1
When do New Technologies Become Economically Feasible?
The Case of Three-Dimensional Television
by
Pei-Sin Ng
Jeffrey L. Funk*
*Contact Author:
Associate Professor
National University of Singapore
Division of Engineering and Technology Management
9 Engineering Drive 1, Singapore 117576
[email protected]; 65-6516-7446
Forthcoming, Technology and Society
2
When do New Technologies Become Economically Feasible?
The Case of Three-Dimensional Television
Abstract
This paper analyzes the timing of a new technology’s economic feasibility using a simple
yet novel approach. While the conventional wisdom that costs fall as cumulative production
increases does not enable us to analyze this timing, the proposed approach enables us to do so
using existing technological trends in the components that form a new technology’s system.
For 3D television, although the concepts that form the basis of 3D television have been
known for many years, improvements in specific components within two-dimensional (2D)
televisions such as the liquid crystal display (LCD) are finally making 3D television
economically feasible. More specifically, improvements in the frame-rates of 2D LCDs are
making it economically feasible to introduce time sequential 3D, which requires special
glasses. Similarly, increases in the number of pixels per area (resolution) will probably make
auto-stereoscopic 3D LCDs economically feasible in the next five to ten years and thus
eliminate the need for special glasses.
Keywords: technological discontinuities; technology paradigms: geometric scaling; technical
feasibility; economic feasibility; three dimensional television: liquid crystal display
3
1. Introduction
Understanding when a new technology might become economically feasible and begin to
diffuse remains an allusive goal. The economics literature focuses on cumulative production
as a key driver of diffusion in that the cost of a new technology falls as cumulative
production increases in a so-called learning or experience curve, According to such a curve,
product costs drop a certain percentage each time cumulative production doubles [1] [2] as
automated manufacturing equipment is introduced and organized into flow lines [3]. However,
if cost reductions primarily come from production, as the learning curve suggests, by
definition cost reductions cannot occur before production occurs thus making it very difficult
to use a learning curve to analyze when a new technology might become economically
feasible and thus begin to diffuse.
The management literature uses the term technological discontinuity to distinguish
between new and old technologies where products defined as discontinuities are based on a
different set of concepts or architectures than are the old technologies [4]. However, while
there is wide agreement on the descriptions and timing of specific technological
discontinuities, most research on technological discontinuities focuses on the existence and
reasons for incumbent failure and in doing so treats these discontinuities as “bolts of lightning”
[5] [6] [7] [8]. For example, the product life cycle, cyclical and disruptive models of
technological change do not address the sources of technological discontinuities and instead
their emphasis on incumbent failure implies that the timing of these discontinuities depends
entirely on cognitive factors and thus cannot be easily analyzed [9] [10].
This paper analyzes the timing of a new technology’s technical and economic feasibility
using a simple yet novel approach. This approach builds from the notion that technologies
can be thought of as a “system” of components [11] [12] [13] [14] [15] where new
technological systems often borrow components from existing technological systems [16].
Thus, this approach focuses on the concepts that form the basis of a new technological system
4
and the levels of performance that are needed in the relevant components before new the new
concepts become technically and economically feasible. This enables us to utilize
technological trends in the relevant components to analyze the timing of economically
feasibility. Since data on technological trends are available for a wide variety of existing
components, the ability to utilize this paper’s approach primarily depends mostly on our
understanding of a technology’s system and components.
This paper demonstrates this approach using three-dimensional (3D) television, a
technology whose basic concepts have been well understood for many years. Building from
one author’s experience with televisions and a second author’s knowledge of technological
change, the key components in a 3D television are identified and analyzed. Such a system
includes LCDs, ICs, and other electronic components where improvements in these
components continue to be made somewhat independently of the existence or introduction of
3D televisions. For LCDs, costs have been falling quite rapidly as firms have gradually
increased the size of the substrate and production equipment. In addition, improvements in
their frame-rates and in the number of pixels are also being made in response to demand from
other electronic products and these improvements are gradually making 3D television
technologically and economically feasible
This paper first describes the ideas that form the basis of this paper’s approach, the
sources of these ideas, and their application to televisions. Second, it briefly describes the
research methodology. Third, it summarizes the improvements in LCD displays and other
electronic components that are making 3D LCD televisions technologically and economically
feasible. Third, it describes how these improvements are improving the technological and
economical feasibility of time-sequential and auto-stereoscopic 3D televisions, which are the
two most discussed methods of achieving 3D television. Time-sequential 3D displays
requires special glasses that include an active or passive LCD display while auto-stereoscopic
3D LCDs do not require glasses. Fourth, this paper speculates on a pattern of diffusion for 3D
5
television
2. Key concepts
Technological discontinuities are typically defined and classified by the extent to which a
new product, when compared to a previous one, involves changes in the core concepts that
form the basis of a product or in the linkages between a product’s key components [4].
