JPEG XR

When approved, the new JPEG compression system submitted earlier this year will be called JPEG XR.

As I mentioned in my 29 April 2007 post on MPEG-A, that week Microsoft had submitted HD Photo as a new possible JPEG standard. Earlier this morning Microsoft sent out a press release announcing that JPEG introduced a new work item for this specification and its tentative name is JPEG XR. In truth, this is actually old news, as JPEG already made this announcement in their 6 July 2007 press release after their 42^nd meeting at the EPFL in Lausanne. If you want to stay tuned, the next date is 16 November 2007, when the press release on the 43^th meeting in Kobe will come out.

If you are a JPEG delegate or have access to one, the relevant working draft document is ISO/IEC JTC 1/SC 29/WG1 n4336, or wg1n4336 for short.

As I mentioned in my 8 March 2007 post, JPEG XR is somewhere between JPEG and JPEG 2000. It is a file format and associated codec specifically designed to for use with all types of continuous tone photographic content. The codec has some reminiscences of MPEG-4 AVC (Advanced Video Coding, a.k.a. ITU-T H.264), like the use of 4×4 blocks. However, the main innovation is the extended range for the numerical encoding of the pixel values.

In the beginning of digital color imaging standardization, the Xerox Color Encoding Standard adopted the concept of arbitrary color gamuts that was already in use in earlier standards for digital monochrome imaging. Such an encoding was used also in the ISO ODA Color Addendum and in color fax (ITU-T T.42, T.81). However, then in 1994 a less brilliant decision was made with sRGB to cast in concrete a specific device gamut based on a byte encoding.

In 1994 that not-so-smart ad hoc decision was plausible from an engineering point of view because at that time digital cameras in the consumer market had only 6 bits of information, and the rest was noise. CRTs typically had 8 bits per channels, and to use a 10 bit CRT you had to have a very clean power source.

Today the situation is quite different. In the consumer market, digital reflex cameras use sensors that have 12 bits of information per channel, and also typical desktop LCD panels have 12 bits of information per channel. Today we use algorithms like retinex (e.g., digital flash) to compress the sensor's dynamic range into 8 bits, but then when the LCD panel's controller logic expands the data again to 12 bits, those 4 bits are gone, and we see contouring.

The overall goal of JPEG XR is to support the greatest possible level of image dynamic range and color precision, maintain forward compatibility with existing formats, and keep the device implementations of the encoder and decoder as simple as possible.

To maintain full compatibility with existing legacy devices and applications, JPEG XR supports a variety of color profiled pixel formats using unsigned integer representations in bit depths of 8 and 16, as well as smaller bit depths for specialized applications. Additionally, JPEG XR also supports a number of advanced pixel formats that avoid many of the limitations and complexity imposed by unsigned integer representations.

Specifically, JPEG XR supports three types of numerical encoding, each at a variety of bit depths: unsigned integer, fixed point, and floating point.

Advanced Human Eye Models

New progress in personalized computer eye models to simulate an individual's vision quality using optical and mechanical engineering software.

If you attend the AIC meetings, you are certainly aware that in the last decade or so there has been a lot of research in Europe and Korea, but especially in Japan, on the the visual effects of aging. For example, at the AIC Meeting in Granada, Dr. Tomoko Obama of Matsushita's Panasonic Design Company in Yokohama presented a revolutionary pair of goggles [pages 38-41] that allow researchers to personally experience age-related conditions of the visual system like cataracts.

Here in the U.S., the country of the young, of the celebrities, and of the sociopaths, where the Constitutional right to happiness translates to supermen living in reality distortion fields, we have not seen much research on the age related issues of vision. Therefore it is refreshing to read about Dr. William J. Donnelly's news article Software advances human eye modeling.

Dr. Donnelly's work at the Breault Research Organization involves combining biometry, gradient index crystalline lenses, spectacles, contact lenses, intraocular lenses, and intraocular scatter into monocular and binocular eye model systems, culminating in the Advanced Human Eye Models (AHEM) modeling capability.

