How Does 3D Actually Work?

Printer-friendly version

Walk around any Digital Signage trade fair and you will see the growing sector that is 3D. Already launched with some success directly to the public as televisions, the question is now whether or not 3D digital screens are going to be part of the next wave of Digital Signage. And if it is, how much do you really know about 3D? Meko, the European Display Market Research specialist, has produced an independent White Paper sponsored by Samsung LCD Business, looking at the different aspects of 3D DS, including how the screens actually work.
 
 
A background to 3D Technology
 
 
Introduction

Depth perception is an important feature of the Human Visual System and displays which convey a realistic impression of depth can enhance and improve the degree of emotional involvement of the viewer in the content. There is a lot of confusion about 3D screens and that confusion is being added to by the development of a number of new approaches to creating depth illusion. This backgrounder was written to try to explain some common misconceptions about 3D and to set out the advantages and disadvantages of different techniques.
 
There are several ways of creating a 3D screen. At present, there is no practical way to create displays that can show Stereo 3D with a strong feeling of depth without the viewer wearing glasses.
 
Autostereoscopic (without glasses) viewing is only currently practical on single viewer devices such as mobile phones or hand‐held consoles.There are two main approaches being used for 3D using glasses, based on ‘shutter glasses’ or on ‘line by line’ polarisation. Each has advantages and disadvantages. Shutter glasses need power so may be heavier, need charging and may be expensive, but deliver full resolution, although with reduced brightness. Polarising glasses are cheaper and lighter but mean that resolution is reduced by a half or more. It is a belief that maintaining the ‘artistic intent’ is a key requirement of any display device and the display should, as closely as possible, show what the content creator saw and intended the viewer to see. Depth should also be an improvement on existing 2D content and performance and not cause a reduction in other aspects of display performance such as resolution or colour.
 
 

 
What is 3D?
 
The term 3D is used in many ways to identify different things. In the world of electronic displays, it is used in three distinct ways:
 
• 3D Graphics – images that represent a scene with depth that may be shown on a simple 2D or some kind of stereo 3D display. This is the kind of image seen when playing a 3D game on a console or PC, and may be seen as 3D even with a simple 2D display.
 
• Stereo 3D (S3D) – this means a display image that presents different images to the left and right eye and is intended to give an enhanced sense of depth to the content being displayed, whether from 3D graphics (above) or S3D movies or broadcasts.
 
• Volumetric or ‘solid’ 3D displays – this kind of display projects an image into space so that the viewer can walk around the object which appears to exist in space. Most of what is talked about as ‘3D’ in the press and other media currently is Stereo 3D (S3D) and that is the focus of this backgrounder.
 
 
Stereopsis
 
The human visual perception system (VPS) is a complex combination of the eyes and optical system and the brain. The VPS uses a lot of different cues to understand depth in a scene, but one of the most powerful is stereopsis – the process of understanding the depth of items in a scene based on the disparity between the images that are seen in each eye.
 
Other depth cues that the VPS uses include:
 
 Relative size of objects
 Perspective
 Depth of field
 Parallax (both static and moving)
 Accomodation by the eye (focus of the eyes)
 Convergence by the eyes
 Occlusion (how one object blocks the view of another)
 Aerial perspective (how distant objects fade in contrast and change colour)
 Peripheral vision
 Blur
 
Really compelling impressions of depth tend to be based on exploitation of a range of cues, not just stereo viewing.
 
 

 
 
Stereopsis is naturally developed in most people in their childhood, although sight problems can cause this development to fail to take place and a significant percentage of people do not see ‘in stereo’.
 
SS Kim of Samsung Mobile Displays said that the development of S3D in entertainment in the future would be talked of as a ‘pre‐Avatar phase’ and a ‘post‐Avatar phase’. One of the reasons for the huge success of Avatar as an S3D presentation was that James Cameron showed a very good understanding of depth cues and used many of them in addition to stereo viewing to create a compelling and involving depth experience.
 
However, in this paper, we will focus on the issues of stereo viewing.
 
 
Creating S3D Displays
 
The fundamental task when creating an S3D electronic display is to ensure that not only are the correct left and right images delivered to the correct eye, but also to ensure that the images are not delivered to the wrong eye. If one eye sees an image intended for the other eye, it’s known as crosstalk and it is, at the very least, a source of discomfort or a breaking of the ‘suspension of disbelief’ that is a critical part of any experience on an electronic display.
 