Radical innovations change both the concepts and the linkages, architectural innovations
change only the linkages between components, and modular innovations change only the core
concepts of a single component. Although some scholars also focus on a technology’s impact
on the linkages between a firm and the market [17], these types of discontinuities, including
so-called disruptive ones, can also be classified as either radical or architectural innovations
[7].
This paper focuses on radical innovations in televisions and in particular it focuses on 3D
televisions. Looking at the concepts that form the basis for electronic displays, the first ones
were cathode ray tubes (CRT) that were initially used in oscilloscopes and only later used in
televisions. In the cathode ray tube, one electrode emits electrons and electrons striking
phosphors cause the phosphors to luminesce [18]. By controlling the direction of the
electrons with an electric field, one can determine the specific phosphors on a glass tube that
will be struck by the electrons. By using three electrodes or so-called electron guns and the
right type of phosphors, color images can also be displayed on the television screen. Like
other forms of electric discharge tubes such as incandescent lights, many improvements in
cathode ray tubes came from finding better phosphors and better materials for the electron
gun. Nevertheless, limits to these improvements began to emerge many years ago. Their
miniaturization (and thus their costs) has been severely constrained by the size of electrodes,
glass bulbs and sockets, and the resolution of them is constrained by similar problems (an
ability to control electrodes, emitted electrons, and impacted phosphors).
6
LCDs are based on a different concept than are CRTs. They use an electric field to control
the orientation of the liquid crystals. Although the properties of liquid crystals had been
identified by the late 19th
century, it was not until scientists were able to control them with an
electric field in the mid-1960s that interest in them emerged. Applying an electric field causes
them to align in the appropriate direction and thus either block or transmit polarized light
from an external source such as a so-called backlight. Applying different electric fields to
different regions, or so-called “pixels,” in a liquid crystal causes light to be either passed or
blocked by the different pixels and thus enables an image to be formed on a display. Finding
the appropriate liquid crystal materials along with materials for polarizers and color filters
took many years of scientific research in the 1960s, 70s, and 80s where these advances were
facilitated by the ability to use semiconductor manufacturing equipment to deposit and form
patterns in these materials. This semiconductor manufacturing equipment also facilitated a
change from so-called passive-matrix to active-matrix LCDs where cost reductions are now
largely driven by increasing the scale of this production equipment along with reducing the
thickness of the materials [19][20][21][22].
A variety of technological discontinuities have been envisioned by scientists and
engineers for television and other displays. One is the replacement of existing backlights,
which are so-called cold cathode fluorescent lights, with light-emitting diodes (LEDs) since
the LEDs are much thinner and have higher luminosity per watt than do cold fluorescent
lights. A second discontinuity is the replacement of the entire LCD with a display constructed
with organic light-emitting diodes (OLEDs). While most LEDs are made from semiconductor
materials, OLEDs are constructed from organic materials in which it is much easier to place
different color polymers on a single substrate (usually glass) using ink jet printing than to
place different color LEDs on a single semiconductor substrate. OLEDs are potentially
thinner, more flexible, and cheaper than are LCDs while the elimination of an external light
source enables OLEDs to use less power and have higher viewing angles than LCDs.
7
A third possible discontinuity is 3D television. 3D television requires a new form of
display and new content. Although 3D televisions can be constructed from either OLED
displays, LCDs, or even plasma displays, this paper focuses on LCDs since they are currently
the dominant form of television display. Two of the most common ways of displaying 3D
images with an LCD are time-sequential and auto-stereoscopic. The first method requires
special glasses that include an active or passive LCD display while auto-stereoscopic 3D
LCDs do not require glasses.
In both types of televisions, a key component is a LCD. One reason for using the term
“component” is to distinguish between components and systems in what can be called a
“nested hierarchy of subsystems.” Systems are composed of sub-systems, sub-systems are
composed of components, and components may be composed of various inputs including
equipment and raw materials [11][12] [13] [14]. This paper will just use the terms systems
and components to simplify the discussion. For example, a system for producing integrated
circuits (ICs) or LCDs is composed of components such as raw materials and manufacturing
equipment.
Other reasons for using the term “component” is that some components experience more
improvements in performance and cost than do other components [23] and when these
components impact strongly on the performance and cost of a system, rapid improvements in
such a “key component” can lead to rapid improvements in the cost and performance of a
system. For example, some argue that improvements in the cost and performance of
computers came directly from improvements in the cost and performance of integrated
circuits (ICs) [23][24] and that increases in the recording capacity of hard disks or magnetic
tape-based systems came directly from improvements in the magnetic recording density of
platters or tape [23] [25]. Taking this argument one step further, some argue that
improvements in ICs, magnetic platters, and magnetic tape led to discontinuities in the design
of computers, hard disks, and magnetic tape-based systems [16] [26]. These arguments are
8
consistent with Nathan Rosenberg’s argument that complementary technologies are often
needed to implement new discontinuities [27] [28] [29] [30].