Dr. Donnelly writes "Advanced Human Eye Models can provide simulations of eye systems including ophthalmic and external optics. They can also predict visual performance with retinal image analyses inclusive of the effects of aberrations, diffraction, and scatter."

Read the article at this link.

Emotional visual memories

Research at the University of Colorado in Boulder indicates that emotional visual memories can be suppressed.

About two decades ago here in northern California we had a few high-profile trials involving persons with a sudden recall of sexual abuse during childhood. A researcher who subsequently showed how false memories can be created, had her life threatened and was persecuted — an indication of how controversial research on the manipulation of memories is.

I do not remember the details, but at that time research indicated how visual memories are not cast in concrete in the form of fixed synaptic connections. Instead, visual memories are constantly relived or revisited, e.g., in the slow wave phase of sleep. A recall entails reliving the memory, i.e., retrieval of memories entails the same regions in the brain that are involved in encoding a memory. False memories can be created by repeatedly inserting false images during this recall process.

The new research at the University of Colorado in Boulder unveils how emotional visual memories can be suppressed via two time-differentiated neural mechanisms:

1. an initial suppression by the right inferior frontal gyrus over regions supporting sensory components of the memory representation (visual cortex, thalamus), followed by
2. right medial frontal gyrus control over regions supporting multimodal and emotional components of the memory representation (hippocampus, amygdala), both of which are influenced by fronto-polar regions.

A paper on this research has been published in Science 13 July 2007.

According to the researchers, their findings may have implications for therapeutic approaches to disorders involving the inability to suppress emotionally distressing memories and thoughts, including PTSD (posttraumatic stress disorder), phobias, ruminative depression/anxiety, and OCD (obsessive-compulsive disorder). They provide the possibility for approaches to controlling memories by suppressing sensory aspects of memory and/or by strengthening cognitive control over memory and emotional processes through repeated practice. Refinement of therapeutic procedures based on these distinct means of manipulating emotional memory might be an exciting and fruitful development in future clinical research.

Chromatic discomfort

Some interesting new research at the University of Essex in Colchester shows how narrow stripes of complementary colors can induce discomfort.

Several times I wrote about the readability of colored text, and in unrelated posts on how for example a flickering display in the periphery of the visual field can cause discomfort. Prof. Arnold Wilkins in the psychology department at the University of Essex in Colchester has researched this class of problems for decades.

For example, regarding readability he has studied how colored overlays can reduce visual stress and increase reading fluency. He has a Web page devoted to this topic.

Last year, Arnold Wilkins completed a joint project with artist and migraine sufferer Debbie Ayles on the discomfort induced by complex color images. The scientific results are presented in a poster. In summary,

• uncomfortable images have an 'un-natural' curved power spectrum
• they tend to have greater power at about 3 cycles/degree
• urban images have slightly more power at about 3 cycles/degrees

Colored stripes with such spatial frequencies

• are uncomfortable to view
• give perceptual distortions
• can provoke eye strain, nausea, headaches, even seizures

Risk Management

This is a summary of the feedback I received on my previous blog entry.

I always had two feedback channels for this blog: the open comments and the personal email. After the system problems in the first half of June, which prevented you from posting comments and me from blogging, the comment channel has closed up. However, the personal email still keeps coming.

Posting comments works again and I much prefer comments to personal email, because this is supposed to be a conversation — with personal email it just looks like I am pontificating. In this post I summarize the feedback I received on the last blog entry, hoping further feedback uses the comment channel, which is up and running.

The emailers agreed among them that I had it all wrong. They observed that Christoph Hauert et al. had the typical European view of a social state, which is proven to be inefficient. Readers pointed out at how efficient the American businesses have become and therefore that one has to study them to learn how to build a successful society.

Using the terminology in Hauert's paper, cooperators tend to become inefficient over time. Defectors are the good guys who identify these inefficiencies and by exploiting them become the punishers of the lazy cooperators and build a better society. Examples given included hedge funds and takeover partnerships like Steel Partners in the Bull-Dog Sauce case.