There are a number of ways to achieve the needed split:
•  Separation by using a different display for each eye
• Separation by directing images at different angles to the eyes
• Separation by time
• Separation by polarisation and glasses
 
The first method is used in ‘virtual reality’ (VR) systems based on headsets where different images are created that are appropriate for each eye. Each eye can only see a single display, so there is no need for any further separation.
 
Different Images at Different Angles
 
There are a number of different techniques for using a single display with dual images that are directed towards different eyes. Techniques used include:
 
• Parallax barriers
• Lenticular and other lenses
• Liquid crystal lenses
 
Parallax barriers are relatively simple to understand. In these displays, alternate pixels in each display line are intended for the left and right eyes and a barrier is integrated which blocks the view of the incorrect pixel from each eye. Sometimes the barrier is switchable and is formed by a liquid crystal layer and that has the advantage of allowing the barrier to be eliminated for 2D operation. This kind of technology has proved popular in mobile phone displays.
 

 

 
With lenticular lenses, the lenses are applied to the surface of the display to direct the light from different pixels to the viewer. This technology has been developed for digital signage displays where there may be multiple viewers, so the lenses may control the light from a number of different pixels to create multiple views.
 

 
 
Liquid crystal lenses may to be used in the same way as lenticular lenses, with the advantage that they are electrically controllable, so the S3D effect can be switched off, with the display reverting to a 2D mode when the lenses are effectively ‘switched off’.
 
Why do we Need Glasses?
 
There are many comments about the possibility of eliminating the wearing of glasses for S3D.
This is really no problem for applications that have only one viewer, using the techniques detailed above, especially if the user can adjust the viewing distance to match their position and get the best 3D effect – for example on a handheld game console or mobile phone.
 
The biggest downside of S3D running without glasses is that each image has only half of the full image resolution.
 
However, if there are multiple viewers, the display needs to be able to create a pair of images for each viewer. Each pair has to be directed to the correct eye. This has three drawbacks.
 
First, the resolution of the display is radically reduced because of the need to create a number of individual image pairs – one for each direction from which the display can be viewed.
 
The second problem is that when a system that uses fixed view points (such as lenticular lenses), the viewer will typically see ‘artefacts’ when the eyes are not directly aligned with the images. Typical artefacts are dark bars in the image, when the eyes are positioned between two pairs of images.
 
The third problem is that the display has to be engineered for a specific viewing distance. Although there may be a depth effect when used at the wrong depth, the effect may not be good or may be uncomfortable.
 
At the time of writing, companies, including Microsoft, are working on systems that use cameras to identify exactly where the viewers’ eyes are and combine this tracking with ‘steerable’ optics to effectively direct the S3D images appropriately to each eye, but they are in the research stage and not yet ready for volume production.
 
 
S3D Displays Using Glasses
 
The fundamental advantage of using glasses for S3D is that the division of the two images needed for S3D can be separated at the point where the viewer is, so that multiple viewers can see the same image, regardless of their position relative to the screen.
 
There are three technologies that are most commonly used to exploit this idea.
• Active ‘shutter’ glasses
• Polarising glasses
• Special colour filter systems
 
The special colour filter systems are used in the cinema only. As this backgrounder is intended to cover TV and displays, we will focus on the first two systems.
 
 
Active Shutter Glasses

Active shutter glasses have been used for S3D for many years in professional applications such as molecular modelling3. The concept is that the left and right images are shown sequentially on the display. The viewer wears glasses that look like sunglasses but each lens contains a layer of liquid crystal which can be switched to be clear or opaque. By synchronising the lens over each eye to match the timing of the two images that are displayed, the appropriate image is directed to the correct eye.
 
There are advantages and disadvantages of this system:
 
Advantages of Active Shutter Glasses

• Each image can be shown over the whole surface of the screen and can be displayed at full resolution.
• There are no particular limitations on viewing angles or the number of viewers beyond the limits of the display itself.
• There is no need to modify the display device itself, so that there is no change in 2D performance.
• If the synchronisation circuitry in a screen is made as a plug‐in option, a suitable screen could be later converted to 3D.
 