There are several reasons why a specific component may incur more improvements than
do other components. These include a greater potential for: 1) improving the efficiency by
which basic concepts and their underlying physical phenomena are exploited; and 2)
geometrical scaling [31][32]. The first method includes finding materials that better exploit
the basic concepts and their underlying physical phenomena. Such materials have been found
for batteries [33], lighting [34], displays [19], vacuum tubes, ICs [35], and magnetic storage
[36] technologies. The realization and exploitation of each physical phenomenon that forms
the basis of these technologies required a specific type of material and finding the best
material has taken many years. The best material exploited the physical phenomenon more
efficiently than did other materials and this higher efficiency also often led to lower costs as
fewer materials were needed. For LCDs, there has been a search for liquid crystal materials
whose orientation better responds to electrical signals than do other materials where recent
searches have focused on crystals with fast response times in order to increase frame rates.
Geometric scaling refers to the relationship between the geometry of a technology, the
scale of it, and the physical laws that govern it. Or as others describe it: the “scale effects are
permanently embedded in the geometry and the physical nature of the world in which we live”
[32]. Some technologies such as integrated circuits (ICs) benefit from reductions in the scale
of specific features such as transistor gate length or metal line widths because these
reductions in scale lead to improvements in both cost and performance. For LCDs, the most
relevant form of scaling is increases in the scale of LCD substrates and their associated
production equipment where large LCD substrates, some are now greater than 10.5 square
meters, are cut into smaller panels for televisions and other electronic products [37] [38]. The
benefits from increases in this scale are somewhat similar to the large benefits that have been
experienced from increasing the scale of chemical and steel plants, engines, and oil tankers
9
[30] [31] [39] [40].
3. Methodology
This paper uses the concept of a nested hierarchy of subsystems, an understanding of a
3D television’s system and component, and Eisenhard’s [41] case study approach to
understand the timing of 3D television’s economic feasibility. In particular, Eisenhardt’s
notion of cross-pattern search was used to interpret data from a wide variety of technical
journals, trade magazines, and technical reports. First, the analysis focused on the factors
preventing 3D television from being implemented. Second, it looked at the components in 3D
televisions and in particular the ones experiencing improvements. Third, to what extent could
these improvements solve the problems that were preventing 3D television from being
implemented? Fourth, when would the components reach the levels of performance and cost
that were needed to make 3D television technologically and economically feasible?
Eisenhardt’s notion of cross-pattern search was appropriate because these issues were
addressed in a recursive method in which increasing levels of details were considered.
4. Improvements in 2D displays and other “components”
2D displays have been and are still being improved in a number of ways and as discussed
below, these improvements facilitate the implementation of 3D displays. First, recent LCD
televisions with LED (light-emitting diode) backlights yield a more comprehensive color
spectrum than do previous LCDs (and CRTs). Second, the sizes of LCD televisions have been
gradually increased and one reason that LCDs replaced CRTs is that the size of flat panel
LCDs can be more easily increased than can the size of CRTs. This is because increases in the
size of CRTs require increases in the thickness of glass, which increases cost and causes
image distortion around the edges of screen. Another advantage of LCD panels is that they
can be produced in a variety of rectangular dimensions and thus are more easily matched with
10
the widescreen format commonly used in films for cinema screens than are CRTs.
Third, increases in the substrate size of LCDs and other improvement in the LCD
production process have reduced the cost of displays. For example, the size of LCD
substrates used in making panels has grown by a factor of 1.8 every 3 years, and doubles
every 3.6 years where each doubling is often labeled by new generations of substrate sizes
and production equipment [42]. This doubling in substrate size is a major reason why costs
fall about 22% for each doubling of cumulative production in terms of area. The reason that
increases in the substrate size led to lower costs is that large LCD production equipment can
more quickly handle and process substrates on a per area basis and they have smaller “edge”
effects than do small equipment1 and thus have lower equipment costs per output than do
small equipment.
For example, the output (substrate area per hour) per dollar of capital costs for one type of
LCD manufacturing equipment was increased by 8.5 times as the substrate size was increased
by almost 16 times from 0.17 (Generation II) to 2.7 square meters (Generation VI) [37]. The
capital cost, this time for a complete facility, per area dropped by 36% as the substrate size
was increased from 1.4 (in Generation V) to 5.3 square meters in Generation VIII [39].
