The papers on this CD-ROM span a decade of research in color reproduction. It has been a turbulent decade. The cold war had come to an end a few years earlier and the mission to “out-brain” the scientists on the other side of the iron curtain became obsolete. The dwindling funding for research brought with it a demise of industrial research labs and the refocusing of academics to more applied research that can attract funding from industry. There has also been a shift to research areas where funding is still available plentiful, such as molecular biology, genomics, and bio-informatics.
In industrial nations, the baby-boom bulge came closer to retirement age and – now that life expectancy became years longer than the legal retirement age – people became concerned about life quality in retirement. This resulted in a surge of savings in search of good returns.
Concomitantly a change of course has occurred in corporate ethics and governance. While previously entrepreneurs were driven to a large extent by the desire to build businesses around new technologies or services in which they had expertise, in the past decade MBA programs in universities hatched an new breed of industry captains who were successful when they were skilled at harnessing the capital available in the savings sector and attracting investors. The telecommunications sector was one where this new trend was prominent.
When new companies can develop their business by attracting investors instead of laboriously creating new inventions in-house, established research organizations suffer because it is more economical to buy new technologies from graduating students. A strategy of mergers and acquisitions reduced risks and shortened the time span in which a new technology impacts positively the quarterly results.
While research as a whole has suffered a setback in the decade spanned in this CD-ROM, development has been very prolific, resulting in a plethora of labor-saving devices. As Marshall McLuhan noted in “Understanding Media – The Extensions of Man,” instead of saving work, labor-saving devices permit everybody to do their own work. In color reproduction, this has meant that professional authors have become able to do their own pre-press work, and the pre-press industry has essentially disappeared.
The large increase in storage space and processor performance for a given cost, made it not only possible to perform pre-press functions on ordinary desktop computers, but also to invent new publishing paradigms that were not possible with traditional electronic pre-press equipment.
The shift from R&D to customization and branding in the computer industry has resulted in the commoditization of desktop computers, at first through OEM (original equipment manufacturing), then through ODM (original design and manufacturing). Today, the most popular computer brand is “Other.” This trend was not followed to such an extent in color reproduction technology, where invention has continued unabated in the decade covered in this CD-ROM, allowing these editors to compile an extraordinary number of papers that have been instrumental, among others, for getting color printers from usable office machines to fine photo-imaging appliances, and for the invention of digital photography.
One last historical note is necessary to appreciate the developments that tie the papers in the collection on this CD-ROM. During the 1960s the generation of baby-boomers was imbibed with a culture of dialectic and the study of many point of views from which each individual would form his or her personal opinion. This culture and the necessity to digest a large amount of detailed information has fostered an information industry that has spun the gamut from the Encyclopaedia Britannica to hypertext through The Pentagon Papers, the MacNeil/Lehrer Newshour and Photo Retouching & Restoration For Dummies.
This contrasts the generation now entering the workforce, which has learned to play video games before reading (you learn to master video games by trial and error, not by reading manuals), has learned that information is found by searching on the World Wide Web, and that buddies can be contacted instantaneously via instant messaging or mobile phone. Instead of asking “did you consider their point of view?” interlocutors now demand “give me the 4-1-1.” Information must be concise, up-to-date, and through multiple rich media channels.
Research activity usually follows a pattern where an area suddenly becomes hot, often due a charismatic promoter or a concomitant historical event that can spark a new meme. Droves of researchers participate in the new area and a plethora of publications appears in journals and conferences. With this large collaborative effort the area then usually reaches a bifurcation point, when all the low hanging fruits have been harvested and two new collective efforts take over. The first one consists of a small number of researchers that attacks the very hard problems remaining after the easy ones have been solved. The second one consists of an army of developers who create commercial applications of the new technology. In our conferences we have tried to identify new potentially hot areas and give them a forum.
The early papers on this CD-ROM are to a large extent on color hardcopy technologies, with relatively few color and image processing papers. As research in color printer technologies reached the bifurcation point, we re-focused our conferences on software and science, redirecting hardware papers towards either the more specialized conferences or towards a more applied conference when they were more application oriented.