 
Disadvantages of Active Shutter Glasses

• By definition, because they are active, there needs to be a power source to operate the glasses, which can make the glasses heavier than polarised glasses. Glasses need to be kept charged.
• The glasses can be expensive and each viewer needs a pair.
• Glasses from particular TV and display brands are not generally usable on sets from other brands. Third party ‘universal’ glasses may not be as well optically optimised to match the set as the set maker’s glasses.
• Brightness and contrast are reduced because of the need to have breaks in the displayed images between the different views.
• If a slow display device is used, there may be ‘crosstalk’ where the image intended for one eye ends up being visible in the other eye.
• If the display area is small relative to the field of view, there may be some flickering in the background that is apparent to the viewer.
 

 
 
Polarising Glasses
 
In systems using polarising glasses, the images are divided by using different polarisation of light. In the cinema, this is achieved by having an active polariser in the projector’s light path to switch the polarisation of the entire image. In TVs and displays using polarisation, a special filter is applied to the front of the screen that polarises alternate lines in the image. The filter may be based on glass or on film. The polarisation can be linear, but this causes problems if the head of the viewer is not vertical, so most displays that use this system use circular polarisation, with alternate lines correspondingly polarised.
 

 

 
The viewer has a pair of glasses that have the appropriate polarisation filter applied to the lens.
 
 
Advantages of Polarised Glasses
 
• The glasses used for this method can be very cheap and light. They can be made of curved glass which is better optically, and keen viewers could get ‘prescription’ glasses made or ‘clip‐ons’ for existing glasses.
• No power is needed in the glasses.
• It is likely that glasses from different brands may be interchangeable.
• There is less light loss, so images are brighter than with shutter glasses.
• There is no potential issue of flicker.
• The low weight of the glasses and elimination of any flicker means that polarisation techniques tend to be the choice of those that have to work with S3D in broadcasting for long periods.
 
 
Disadvantages of Polarised 3D
• The cost of the polariser has to be added to the set cost, even if the buyer does not plan to watch any S3D.
• The vertical resolution of the display is halved and watching broadcast S3D in Europe is likely to mean that the resolution is down to a quarter – going back to Standard Definition.
 • Depending on the viewing distance and the eyesight of the viewer, horizontal scan lines may be visible on the display, like a large old fashioned CRT showing standard definition images.
• The brightness and contrast may be affected by the vertical viewing angle (if viewed from above or below).
• Current ‘universal’ S3D glasses designed for shutter glass applications will not work.
 
 
Resolution Issues of Polarisation

While the writer is not convinced of the argument that by showing 540 different lines to each eye produces 1080 resolution, there is another problem with this approach.
 
S3D content coming from 3D broadcasts is currently being transmitted in what are known as ‘frame compatible’ formats, that is to say both the left and right images are transmitted within a single frame as they would be for existing HD transmissions. This is necessary to allow existing set top boxes and digital HD TVs to decode the signal. The two main optional formats are ‘side by side’ and ‘top and bottom’ transmission. Both techniques effectively compress the two images into a single HD frame by halving the resolution.
 
If the broadcaster was to use ‘Top and Bottom’ format, then the polarised solution would show the half resolution image as it came from the broadcaster without further resolution reduction.
 
 

 

 
However, if the broadcaster uses the ‘side by side’ format for the S3D (and as far as we know, this is the format used by all European broadcasters at the time of writing), then the broadcaster reduces the resolution by half in the horizontal direction. In the writer’s view, this combined with the reduction in vertical resolution will not produce much more than quarter resolution, effectively moving S3D over broadcast back to standard definition, a retrograde step.
 

 



Alternative Polarising Approach
One alternative being investigated to avoid the resolution loss that comes from line by line polarisation is to add a controllable optical layer over the whole display area that can switch the polarisation of the whole image, allowing the use of passive polarised glasses. This method keeps all the advantages of the polarised glass system but maintains full resolution. The main disadvantage is that the extra liquid crystal layer is relatively expensive – and the technique is not in mass production at the time of writing.
 
 
Conclusion

The aim of this paper is to summarise the current state of development of 3D screens.
While everyone would like to see ‘no‐glasses’ S3D, in practice that is not currently possible except where there is a single viewer.
 
 
There are two alternative technologies that will be in the market over the next couple of years. Each has advantages and disadvantages. Systems using shutter glasses preserve the full resolution of the image but have more expensive and heavier glasses. Systems using polarisation have lower cost glasses but at the current ‘state of the art’ will lead to some loss of resolution especially on broadcast TV content.
 
 
Our view over many years is that the job of any screen is to match the ‘artistic intent’ of the content creator. That is to say that as a base level of performance, the display should show images as close as possible to the images as they were seen by the creator or broadcaster.