Generation XI panels are now being produced in sizes of 10.5 square meters. Furthermore,
the most important material in LCDs, glass, also benefits from increases in the scale of their
production equipment [37]. The result is that the average selling price of large LCD
televisions on a per meter squared basis dropped 18.8% a year between the first quarter of
1998 and the first quarter of 2007 [43].
Fourth, improvements in liquid crystal response time enable increases in frame rate (See
Figure 2) and these improvements facilitate the technical and economic feasibility for one
1 Like IC wafers, large LCD substrates have smaller edge effects than do small substrates since the LCD production equipment must be
wider than the substrate in order to have consistent processing across the substrate. Thus, the extra width of the production equipment as a
percentage of the substrate width declines as the width of the substrate is increased.
11
form of 3D television called time sequential with active glasses. These improvements are
driven by the use of new materials and the use of higher voltages [44]. By enabling faster
frame rates in both televisions and glasses, these improvements eliminate the blurring that
can occur when an image for the left-eye is followed sequentially by an image for a right eye
and when these images are assigned to each eye by glasses that also contain LCDs. As shown
in Figure 2, a rate of 120 frames per second was achieved in 2010 and this is considered the
minimum necessary to prevent blurring.
Fifth, improvements in pixel density (See Figure 3) facilitate the technical and economic
feasibility for a second form of 3D television called auto-stereoscopy where pixel count has
increased by four-times every 3 years [42]. These improvements are being driven by the use
of better photolithography and etching equipment that are being borrowed from the
semiconductor industry. By reducing the feature sizes for transistors and thus pixels, this
better equipment enables improvements in pixel density and thus resolution where increases
in resolution are needed to implement auto-stereoscopy. The reason these increases in
resolution are needed to implement auto-stereoscopy is because pixel elements on the display
need to be divided into ones for the left and right eye and for each necessary “viewing zone”
(See details below).
Sixth, improvements in ICs such as graphic processing units (GPUs) (See Figure 4) and
digital storage facilitate the introduction of 3D television because 3D images require more
data processing than do 2D ones. For example, stereoscopy requires the processing of two
stereoscopic streams of video and thus requires more data processing than do 2D television.
Furthermore, improvements in GPUs enable the use of more complex software algorithms for
converting existing 2D movies into 3D ones. For example, Samsung 3D television includes
software programs for converting regular 2D content into 3D [45].
The importance of improvements in GPUs is also relevant for animation where millions
of motion control points and polygons are used to represent images. One expert estimated
12
that the "reality threshold" (simulations indistinguishable from ordinary human vision) of
computer animation is about 80 million polygons per frame [46]. Polygon count generated by
leading video game hardware doubles roughly every 2 years. If these trends continue, the
reality threshold may be achieved by 2014.
Seventh, improvements in digital storage include improvements in magnetic (See Figure 5),
optical and semiconductor storage and these improvements enable the cheaper and more
compact storage of video. For example, while a DVD disc containing two hours of video
(720p, MPEG-2) contains 4 gigabytes of storage, a blue-ray disc containing 9 hours of HD
video has 50 gigabytes of storage. The result of these improvements is that the “street” price
per gigabyte of optical and magnetic hard disk storage had fallen below $0.10 per GB by
2010.
Eighth, improvements in Internet bandwidth facilitate the introduction of 3D television
because these improvements facilitate a move towards Internet downloads of both 2D and 3D
movies. According to Nielsen’s Law of Internet Bandwidth [47], the connection speed of
high-end Internet users grows by 50% annually, or double every 21 months. By 2011, Internet
bandwidth for high-end users would exceed 40Mbps, the average bitrate of high-quality 3D
HD video streams, enabling high-quality 3D-HD content to be distributed over the Internet.
Ninth, standardization of compression and digital broadcasting methods also facilitate 3D
television and this standardization is partly driven by the improvements discussed above. For
example, although many countries have adopted different forms of digital broadcast standards
(DVB/T, ATSC, ISDB-T, DMB-T/H), most of these standards include a common MPEG-4
video compression standard with an extension called Multiview Video Coding (MVC). This
means that most of the digital broadcast standards are “3D ready” [48] and since most
countries already broadcast partial or all in digital, a lack of agreement on standards will
probably not slow diffusion [49].
Tenth, since many of the above-mentioned improvements facilitate the introduction of 3D
13
content, together they facilitate a growth in 3D content. As shown in Figure 6, the number of
3D movies has grown quickly since 2005. Although most of these movies have been viewed
in theaters with special glasses, these movies can provide the first content for 3D televisions
as the televisions become economically feasible. Similar things can be found in video games
where video game consoles typically include faster graphic processing units than do personal
computers and thus 3D content is probably diffusing more rapidly with video games than
with 3d movies. For example, in late 2010, more than 500 3D PC games titles were listed on
Nvidia’s 3D Vision website [50].
