3D TV—is it good enough?

The human visual system uses various techniques for discerning depth in order to perceive objects in three dimensions, such as:

Typical systems for displaying images in three dimensions

Projections of three dimensional images into two dimensional space, using techniques such as perspective drawing with vanishing points, forms the basis of artwork dating back many centuriesMore accurate representations of perspective can be achieved through matrix transformations of 3D space in to 2D space, which is feasible due to the general availability of powerful computers.

Stereoscopic images (which exploit stereopsis) have been produced commercially for almost as long as photographs, and continue to be mass-produced today.  Some of the first stereoscopic colour positives mass-produced as Viewmaster reels still look vibrant today and convincingly fool the brain into “seeing in 3D”; although a careful examination of Viewmaster images reveals a basic limitation of the technology, that the eye cannot focus on what the viewer chooses, but rather is restricted to the plane and depth of focus selected by the photographer or artist.

Lenticular lenses and parallax barriers enable multiple images to be displayed on a single surface, so that different images are seen when the image is viewed from different angles.  With precision-manufactured lenticular prints, it may be possible to create convincing three dimensional effects exploiting perspective, stereopsis and focus.

Laser holographic images reproduce all the effects we associate with three dimensional imagery, but are deficient in resolution and colour characteristics and inconvenient to produce.

The challenges of comfortably presenting a convincing 3D illusion

Realistic focus effects are difficult to recreate since they are formed by small deviations in the angles at which light diverges from its origin (or virtual origin).  To enable the viewer to focus at will upon different parts of a recorded or synthesised image, each pixel on a two-dimensional display must present different colours at subtly different angles (so that each pixel itself consists of a two dimensional array of sub-pixels).

The resolution of the underlying image is a finite resource for the presentation of various effects:

  • 2D Pixel count, from the perspective of a fixed observer,
  • Parallax and focus (angular differentiation of the image, through the presentation of different colours at different angles from any given “pixel”)

Creating a comfortable still 3D illusion requires an underlying image of sufficient resolution for presenting both of these effects with a high degree of fidelity.  Where a still printed image is used (e.g. a large colour positive transparency printed with a laser and chemical process; a medium with excellent colour fidelity), very high resolutions can be achieved.  Such images can take several minutes each to produce, and might form the basis of convincing 3D illusions.

Animated 3D images would require, in addition to these qualities; rapid image generation, high-bandwidth transmission and high-fidelity presentation.  A further fundamental problem with moving 3D images is that the more convincing they are and the longer they engage their audience for (or, the more successful the display technology is), the more chance there is of viewers suffering from motion sickness (sea-sickness).  This happens because of the disparity between motion parallax information received by the brain via the optic nerves, and the acceleration & balance signals received from the inner-ear.  The more motion-parallax is exploited by the creators of a 3D movie, the more its viewers will suffer from sea-sickness.  (The designers of 3D display devices should consider themselves lucky most humans are not tetrachromatic, and that we do not have magnetic senses as some birds and sea creatures do—since if we had such senses, issues of sensory disparity would be even more challenging to overcome).

Worse still for 3D movies, if the display technology does not faithfully replicate all visual effects the human eye is capable of discerning, extended viewing periods may impair viewers' eyesight (for example, by encouraging laziness of the focusing muscles).  Staring at 3D images for long periods of time is not advisable since the brain continually calibrates control of eye focussing muscles and its mechanisms for interpreting visual images according to the visual signal presented through the optical nerve (so that looking at false 3D images effectively reprograms the brain's subconscious mechanisms for controlling important muscles and interpreting real 3D images that are normally important to our personal safety.)  This is not a significant problem when viewing still 3D images for a few seconds each in a session of up to about 15 minutes; but might be problematic for those viewing 3D images for longer periods (such as might be the case for animations).

What is 3D TV/ cinema?

Experiments in 3D TV and cinema began in the 1940′s before being largely rejected as an impractical novelty for specialist applications and an inferior medium for the presentation of a theatrical experience.  The technology was developed in specialist IMAX 3D theatres with films designed for the specific advantages of IMAX 3D, which gained a niche following.  With the advent of digital cinema and more practical & economical projection equipment, 3D cinema is now returning to the entertainment market.

3D cinema and TV has used a variety of technologies, for example:

  • Colour based image multiplexing (red/green or red/blue) so that different images are presented to the left and right eyes (this can affect colour perception for some time after a viewing session)
  • Multiplexing through light polarisation (as applied in IMAX 3D and many 3D TV systems)
  • Displays with a lenticular overlay.
  • Displays exploiting optical interference (or holographic) effects.