This did not prevent us from missing one important topic – color appearance modeling, on which this collection contains only a few papers. At the time, we thought that having past the bifurcation point, the proposed models were sufficiently sophisticated that research on the models belonged on more specialistic conferences, while the applications were too straightforward for the Electronic Imaging Symposium. In retrospect, we were wrong in this assessment and color appearance models are today an active area of research and the practical applications are not obvious.
A trend we did not miss was the Internet tsunami. Although there are numerous papers on this topic, most progress was actually done in the panel discussions, which unfortunately were not recorded (conference proceedings are never a substitute for personal attendance). In essence, over the years these editors and the panelists concluded that printing and publishing can be broken into four eras: manual, mechanical, electronic, and digital. We took it as one of our duties to nourish the revolution from electronic to digital publishing.
Manual printing refers to a process in which each step a single copy of the publication is reproduced, and for an additional copy the same process must be repeated in its entirety. An example are the scribes of ancient times.
Mechanical printing refers to the preparation of a master or matrix from which multiple copies can be printed and distributed. A first step was the carving of a page in a wood block, which can be repeatedly inked and used to print multiple copies, such as in 868 for the Chinese “Diamond Sutra.” A second step was the 1454 invention of movable type and the mass-production of type. From then on the emphasis was on improving workflows.
Electronic printing and publishing refers to the substitution of mechanical steps with electronic means to increase the efficiency of workflows, which themselves remain unaltered. Note that printing refers to the process of fabricating a publication, while publishing – invented 1500 – encompasses the editorial process to decide what to publish, print planning, as well as marketing and distribution of the publication. When the distribution is electronic, the “printed” artifact can be a CD-ROM, a file transmitted over TCP/IP, an e-mail message, or a posting on the World Wide Web, so we prefer to switch from using “printing” to using “publishing,” because research in color can no longer focus on the printing process but must also take into account the other publishing aspects. An example are the issues of color reproduction across media.
Digital publishing refers to the complex process of acquiring and maintaining a corpus of information consisting of document chunks. This corpus is in a raw “unrendered” state. When a need for information arises, a story is woven from the chunks and printed or “rendered” on the fly in the most appropriate medium. As we have seen earlier, this is not only made possible by new technology, but also demanded by today’s new adults.
The concomitant requirement of instant concise information, the popularization of Xerox PARC’s Bravo through Microsoft’s ever-evolving Word, and the revolution from expensive clunky typewriters to inexpensive photographic quality ink jet printers, have brought about the relegation into obsolescence of the whole pre-press industry, in a time shorter than the big bang and without a whimper, by allowing everybody to do their own copy. Today we have blogs and digital publishers have the opportunity to harvest and distill valuable content from this amalgam of undigested brain dumps.
In the collection on this CD-ROM you can find a continuous stream of papers informing on standards. During this decade the industry has transitioned from the concept of locking in customers while locking out competitors – which required proprietary technology and research laboratories to invent it – to commoditisation, which requires standardization and interchangeable suppliers to allow competition in speed to market and price.
Similar to standards, gamut mapping is a topic that has remained a favorite over the decade spanned in this collection. In his seminal 1943 JOSA paper, R.M. Evans introduced the consistency principle stating that what is important is the relation among the colors in a reproduced image, not their absolute colorimetry. The first gamut mapping paper in this collection builds on the consistency principle and the latest papers describe the use of Retinex for gamut mapping. Yet a general solution still is elusive.
The increasing computer power available on desktop computers, paired with progress in sensor technology, has allowed to evolve image acquisition from device counts through CIE colorimetry and principal components analysis based multiband sensors to end up with spectral color reproduction at the end of the decade. Today the main applications are in art conservation and forensic imaging, with general applications still awaiting a killer application.
Especially for art conservation applications, the next step has been the incorporation of geometric appearance, which is particularly important for the study of oil paintings. Other applications are the imaging of archeological artifacts like tablets and coins.