5. Impact of improvements in 2D displays on achieving 3D displays
While various methods of implementing 3D displays exist, this paper focuses on two of
them: 1) Time-sequential 3D with active shutter 3D glasses (or passive ones); and 2.)
Development of “Glasses-free” auto-stereoscopic 3D displays. 3D televisions using the first
technique were introduced in 2010 and they sold 1.1 million units and they are expected to
sell just under two million units in 2011 [51].
5.1 Time-Sequential 3D
In time-sequential 3D, a special LCD and glasses containing a similar LCD are used to
create the illusion of 3D images for the viewer. Separate streams of images for the left and
right eyes are displayed sequentially on the display, i.e., a frame for left eye followed by
another for the right eye, and by synchronizing the LCD television and the LCD in the
glasses, the appropriate images are presented to the right and left eyes. In order to prevent
blurring, improvements in the frame-rate of 2D LCD displays were needed before this could
be achieved. As shown in Figure 1, the frame-rate of 2D LCD displays surpassed the content
frame rate of 120 frames per seconds in 2010.
The shutter glasses incorporate a liquid crystal that selects appropriate images for the left
14
and right eye. Faster liquid crystal response times were also needed for these glasses to work
effectively in that the active 3D glasses must quickly process the streams of stereoscopic
images to prevent blurring. The disadvantage of these glasses is that their estimated retail
price is about US$100 due to use of a display in the lens, an infrared or radio frequency frame
synchronizing circuit, and a power source to operate the shutter [52].
On the other hand, the falling cost of LCDs will probably reduce the cost of these glasses
and reduce the cost of the overall 3D television to a point at which the 3D television is
cheaper than current 2D televisions for the same size display. The estimated cost of adding
3D capability to an LCD television was about 10 to 30% in 2010 the cost of the television.
Since the cost of large screen LCD televisions fell 18.8% a year between 2003 and 2007 [43],
it is likely that 3D televisions will become cheaper than the current 2D televisions for the
same display size over the next few years even if this rate of price reduction falls.
Another option is passive glasses that do not require a power source and that are expected
to cost less than 10 USD. Although images for the left and right eye are displayed
sequentially on such a display as in time-sequential 3D, the images are polarized by an
additional active polarizing filter before leaving the LCD displays. These filters also depend
on improvements in the frame-rate of the displays so that the displays can process the images
fast enough to prevent blurring. Polarized glasses then filter the images for each eye thus
creating an illusion of 3D images. Polarize-filtered 3D glasses are smaller and lighter and
thus more comfortable and affordable to users.
5.2 Auto-stereoscopic 3D
Improvements in pixel density are needed to make auto-stereoscopic 3D displays
technically and economically feasible and thus eliminate the need for glasses. In
auto-stereoscopy, pixels are divided into two groups -- one for displaying left-eye images,
another group for displaying right-eye images. A filter element in the LCD is used to focus
15
left and right images into appropriate “viewing zones” where respective eyes of the observer
should be located, as shown in Figure 6.
Manufacturers estimate that more than 128 million pixels per square inch are needed to
make auto-stereoscopic 3D technically possible. This is because 8.3 million pixels are needed
to enjoy the full benefits of high-definition television and an auto-stereoscopic 3D television
should have about eight viewing zones in order to accommodate head movements. Since each
viewing zone requires two sets of pixels, about 128 million pixels per square inch are needed
before auto-stereoscopy 3D television is technically feasible.
The best auto-stereoscopic 3D display panel exhibited at the Consumer Electronics Show
in 2011 [53] had a pixel density of 8.3 million pixels per square inch. If pixel density
continues to increase four-times every three years, it will be two more cycles or 2017 before
pixel density reaches 128 million pixels per square inch and thus auto-stereoscopic 3D
displays become technically feasible. As for economic feasibility, this depends on the
incremental cost of the higher densities. If the incremental cost is small, they will probably
become economically feasible before 2020.
6. Diffusion of 3D Televisions
The cost of 3D televisions and both the cost of making and distributing of 3D content are
gradually falling and thus becoming more economically feasible as the cost and performance
of 2D displays, ICs, various storage technologies, and the Internet are gradually improved.
The cost points at which 3D televisions and 3D content begin to diffuse is more difficult to
analyze. The fact that 3D movies are popular in theaters, albeit that popularity has dropped as
the quality of 3D movies have fallen [50], suggests that many consumers would like to watch
3D programs in their homes. But how much do they want to watch them? And how much are
they willing to pay for these 3D televisions and content. These questions are difficult to
answer and are not addressed by this paper’s approach.