The question then is about the morality of defectors punishing cooperators to create a better society. The answer to that was that every cooperator has to have a risk management strategy in place to be useful to society. Hence, the risk manager should replace the abstainer proposed by Hauert et al. An example of risk management in business is the poison pill, but the emails I received mentioned this as a bad example, without giving me good examples.

Anyway, what does risk management mean for us in color science? We do research, which is about taking risk. None of the emails I received touched this topic.

Some classical strategies to manage risk include transferring the risk to an outside party (i.e., out-source research to a university), and reducing the negative effect of the risk. The risk is that the invention cannot be transferred into a product, hence researchers should only take on projects proposed by marketing as a result of a customer study. In our jargon we call this going only for the low-hanging fruit, and the associated tactic is to stick to incremental product improvements.

Practically this means we should stop research in solid state light sources and focus instead on improving the color rendering index of fluorescent and halogen discharge lamps. Similarly, instead of spectral imaging, color Fourier imaging, or high-dymanic-range imaging we should keep improving look-up table methods, add parameters to ICC profiles, and measure more color patches.

Do you have a better understanding you can share?

Contributors, slackers, quitters, and Bull-Dog Sauce

If you saw the news on Japanese TV yesterday, the top story was the Bull-Dog Sauce take-over and specifically the morality of the High Court decision in this matter. It just so happens that the latest Science issue has a paper on a new model of collective action showing how socially beneficial punishment can arise and evolve. Color science relies heavily on mathematical models, so this is an interesting case for us.

In well oiled successful societies, there is a strong sense of community and solidarity. Periodically, this sense of community can weaken and a subculture of social parasites or slackers (scientific term: defectors) can emerge an grow. Society can then punish the slackers to remain competitive.

A historical example was the aristocracy in XVIII century France. When at the end the illuminated created a philosophy and Dr. Joseph-Ignace Guillotin popularized an efficient punishment tool, the problem was tamed. We can explain this superficially by arguing the contributors got fed up and boiled over, but this is not a scientifically tenable explanation, because punishers pay a social price and are not motivated.

Worldwide, from Bucharest to Tokyo, but especially here in the US we see a rapid increase of these defectors. This takes many forms, from insurance fraud to frivolous litigation, elder abuse, family member blackmail, phishing, pharming, and identity theft. Some are legal and some are not, but the common thread is that currently the defectors usually are not punished. At the moment there is no effective punishment mechanism and the situation is worsening, so there is a collective interest in understanding how punishment evolves.

Christoph Hauert et al. have published new research results in their article Via Freedom to Coercion: The Emergence of Costly Punishment in the 29 June edition of Science, who also published the perspective A Narrow Road to Cooperation in the same issue. The editors summarize the research as follows:

Collective endeavors among individuals are often accompanied by risk. Defectors (those who do not invest but who share in the return) fare better than cooperators (who do invest), but a third type of participant, the punisher, who acts against the defectors, can stabilize a cooperative group of individuals. The Science paper now now provides a theoretical basis for the emergence of such punishers, who incur costs that mere cooperators do not and would thus be expected to suffer in evolutionary terms. Allowing for a fourth type of individual — the abstainer — leads to population dynamics where punishers flourish. In essence, it appears that voluntary submission to social norms is a prosocial act.

The mathematics behind this paper is elegant and beautiful, but does a slick mathematical model entail an accurate and useful description of reality? Can it explain the Bull-Dog Sauce debacle? Can it differentiate between the Japanese perception of Bull-Dog Sauce as victims versus the American perception of Bull-Dog Sauce as whiners and loosers?

Personally, this paper leaves me with more questions than answers. How about statesmanship, idealism, passion, etc.? How about the current population pressures of flourishing societies with rapidly decreasing birth rates versus less efficient societies with high birth rates?

And what does it mean for us in research? I do not believe large collaborative research efforts in research labs have disappeared because of defecting researchers. And the abstainers, are those the researchers who are unemployed and working out of their Silicon Valley garages? I am not even sure abstainers are socially relevant, as an abstainer would still need a good connection to a venture capitalist to fund his work and eventually sell it.

What are your thoughts? How does this affect your research, your personal life?