The collection of papers on this CD-ROM concludes with the contributions to the mini-symposium on Retinex at 40. This technique has been very controversial, because there is not yet a psycho-physiological basis, the algorithm tends to fail catastrophically, and it was not well known how to implement it. Today only the last issue has been solved, and it has taken most of the decade. Digital photography with its dynamic bandwidth problems has been a strong motivator for unabated research on Retinex and today a number of Retinex-like algorithms is part of commercial solutions, forming the basis for heuristic automatic image enhancement.
Last but not least, we have papers on digital photography. At the onset of the decade digital cameras used video sensors and recorded a video signal, limiting their application to claims adjusters and similar fields where image quality was not important. The workflow for all other digital imaging started with a traditional AgX image that would get scanned onto a PhotoCD disk and then processed digitally from there on. About half-way in this collection the first true digital cameras became available, but their image quality was still well inferior to AgX. The hard research problem was to follow the path of early AgX technology from mosaics to layers, thus avoiding the image reconstruction problem, a path you can trace in this collection.
We have mentioned only a few of the threads you can weave in this CD-ROM. An example of an area not mentioned is halftoning, where the papers are of high quality and progress has continued at a steady pace.
We already dwelled on the progressive decrease of papers on printers during this decade. Today printers – especially printer consumables – are still a very profitable business. However, the writing is on the wall – printing is gradually fading away and going the way of pre-press. Part of this is generational, as elucidated above, and people no longer need printed reference material.
The other part is that commercial printing is manufacturing. The manufacturing of printers and computers has moved to the low wage regions of the world, and the printing business is following. The artwork can easily be transmitted to the print plant over the available Internet, and efficient logistics companies can get the printed artifact to the point of use quickly and inexpensively.
One form of printing that will survive is the personal printing that is performed to take advantage of the convenience of paper as a portable medium.
CRTs are slowly disappearing, ending an era of 60 cd/m2 devices that require careful calibration and well controlled dim viewing conditions for critical color work. At this writing, LCD displays achieve a brightness of 300 cd/m2 and we can expect it to grow probably to about 500 cd/m2. The next displays on the horizon are based on OLEDs, which will have twice the brightness. Finally, the following technology available today is based on iridescence obtained with MEMS, where the display brightness increases with the ambient illuminance, and resolution reaches 1,000 dpi.
As mentioned in Evans’ paper cited above, with increasing brightness the appearance mode of the human visual system switches to film mode and the color constancy mechanisms take care of color fidelity, as long as color consistency is preserved. Understanding Retinex is becoming more important, as is research in phenomena like the perception of transparency.
There is a large amount of dark fiber all over the world that awaits to be lit. Digital Cinema is around the corner and the media conglomerates cannot wait for video on demand to become viable. The hardware and software companies are ready to provide the rendering devices, the imaging software, and the services to build and maintain these networked systems.
Digital images like to be copied, and their transfer likes to be tracked. The ensuing issues of digital rights management (who gets paid for what, who can decrypt an image) and privacy (what personal data can the sender collect on the receiver) are still largely unsolved. ISO’s MPEG-21 is an effort to build a framework addressing these issues.
The digitalization of imaging does not necessarily lead to a world dominated by a few large conglomerates owning all contents and enslaving image consumers in feudal servitude. In fact, digital imaging allows any talented cinematographer to build a professional movie production studio for a few thousand dollars. The same distribution channels are available as for the conglomerates.
While conglomerates strive to become virtual corporations with few permanent employees, digital imaging allows everybody to become an independent imaging professional and make a good living in a world of peers.
The key is to spread the knowledge and create an educated society. Despite the Internet, knowledge is still transferred most efficiently face to face, so while this CD-ROM you are holding in your hands allows you to leverage on past research, to become successful in the future there is no way around attending SPIE conferences on color imaging and network with other scientists.
[ order information | imaging | EI | inventoland ]