16
A second set of questions revolves around the interaction between the availability of 3D
programs and the diffusion of 3D televisions. A large body of research suggests that a critical
mass of users, content, and hardware must be created for growth to occur in those industries
in which strong network effects exist [54][55][56]. Will the availability of 3D movies for
theater viewing reduce the challenges of creating a critical mass by making a sufficient
number of movies available before 3D televisions begin to diffuse? The answer is probably
yes but to what extent? Furthermore, even if the availability of 3D movies for theater viewing
reduce these challenges, significant increases in the amount of 3D content are still needed for
most consumers to purchase 3D televisions. How fast might these increases occur?
A third set of questions revolves around whether users will be willing to wear glasses to
watch television. Surveys of consumers have found that users do not like glasses and worry
about eye strain, nausea, and fatigue [50]. Thus, the diffusion of 3D television may have to
wait for auto-stereoscopic ones. The second and third sets of questions are not addressed by
this paper’s approach
Nevertheless, it is likely that 3D television will diffuse over the next 5 to 10 years (from
2011). Improvements in frame rate, pixel density, and overall costs continue to be improved
since they are being made for 2D LCDs and thus these improvements will continue to be
made even if 3D television is not implemented. In other words, 3D television is benefiting
from spillovers from 2D displays and the economic feasibility of 3D television will continue
to improve even if 3D television is not implemented. In particular, as long as the prices of
large screen LCD televisions continue to fall through increases in substrate size, reductions in
material thicknesses, and other improvements, consumers will upgrade to larger screens,
higher-definition, and eventually 3D televisions. At the same time, as more movies and
television programs move to the Internet, it is likely that 3D displays will begin diffusing in
the personal computer market and thus drive a move to 3D content for many media.
17
7. Discussion
This paper analyzes the timing of a new technology’s economic feasibility using a simple
yet novel approach. While the conventional wisdom that the cost of a new technology falls as
cumulative production increases cannot address economic feasibility until a technology has
been introduced, this paper’s approach enables us to analyze economic feasibility long before
a technology has been introduced. This approach builds from the notion that technologies can
be thought of as a “system” of components [11][12] [13] [14] [15] where new technologies
often borrow components from existing technologies [16]. Thus, this approach focuses on the
concepts that form the basis of new technologies and the levels of performance that are
needed in the relevant components before new concepts become technically and
economically feasible.
For a 3D television “system,” key components include LCDs, ICs, hard disks, and the
Internet. Key dimensions of performance for them include the frame rate and pixel density of
the LCDs and various dimensions of performance for ICs, hard disk storage, and the Internet.
In particular, we were able to identify specific levels of performance that are needed in the
frame rate and pixel density of LCDs before time sequential and auto-stereoscopic 3D
televisions will respectively become technically feasible. Based on this analysis, we were
then able to identify other improvements in costs that are occurring and that might
compensate for the cost disadvantages that may result from the implementation of higher
frame rates and pixel densities.
This paper’s approach to understanding economic feasibility highlights two problems
with the economics literature’s use of learning curves. One problem is that many applications
of learning curves assume that all of the components in a new technology’s system are unique
to that technology and thus any cost reductions in these components come from production of
this technology’s system. This is clearly not the case in 3D television and in many other
electronic products such as computers and mobile phones [15][16][26][30].
18
A second problem is that many applications of learning curves assume that the cost
reductions come from activities in a factory [1] [2] [3].. Building from other research, this
paper considered several other ways in which 3D television related technologies are being
improved. These include: 1) improving the efficiency by which basic concepts and their
underlying physical phenomena are exploited; and 2) geometrical scaling. For the first
method, there has been a search for liquid crystals whose orientation effectively responds to
electrical signals where recent searches have focused on crystals with fast response times. For
geometric scaling, LCDs have benefited from increases in the scale of the production
equipment for them where increases in substrate size have facilitated the realization of these
benefits. In showing how these improvements are being made in LCDs, this paper also shows
some of the limitations with assuming that cost reductions are mostly from cumulative
production as automated equipment is implemented and organized into flow lines.
Some readers might argue that this paper’s approach is too simple. However, the author’s
argue that we need simple approaches that can be used by managers and students to
understand technological change and how this change leads to the emergence of
technological discontinuities. In particular, we need simple methods that students can use to
look for opportunities in the university courses that are intended to help students look for
opportunities. This paper (and a forthcoming book) provides a simple method for analyzing
the economic feasibility of new technologies and thus provides a simple method that students
can use to look for new opportunities. It demonstrates the method using an analysis of 3D
television, which is the type of near-term technology that we would like students to analyze.
Furthermore, by going beyond the notion that costs fall as automated equipment is
implemented and organized into flow lines and illuminating two other methods of achieving
improvements in cost and performance, this paper can help students and managers analyze
other technologies. A forthcoming book by the one of the author provides additional details
on this methodology.