Is current 3D display technology good enough?

Some quick calculations (based on studies the author conducted in the early to mid 1990's) may help illustrate why this technology will not be available in a worthwhile and economical form for some years to come:

  • The maximum resolution for common computer displays is about 2560×1600.
    This limitation arises from:
    • Semiconductor technology for current graphics chipsets
    • Display cable bandwidth
    • Manufacturing constraints for the screens themselves
  • The author estimates that to produce an image comfortable and safe enough for two-hour viewing sessions, the screen should exceed the following specifications:
    • Apparent image resolution ≥ 1280×800 pixels in size
    • > 75cm diagonal screen size (i.e., > 64cm×40cm)
    • Typical angular resolution (both horizontal and vertical) < 0.04°
    • Viewing angle of > ±20° horizontally or vertically from any pixel

To achieve these specifications for multiple simultaneous viewers (particularly where the viewers are mobile), the underlying intrinsic screen resolution would need to be at least 640,000 pixels horizontally by 400,000 pixels vertically—i.e., 250× better horizontal & vertical screen resolution than is commercially possible today.  Using technology available in 2011, it is simply not possible to manufacture such a screen, nor the control technology that such a screen would require.  The artificial reproduction of a three dimensional image with no discernible difference from the real scene in terms of parallax, focus, colour and viewing angle, which would require a screen substantially better than the one described above, is certainly impossible using the technology available in 2011, and is not even close to becoming possible, let alone commercially viable.

The true goal of current 3D display technology must therefore be to merely fool a willing observer for a limited period of time into believing that an image is three dimensional—to make it sufficiently smooth and convincing that a cooperative observer will not suffer any serious discomfort as a result of their viewing.

Predictions based on industrial trends

Until high-resolution 3D displays become available, and until feasible methods are devised for economically acquiring real-world data for display on such devices; 3D displays are likely to remain a niche technology for:

  • Displaying scientific data and design engineering concepts;
  • Visualisation in industries such as space, mining, nanotechnology, astronomy, medical sciences, military photoreconnaissance and other fields where human presence or direct visualisation can be dangerous, expensive, inconvenient or impossible;
  • Computer games, fairground rides and advertisements.

Are we on the verge of 3D display adoption becoming commonplace, phasing out less useful and marginally less expensive 2D displays?  It is the author’s opinion that we are not, and that the next 50 years will see continued periodic flirtation with updated versions of the same 3D display technology that has largely been a commercial failure for the last fifty years.

Beyond the next half-century, we might see semiconductor-based displays where the graphics chip (display control hardware) is integrated into the same substrate as the light-emitting (laser-based or LED-based) components.  We might ultimately create a screen possessing appropriate resolution, colour and information management characteristics.  However, based on pragmatic calculations and established industrial trends, we feel sure that this goal will not be realised for at least another decade, and that such technology will probably not be available commercially on a large scale for another three decades.


Disclosure & invitation for remarks

The author has studied computer science, physics, semiconductor technology, computer graphics, technical drawing and photography at various levels; and has had some contact with the large format digital printing industry.  While the author has a personal interest in printing technology and 3D display technology and has investigated this from a technological perspective, he has not personally reviewed many of the latest 3D display devices and has no intention of conducting any rigorous usability testing as he sometimes suffers from mild motion-sickness when viewing 3D illusions, and has plenty of other things he would rather do than reexperience nausea.  The author has experienced IMAX 3D, and considers well-made 3D movies to be an excellent art-form.  Comments would be especially welcome from people who have personally experienced other 3D media and display technologies, or those who have a sound understanding of how any particular 3D technology works and who wish to challenging the author's figures and estimates from a pragmatic perspective.

4 thoughts on “3D TV—is it good enough?

  1. In today’s news:

    “Thunderbolt smokes USB, FireWire with 10Gbps throughput”
    http://arstechnica.com/apple/news/2011/02/thunderbolt-smokes-usb-firewire-with-10gbps-throughput.ars

    This standard probably will be upgraded to include 100Gbps fibre-optical options within a decade. Now, we just need some new compression standards for true 3D video/ images, in order to economically transmit this bandwidth-intensive media from computer to display…

  2. Pingback: Nintendo 3DS Teardown - Gadget Venue

Comments are closed.