Chris Chinnock, Insight Media

A 3D display can offer a truly magical experience. You can be a diver, wonderfully immersed in an underwater world of tropical fish, or you can accompany an astronaut on a space journey as if you were almost weightless yourself. Giant Screen Cinema has shown such a potential of 3D to awe its audiences. But the flip side is that the current technical limitations cause distractions, that in the longer run seem to outweigh the wow factor. Maybe this is the reason why 3D cinema has a history of coming back in waves, as new generations want to experience the wow factor. We need to solve the technical issues if we want 3D to stick around and become a feature of home entertainment displays in the future.

In this article we will discuss the market needs and technical requirements for good 3D displays. This content is part of a new white paper on Realfiction’s, recent 3D display technology that was developed with these requirements in mind. You can access the white paper HERE.

Some of the considerations a 3D display developer must consider are detailed below:

• Required viewing sweet spot
• Improved horizontal parallax
• Vergence and accommodation conflict
• Vergence and motion-parallax conflict
• Miniaturization effect
• Limitation of discreet perspective views
• Moire-like interference
• Immersiveness from large display sizes
• Insufficient frame rate
• Cross talk

Required viewing sweet spot

Consumers and professionals want a 3D display that provides natural 3D images without glasses. Some applications are for single viewers, but many applications need to support multiple viewers with good 3D images. So far, glasses-free 3D displays have always had a viewing “sweet spot”; a zone that has a certain horizontal and vertical viewing range as well as an ideal depth range from the display. Outside the sweet spot the 3D image can fall apart. Clearly, a technology that eliminates the sweet spot is desirable.

Improved horizontal parallax

Auto-stereoscopic displays are glasses-free displays that can provide a few different stereoscopic views in the horizontal direction (horizontal parallax). Super-multi-view displays are auto-stereoscopic displays that overlap multiple views to provide a more seamless viewing experience (improved horizontal parallax). This can provide a decent look around capability with fairly-high fidelity. Such displays may or may not use eye tracking. Applications like TV, gaming, automotive, surgery, conference rooms and digital signage all desire the ability to support multiple users with high fidelity 3D images.

Vergence and accommodation conflict

It is also desirable for 3D displays to minimize any eye strain and nausea. This is often caused by the conflict between vergence (how your eyes tow in or out when focusing on an object) and accommodation (where your eyes focus). In stereoscopic or auto-stereoscopic displays, the image is always in focus at the display surface, while the stereo image pairs may place the virtual object in front of or behind the screen surface. This is not a natural condition, and it is the cause of the vergence-accommodation conflict. Addressing this conflict is important in any 3D display design, especially in personal displays for close viewing, where the disparity between focus and vergence is pronounced for typical content like, for example, a 3D movie. For larger displays observed at greater distances, it is still important to keep this conflict in mind when designing content and limit the duration of “in your face” type effects.

Vergence and motion-parallax conflict

Another factor which may cause discomfort is the conflict between vergence and motion-parallax. For example, when watching a 3D movie in a cinema and sitting still, you can have a realistic feeling of depth. However, when you move your head the view perspective does not reflect the head position change. This is due to a lack of look-around capability. It gives an uncomfortable feeling that an object or a whole scene is somehow strangely “strapped to your head” and always turns towards you, which can make you feel quite seasick if you don’t sit still.

Miniaturization effect

A cause of distraction is also the so-called “miniaturization effect”, which is the artifact that people and objects look smaller than in real life, giving a feeling that your favorite actors are characters in a puppet theater. The problem is especially pronounced when content is shown on screens smaller than the content was originally rendered for (like watching a movie rendered for theatrical on a TV screen). When watching 2D, like a painting or a photo, we are so used to a person being much smaller than in real life that we are not offended by it. But in 3D, when we have depth and when objects appear solid in front of us, our brain automatically correlates the extension of objects on our retinas to physical sizes of the objects. Miniaturization can be avoided when you know the display size and the viewing distance for each viewer in the audience by adjusting stereo parameters such as camera distance and left/right eye image horizontal offset (parallax). But if viewers are not located at the same distance from the display, it requires delivery of an individual 3D image to each viewer.