19
8. References
[1] Arrow K, The economic implications of learning by doing, The review of economic
studies 1962; 29(3): 155-173.
[2] Argote L, Epple D. Learning Curves in Manufacturing, Science 1990; 247(4945): 920 –
924.
[3] Utterback J. Mastering the dynamics of innovation. Boston: Harvard Business School
Press, 1994.
[4] Henderson R, Clark K. Architectural Innovation: The Reconfiguration of Existing Product
Technologies and the Failure of Established Firms. Administrative Science Quarterly 1990;
35: 9-30.
[5] Anderson P, Tushman M. Technological discontinuities and dominant designs: A cyclical
model of technological change, Administrative Science Quarterly 1990; 35: 604-633.
[6] Utterback J. Mastering the dynamics of innovation. Boston: Harvard Business School
Press; 1994.
[7] Christensen C. The innovator’s dilemma. Harvard Business School Press. Boston, MA,
1997.
[8] Christensen C, Bower J. Customer power, strategic investment, and the failure of leading
firms. Strategic Management Journal 1996; 17: 197-218.
[9] Kaplan S, Tripsas M. Thinking about Technology: Applying a cognitive lens to technical
change. Research Policy; 2008: 37(5): 790-805.
[10] Dyer J, Gregersen H, Christensen C. 2011. The Innovator's DNA: Mastering the Five
Skills of Disruptive Innovators. Boston: Harvard Business School Press; 2011.
[11] Simon H. The Architecture of Complexity, Proceedings of the American Philosophical
Society 1962; 106(6): 467-482.
[12] Alexander C. Notes on the Synthesis of Form. Cambridge: Harvard University Press;
1964.
20
[13] Tushman M, Murmann J. Dominant Designs, Technology Cycles, and Organizational
Outcomes, Research in Organizational Behavior 1998; 20: 231-266.
[14] Tushman M, Rosenkopf L. On the organizational determinants of technological change:
towards a sociology of technological evolution, Research in Organizational Behavior,
Cummings, L. and Staw, B. (ed), Greenwich, CT: JAI Press 1992: 14: 311-347.
[15] Malerba F, Nelson R, Orsenigo L, Winter S. History-Friendly Models of Industry
Evolution: The Computer Industry. Industrial and Corporate Change 1999; 8: 3-40.
[16] Funk, J. Systems, Components, and Technological Discontinuities: The case of magnetic
recording and playback equipment, Research Policy 2009: 38(7): 1079-1216.
[17] Abernathy W, Clark K. Innovation: Mapping the Winds of Creative Destruction,
Research Policy 1985: 14: 3-22.
[18] Wikipedia, 2011a. http://en.wikipedia.org/wiki/Phosphor. Last accessed on October 6,
2011
[19] Castellano J. Liquid Gold: The Story Of Liquid Crystal Displays and the Creation of an
Industry, London: World Scientific Publishing Company, 2005.
[20] Numagami T. History of Technological Innovation in TFT-LCD Display. Tokyo: Hakuto,
1999 (in Japanese)
[21] den Boer W. Active Matrix Liquid Crystal Displays. NY: Elsevier, 2005.
[22] Gay C. Applied Materials: Accelerating Solar Power Cost Reduction, June 12, 2008.
[23] Kurzwell R. The Singularity is Near, NY: Penguin Books, 2005.
[24] Smith R. A Historical Overview of Computer Architecture, IEEE Annals of the
History of Computing 1988; 10(4): 277-303.
[25] Molstad R, Langlois D, and Johnson D. Linear tape servo writing enables increased
track density, Data Storage, 2002, online at:
http://www.imationltd.co.uk/products/pdfs/Network_Tape_article_servo.pdf
[26] Funk J. Components, Systems, and Technological Discontinuities: Lessons from the IT
21
Sector Long Range Planning 2008; 41(5): 555-573.
[27] Rosenberg N. Technological Change in the Machine Tool Industry, 1840-1910, The
Journal of Economic History 1963; 23 (4)P: 414-443.
[28] Rosenberg N. The Direction of Technological Change: Inducement Mechanisms and
Focusing Devices, Economic Development and Cultural Change 1969; 18 (1): 1-24.
[29] Rosenberg N. Some critical episodes in the progress of medical innovation: An Anglo-
American perspective, Research Policy 2009; 38: 234-242.
[30] Mowery D and Rosenberg N. Paths of Innovation. NY: Cambridge University Press,
1998.
[31] Winter S. Scaling heuristics shape technology! Should economic theory take notice?
Industrial and Corporate Change 2008; 17(3): 513–531.
[32] Lipsey R, Carlaw K, Bekar C. Economic Transformations. NY: Oxford University Press,
2005.