Cinema al fresco

Our yearly heat wave is coming, with temperatures predicted to reach 30º C in Palo Alto. A popular escape is to hit the road for the foggy beaches, but with the concomitant 4^th of July holiday, staying at home may be a safer choice. Here is an HP way to weather the heat wave at home…

A favorite spot to stay cool is to hide in an air conditioned movie theater. However, since the beginning of cinematography, in Mediterranean countries it has been popular to enjoy movies al fresco in the gran piazza. You may have seen it in Cinema Paradiso or you may even have attended the famous International Film Festival in Locarno, which celebrates its 60^th birthday this year.

Here in the Silicon Valley quite a few people own portable video projectors they use to give presentations when they are on the road visiting customers. On hot evenings many of them check out a movie at the library and enjoy it with their families in the back yard.

The latest latest news story on HP Labs' Web site teaches you how HP technology allows you einjoy your al fresco movies in grand style, on a 16-foot screen and with bright vivid colors. HP Labs' Nelson Chang and Niranjan Damera-Venkata had a hunch they could make a cheap projection system by combining the outputs of several smaller projectors to create a single, high-quality image.

Essentially you invite a dozen friends who own an HP display projector and have them bring their projectors along. All you do is then is to use your HP digital camera to gang together all the projectors using a version of HP's resolution enhancement technology, and voilà, you can flood your 16-foot screen with the combined photons of all the projectors.

Here is the link again, and do not forget the ice cream!

The end of PDF

This post echoes my 8 March 2007 post The end of JPEG, this time pointing to the standardization of print (or paper) specifications. I am reacting to last Friday's post on Andy Updegrove's Standards blog about Microsoft's submission of its XML Paper Specification (XPS) to ECMA International.

PDF is a direct descendant of Interpress, a device independent page description language originally designed at PARC and released in 1980. Although I am linking into Wikipedia, the information there is inaccurate: JaM was indeed based on Forth, and PostScript is a direct descendant of JaM, but Interpress is not really a direct descendant of JaM and PDF is not really a direct descendant of PostScript.

What do I mean by that? Although the same people where involved, JaM and PostScript are philosophically very different from Interpress and PDF. The former was a top notch and very elegant engineering effort by John Warnock to create a real world device independent page description language to replace the device dependent Press page description language.

After JaM was successful, PARC's computer language gurus that were behind Cedar did a clean-sheet design of a device independent page description language carried out with the methodology of specifying a computer language. Although is took advantage of the JaM experience it's author's collaboration, Interpress is much cleaner, more powerful, and more efficient. The main driving forces behind this effort were Butler Lampson and Bob Sproull.

In my view, PDF is more the sibling of Interpress than the son of Postscript, and I believe it is mostly the merit of Ed Taft that PDF has remained a pure and clean language to these days.

Those were the Seventies. If PDF would be specified today, it would very likely be as one of the XML languages. In fact, if one considers the evolution of PDF over the releases, one can say that it is somehow converging to XML.

When they were done with Interpress, what problem did PARC's computer language gurus tackle? In those days there where three computer systems in use at PARC — Cedar, Interlisp, and Smalltalk — and each had their own incompatible document preparation system. Therefore, there was a need for a system independent editable document description language, which would do for WYSIWG editors what Interpress did for printers. The outcome was Interscript and the main authors were Bob Ayers, Jim Horning, Butler Lampson, and Jim Mitchell.

Unfortunately, 1983, when Interscript was finished, was a very turbulent year at PARC and Interpress fell between the cracks. It lingered along in the Spinnaker project and at INRIA Sophia Antipolis, eventually strongly influencing ODA (Open Document Architecture), but there have not been pure-bred descendants with immaculate pedigree.

Although there was no "Ed Taft" for Interscript, through the years it has left its mark on SGML and XML. And this is where I would like to make my point. If a new standard language is proposed today, the inspiration should not come from Interpress but from Interscript.

Useful links:

• Andy Updegrove's blog post
• Microsoft's XML Paper Specification (XPS) site
• Microsoft's 2006 document interoperability press release
• Standard ECMA-376