Limitation of discreet perspective views

Another common cause of reduced image fidelity in super-multi-view displays is a limited number of discreet perspective views. Even with 100 individual perspective views you will notice distinctive “jumps” in perspective when moving your head around. One way to mitigate this is to have overlapping, blended perspective views, but this results in smearing of pixels and blurring of the image, especially for objects appearing at large depth in front of the display. Another way to mitigate perspective “jumps” is to track the position of the viewer and render the views in real time, either by a ray tracing system like a game engine or by AI-enhanced interpolation of perspective views. This requires delivery of a dedicated 3D perspective image to each viewer.

Moire-like interference

Many autostereoscopic displays today use slanted lenticular arrays. These have the limitation that moire-like interference effects between the color mask and the slanted lenticular array reduces clarity of fine details such as small text. This has been mitigated by some manufacturers by a clever electrically controlled lenticular array. This can be switched off to act like a clear window, allowing the display to operate in full-quality 2D when using a text editor or spread sheet, for example. However, it does not change the fact that in 3D mode, the clarity of fine details is reduced, which effectively excludes many professional use cases of lenticular-based 3D displays.

Immersiveness from large display sizes

To have a truly magic, immersive and still comfortable 3D experience, a large display is important to avoid the feeling of either looking out through a small window or into an aquarium. A large display can be located so the display edges are outside the central part of your vision. This is important, not only for immersion, but also to avoid frame-violations, i.e. objects stereoscopically appearing in front of the display, but being cropped by the display frame. This creates conflicting depth cues for the brain, which not only reduces the depth experience, but also causes discomfort. While a large display can mitigate this, consumer willingness to pay for a large display would most likely depend on the ability to watch it with a group of friends or your family, hence requires multi-viewer technology.

Insufficient frame rate

The frame rate of the display is the time needed to provide new images to the left and right eyes. It is generally accepted that a frame rate of 60 Hz is needed as a minimum to minimize motion artifacts and flicker. That means the display must be refreshed at a minimum of 120 Hz per stereoscopic 3D image pair in order to meet the 60 Hz frame rate requirement. Some 3D displays do not meet this requirement, and others meet the requirement but are limited to single viewer use.

Crosstalk

All stereoscopic displays exhibit some level of crosstalk, i.e. one eye sees a faint overlay of the image intended for the other eye, sometimes also referred to as a “ghost image”. So-called “ghost busting” algorithms (yes, that is the term used in the cinema industry) can digitally pre-process images and mitigate this to some extent, but it is still important to keep the native crosstalk low, preferably at 1 or a few percent. A lot of factors contribute to crosstalk. For time-multiplexed LCD displays, pixel response time is a major contributor, for example.

Perspective and summary of limitations

Many companies today market their 3D displays as “holographic”. In fact, most are not true holographic displays. They are mostly super-multi-view type displays. True, large holographic displays with moving images in full natural color will address the needs discussed above, but they remain in a development stage. On the other hand, a term like “3D TV” has become associated with products suffering from all or most of the limitations discussed above plus reduced resolution, crosstalk, flicker and a need for glasses, so this is not a good name for a new type of display either. There is a need for a new category name for a new and improved super-multi-view 3D display.

Today, no 3D displays, even glasses-free ones, address all the needs discussed above. Realfiction’s Hybrid Scan Display addresses most, but not all these market needs. There is still a viewing sweet spot with some clever ways to mitigate, but not solve, the vergence-accommodation issue. The big innovation is the ability to support multiple users simultaneously with the same or separate stereoscopic images. Such a capability helps to mitigate the conflict between vergence and motion-parallax and the miniaturization effect. Content is rendered by a Unity game engine thus helping to reduce any perspective jumps. And the technology is not lenticular based, so there is no loss of resolution in changing from 2D to 3D. All these factors address the main issues with the acceptance of 3D displays and will enable many professional and consumer use cases for larger displays.

About Chris Chinnock:

Chinnock holds a BSEE from the University of Colorado and has previously worked for companies such as General Electric, Honeywell, MIT Lincoln Labs and Barnes Engineering. In 1998 he founded Insight Media, which published newsletters (including Display Daily) and market reports plus custom consulting on the emerging parts of the display industry. Insight Media also produced many display events such Projection Summit, Display Summit, QLED and Advanced Display Summit, Streaming Media for Field of Light Displays (SMFoLD), and more. Chinnock also helped found and lead two display-focused industry organizations: the 3D @ Home Consortium and the 8K Association. He currently provides custom writing and marketing services for clients with advanced display-based technology.