[33] Tarascon J. Key challenges in future Li-battery research, Philosophical Transactions of
the Royal Society 2010; 368: 3227-3241.
[34] Williamson R. Strategies for the Future of Lighting, Unpublished M.S. Dissertation in
Engineering and Management. MIT, 2010.
[35] Tilton J. The International Diffusion of Technology: The Case of Semiconductors,
Washington D.C: Brookings Institution, 1971.
[36] Daniel E, Mee D, Clark M. Magnetic recording: The first 100 years. NY: IEEE Press,
1999.
[37] Keshner M, Arya R. Study of Potential Cost Reductions Resulting from Super-Large
Scale Manufacturing of PV Modules, Final Report for National Renewable Energy
Laboratory (NREL/SR-520-36846), 2004.
[38] DisplaySearch, Flat Panel Display Market Outlook,
http://www.docstoc.com/docs/53390734/Flat-Panel-Display-Market-Outloo
22
[39] Axelrod L, Daze R. Wickham H. The large plant concept, Chemical Engineering
Progress 1968; 64(7): 17
[40] Mannan S. Lee’s Loss Prevention in the Process Industries, Vol. 1. Burlington,
MA: Elsevier Butterworth-Heinemann, 2005.
[41] Eisenhardt K. Building Theories from Case Study Research, Academy of Management
Review 1989; 14(4): 532-550.
[42] Multiverse, 2011. http://metaverseroadmap.org/inputs.html. last accessed on October 6,
2011
[43] Zhang B. Flat Panel TV Cost Analysis & Panel Supply-Demand, Presented at the 2008
DisplaySearch China HDTV Conference, May 20-21.
[44] See discussion of response time in the following Wikipedia entry.
http://en.wikipedia.org/wiki/LCD_television
[45] Samsung, 2011. http://www.samsung.com/us/learningresources3D/. Last accessed on
October 6, 2011.
[46] Smith A. Digital Humans Wait in the Wings, Scientific American, November 2000,
72-76, http://alvyray.com/Papers/CameralessMoviesv1.13.pdf
[47] Use it.com, 2011. Nielsen’s law of internet bandwidth,
www.useit.com/alertbox/980405.html. last accessed on October 6, 2011.
[48] Ozaktas H, Onural L. Three-Dimensional Television: Capture, Transmission, and
Display. Berlin: Springer-Verlag, 2008
[49] Digital Television Transition”, Wikipedia,
http://en.wikipedia.org/wiki/Digital_television_transition
[50] Nvidia, 2011. www.nvidia.com/object/3d-vision-3d-games.html. Accessed last on
October 6, 2011.
[51] Bonnington C. Poor 3D TV sales? Panasonic blames Hollywood, Wired.com, July 11,
2001. http://www.wired.co.uk/news/archive/2011-07/11/3dtv-sales-hollywood. Last
23
Accessed on October 6, 2011.
[52] http://www.xpand.me/shop/. Last accessed on October 6, 2011.
[53] 3D Radar, 2011. Toshiba to show off auto-stereoscopic screens at CES 2011, December
2010, http://3dradar.techradar.com/3d-tech/tosh-show-autostereoscopic-screens-ces-2011-
21-12-2010. Last accessed on October 6, 2011.
[54] Grindley P. Standards Strategy and Policy: Cases and Stories. Oxford: Oxford University
Press, 1995.
[55] Rohlfs J. Bandwagon Effects in High-Technology Industries. Cambridge: MIT Press,
2001.
[56] Shapiro C, Varian H. Information Rules. Boston: Harvard Business School Press, 1999.
[57] [http://en.wikipedia.org/wiki/Flicker_fusion]
[58] Nvidia, 2010. http://www.nvidia.com/docs/io/43395/tesla-brochure-12-lr.pdf
Accessed last on October 6, 2011.
[59] http://www.videophill.com/blog/?p=159, last accessed on October 6, 2011.
[60] Aubusson M, Teulade V. Eyes wide open – 3D Tipping Points Loom, Report by
PricewaterhouseCoopers, October, 2009
[61] Dudnikov S and Melnikov Y. Review of current technologies for 3D acquisition and
display, Chaos and innovation research unit, 2010. Last accessed on October 6, 2011.
http://www.docstoc.com/docs/53390734/Flat-Panel-Display-Market-Outloo.
24
Figure 1. Improvements in display frame rate. Sources: [44], [57], author’s analysis.
Figure 2. Increasing pixel density for LCD. Sources: [42] and author’s analysis.
25
Figure 3. Increasing performance of graphic processor units. Source: [58]
Figure 4. Reductions in price per gigabyte of hard disk drive storage. Source: [59]
26
Figure 5. Increased 3D Content. Source: [62]
Figure 6.Concept of Auto-Stereoscopic 3D Display. Source: [63]