example.

Anaglyph Jasmine Tea

This also demonstrates close up macro 3D with a very small baseline, 3mm)

It sounds like a very bad idea to use anaglyph with full colour images, because the colours get very muddy and confused.    However there is a modern more sophisticated system  (Dolby colour, originally called Infitec) which works very well. It may be a popular method in the future.

To get full resolution in the home, the best way now is to use a 3DTV.  Modern digital 3D cameras can be connected directly to the TV, and there are controls to adjust the images to suit the TV and your eyes. The format needed for 3D images is .mpo, the camera makes them and the TV recognises them.   You can put any images into .mpo using the program Stereophotomaker.

One other way is to use a mirror viewer, which has mirrors at 45 degrees set wider than 65mm - just like a military rangefinder.  Another way is to use cross eyed viewing of prints, but not everyone can manage this, it is harder than freeviewing with parallel eyelines. 

2.  Capturing the world by holography

There is currently no way to put on a hologram on a website, but here is a 3D side by side pair of a hologram by Martin Richardson. The hologram from which it is taken is quite large - about 40cm  x  30cm.   The 3D version can be freeviewed; it is convenient, but it is a poor substitute. 

3D Image of Hologram

A real hologram contains a huge amount of detail, and exhibits parallax in all directions,  whereas a 3D pair only records the horizontal parallax.   To see real holograms, find a hologram exhibition, there are a few per year - or join the 3D and holography group! 

A normal photograph records intensities, but a hologram records phases as well.   Since the wavelength of visible light is less than a micrometre the resolution of the special film used for recording the hologram has to be considerably less than a micrometre. Holograms therefore have a huge amount of detail, as a by product of the way they are made.   The idea of digital hologram, manipulated on a PC, sounds, at first, pretty hopeless.   But many commonly used sensors already have sub-micrometre pixels so there is hope, and the dream of holographic TV gives the incentive.

3.  Creating 3D Drawings

A  3D pair of images does not have to be photographs, it can be drawings.   That takes 3D into the realm of purely imagined scenes, a big leap. But making two drawings with just the correct amount of horizontal displacement everywhere is extremely skilled.  Some large and rather complicated machines were built to help.  Nowadays computer aided drawing makes it much easier, but all the depth coordinates have to be fed into the programme.  (The next section, fly’s eye stereograms gives a short cut method). Any such CAD  programme will automatically create the view from any angle.   At the professional end, they are used to make films such as Avatar.  At the amateur end there is a programme called SketchUp which is a sort of graphical Lego - it is quick and easy, especially  for anything with square corners), and you can give your imagination free reign.    Here's an example, an aeroplane (which does not have square corners!) made into an anaglyph so that it comes forward in front of the clouds.

Spitfire Anaglyph SketchUp

A few stereographers, notably John Hart, have used fractal shapes as much more sophisticated building blocks, creating very beautiful shapes which grow and move.   The recursion relations which define the fractal nature are supplemented by rules which create depth, colour and movement.    John Hart has also used multidimensional orbits as translucent building blocks.   The results are wonderful, but they arise from supercomputers or banks of PC's running for months or even years.  Don't expect to knock one off on a wet Sunday! 

4.  Playing with depth - fly's eye stereograms. 

When you have got your eyes used to freeviewing, try the following – you don’t need a viewer.   Concentrate on the image and a section shaped something like Saturn will detach and float to the foreground)

Magic Eye - Planet of Flowers

Like freeviewing, it needs the eyes looking along parallel axes, as if to infinity.   The encoded image stands out from the background by virtue of having had small horizontal displacements added to its elements, like a 3D image.   This would be impossibly tedious but for a Photoshop trick - it can be induced to interpret tonal values as displacements.  A tonal gradient pattern therefore becomes a depth map (or vice versa).  

Here is the depth map for the flowery planet.

Saturn depth map for M Eye SketchUp

Depth maps are very important, they are the link between the two pillars of 3D, stereo photography and laser scanning.   They are complementary, stereo for small scenes, laser scanning for big ones.