Sometimes it amazes me how quickly physics moves from fundamental ideas to producing a new toy. The latest example comes from a number of experiments and theories about how opaque materials affect light passing through them, a topic we’ve covered extensively in the past. The work had the catnip qualities of being cute and simple and exploring some pretty basic physics ideas.
The idea behind the study was simple. Diffusing materials, such as white paint or sugar cubes, turn light into a chaotic mess. But if we could control how they scatter light, we could turn them into useful things like focusing devices. I know the researchers who pioneered this idea, and they were all rather conservative about possible applications. And that was appropriate; the ideas they had – medical imaging, high-resolution imaging and security applications – have all (with the exception of security) proved possible, but very difficult.
So I was a little surprised to see the ideas applied to holographic displays. I have to admit I never thought of it, but once you see the idea it’s like being beaten crazy by Captain Obvious.
Why holographic displays are bad
Holographic displays are worthless. This is a fundamental truth. How fundamental? It is actually a law of physics.
Perhaps the best way to illustrate this is to compare a 2D projection system to a holographic projection system. In a 2D projection system, a distorted light field carries the information across the image. A lens system mimics that image at some distance, where it hits a white screen. The white screen scatters the light in all directions, so no matter where you are in the cinema, you can see the picture (after bending your head to avoid Which dude). The size of the image depends only on the magnification of the imaging system – there is no relationship between the size of the image and the angle of view.
In holography, the first part is the same: a distorted wavefront carries the information to create a 3D image. But that recreation takes place at the focal point of the imaging system. To view it, you must stand in the path of the projection system. So an imaging system with a very tight focus (a lens with a short focal length) creates a small image that can be viewed from many different angles. An imaging system with a looser focus (a lens with a long focal length) generates a larger image, but can only be viewed by Which dude– bend your head and you’re out of sight.
This is actually a basic geometric principle. Sure, you can beat it with multiple lenses projecting from different angles or some other terribly complicated idea. But the added complexity only makes for an expensive and unreliable projector.
Scattering to the rescue
Scattering in itself is not useful. Imagine setting up a projection system that incorporates your bathroom window: the light comes through, but it turns your beautiful film into the world’s least interesting series of blurry blobs. This is because the glass is full of randomly oriented interfaces and air bubbles, which scatter the light in all directions.
If you shine a single color of light on such a scattering medium and look at the light after it passes through the scattering medium, you will not find a homogeneous, even haze. Instead, it will be quite bright in some places and exceptionally dark in others. Inadvertently, the light fields that have taken a lot of different paths through the scattering medium will add up in such a way that they create bright and dark spots.
The surfaces that spread in the glass are fixed – they do not change. So if we carefully control the input light field, we can make it go through paths that add up exactly. This allows researchers to use scattering materials as a lens, creating a sharp focus at a random location behind the scattering medium.
What a group of Korean researchers realized is that the image size and viewing angle are no longer linked. And by quickly scanning a focus, images can be drawn in the air and viewed from a wide variety of angles.
In an ordinary imaging system, the contributions of each spatial location of the object being imaged stick together. As a result, they all have similar angles when they exit the image volume and cannot be viewed widely. The scattering medium spreads all that information over the entire area of the scattering medium, so contributions to each part of the image come from anywhere on the output surface of the scattering medium. As a result, all the light arrives from all possible angles every part of the image.
In effect, the scattering medium acts like many different lenses in many different positions at the same time. Instead of a single image being taken, many are taken, each with a narrow angle of view. Because all of these images overlap, they produce one image that can be viewed from different angles. The viewing angle for the image is now determined by the size of the image volume and the width of the scattering medium. Therefore, to increase the image size and viewing angle, one needs a larger sheet of scattering glass, correspondingly larger optics and a brighter set of lasers.
Behind all this awesomeness is something even more amazing. For all of this to work, the researchers need to understand their scattering medium well enough that they can set the incoming electromagnetic wave to focus in just the right place in a 3D volume. But they know it so well that they can also scan that focal point around the volume quickly and controllably. Images can be created by performing a 3D raster scan. A three-color projection system allows them to create color images (in the newspaper they only demonstrated red and green separately).
In their demonstration, the team used spiral dots and letters projected at different distances to show that both the viewing angle and image size of the hologram have been increased. The image size is still quite small – a cylindrical volume 2 cm long and slightly less than 1 cm in diameter. Despite calling it HD, I would like to see a little less blur.
But the viewing angle is huge in relation to the image size. By simply using the standard optics in the imaging system, i.e. no scattering, a 0.5mm sized image has a viewing angle of just over 0.5 degrees. The product of viewing angle and image size remains constant, so to get a 30-degree viewing angle, the image size would be reduced to just a few micrometers.
By introducing scattering, the viewing angle was increased to 60 degrees and the image size was increased to 4.2 mm, which corresponds to an increase in the product of viewing angle and image size of approximately 1,000. In tests with a different spreading medium, they increased the product to over 2,000.
Now I should note (as the authors do) that there are other holographic display systems that achieve similar combinations of viewing angle and size, but they all suffer from combinations of shortcomings. Some are not dynamic, so only images are displayed. Some do not have a full color display view. Others are terribly complicated. The projection system here isn’t perfect, but given the simplicity of the system, it’s a remarkable result.
It is clear that the researchers hope that this will become a commercial display technology, but there are problems to overcome. First of all, the resolution of an image depends on the pixel density of the controller. In this version, the relationship between the number of pixels and the image resolution is complicated. The researchers used a deformable mirror with an array of 64×64 pixels. In an ordinary imaging system, this would be laughable. But this is only to control light propagation through the scattering medium – the actual light source may have a higher resolution. These 4,000 pixels determine how well defined the grid point is and how precisely it is positioned. It is not yet clear to me whether this is good enough; more pixels may be needed.
The second problem is that I know these systems to a certain extent and they are very sensitive to small changes in tuning. This cannot be changed, but it can be taken into account by an automatic calibration system. As far as I know no one has done anything like this before and it’s unclear how often it needs to be recalibrated in a house where the temperature changes a lot on a daily basis. Quite a task, I think.
Nature photonics2017, DOI: 10.1038/NPHOTON.2016.272