Division 1: Vision and Colour

DIVISION 1: VISION AND COLOUR

MINUTES of the 1st Meeting of the Luo Term

COMMISSION INTERNATIONALE DE L'ECLAIRAGE Sunday 15 June 2008 INTERNATIONAL COMMISSION ON ILLUMINATION INTERNATIONALE BELEUCHTUNGSKOMMISSION Stockholm, Sweden

ATTENDANCE

Officers Ronnier Luo GB DD – Director Miyoshi Ayama JP AD Vision Ellen Carter US AD Colour Mike Pointer GB DS – Secretary

Country Peter Hanselaer Belgium Representatives Juliana de Freitas Gomes Brazil Marjukka Eloholma Finland Françoise Viénot France Gerhard Rössler Germany Michael Pointer Great Britain Janos Schanda* Hungary Miyoshi Ayama* Japan Chang Kim Soon Korea (Republic) Peter van der Burgt Netherlands Thorstein Seim* Norway Elsie Coetzee South Africa Siv Lindberg Sweden Rengin Ünver Turkey Paula Alessi USA *Nominated representative

Technical Paula Alessi US TC1-27 Committee Françoise Viénot FR TC1-36 Chairs Ken Sagawa JP TC1-37, TC1-54 Miyoshi Ayama JP TC1-42 Robert Hirshler HU TC1-44 Taiichiro Ishida JP TC1-61 Klaus Richter DE TC1-63 Janos Schanda HU TC1-66 Hiroyasu Ujike JP TC1-67 Peter Bodrogi HU TC1-68 Wendy Davis US TC1-69 Changjun Li CN TC1-71 Mike Pointer GB TC1-72

Reporters Mike Pointer GB R1-39 Boris Oicherman IS R1-41, R1-43 Changjun Li CN R1-42

Guests In addition there were approximately 8 guests present.

Apologies Mike Brill TC1-56 Andrew Chalmers NZ Vibeke Clausen DK Gunilla Derefeldt R1-32 Marta K Gunde SI Jack Holm R1-40, L1-5 Eugenio Martinez-Uriegas TC1-60 Sharon McFadden TC1-64 Manuel Melgosa ES, TC1-55 Inna Nissenbaum IS Danny Rich L1-6 Alan Robertson TC1-57 Michael Stock L1-2 Ian Tutt L1-8 Klara Wenzel HU Joanne Zwinkels L1-3

Total attendance: Approximately 32 persons

1. WELCOME The Division Director, Ronnier Luo, welcomed all those present and thanked the hosts, the Swedish Colour Institute, for providing facilities for the meetings.

2. ATTENDENCE 15 countries were officially represented. However not all representatives were present for the whole of the Division meeting.

3. MEMBERSHIP The following changes in national representative were noted: China: Guan-Rong Ye Finland: Marjukka Eloholma France: Françoise Viénot Hong Kong: Patrick Wong South Africa: Elsie Coetzee Spain: Manuel Melgosa Sweden: Siv Lindberg

4. CONFIRMATION OF THE AGENDA It was agreed that Agenda items 14 (New Work Items) and 15 (Liaison Reports) would be changed round. A copy of the amended agenda can be found as Appendix 1.

5. MINUTES The Minutes of the 2007 meeting held in Beijing, China, were approved with one change: it was noted that the Netherlands’ representative, Peter van der Burgt, was present.

6. MATTERS ARISING FROM THESE MINUTES There were no matters arising not covered by items on the agenda.

7. DIVISION OFFICER REPORTS

7.1 Director The Director highlighted the following points: · CIE Standard S 014-4/E:2007 Colorimetry - Part 4: CIE 1976 L*a*b* Colour Spaces has been published in September 2007.

· The book Colorimetry: Understanding the CIE System is now available from Wiley.

· Three new TCs were approved by the BA following the 2007 Beijing meeting:

TC1-70 (C) Metameric Samples for Indoor Daylight Evaluation Chairman: Balázs Kranicz, HU Terms of Reference: To investigate the derivation of a set of metameric samples to enable the evaluation of indoor daylight simulators.

TC 1-71 (C) Tristimulus Integration Chairman: Changjun Li (CN) Terms of Reference: To investigate methods for computing weighting tables for the calculation of tristimulus values from abridged data.

TC 1-72 (C) Measurement of Appearance Network: MApNet Chairman: Mike Pointer GB Terms of Reference: i. To establish a network of those interested in the measurement of visual appearance. ii. The network shall be under the direction and guidance of a group of at least four Technical Leaders each responsible for a particular aspect of the subject. iii. Each Technical Leader shall provide substantial periodic reports in a form that might be published. iv. A second Expert Symposium on Appearance shall be organised at an appropriate time within the next 4 years. v. A database of relevant published work shall be maintained. vi. Consideration shall be given to the establishment of separate Technical Committees when appropriate.

· Five new Reporterships were approved following the 2007 Beijing meeting:

R1-41 (C) Adaptation Transforms: B. Oicherman IS Terms of Reference: To investigate and report, in one year, on the state-of-the-art of adaptation transforms.

R1-42 (C) Extensions of CIECAM02: C Li CN Terms of Reference: To evaluate potential additions to CIECAM02 in liaison with Division 8 and to include: Those published in the literature; Extension to include unrelated colours; Extension of the range down to scotopic levels.

R1-43 (V) Standard Deviate Observer: B. Oicherman IS Terms of Reference: To document available databases that could yield a definition of a new standard deviate observer.

R1-44 (V) Limits of Normal Colour Vision: S. McFadden US Terms of Reference: To review the literature to see what information is available to establish the limits of normal colour vision.

R1-45 (V) Luminous Efficiency Functions: Y. Nakano JP Terms of Reference: To provide definitions and tables of the existing functions Vb,point(l), Vb,2(l), and Vb,10(l).

· The next CIE Midterm Meeting will be held in May-June 2009 in Budapest, Hungary.

· The next CIE Session will be held in July 2011 in Sun City, South Africa.

7.2 Editor In the absence of the DE, the DS reported that the Activity Report had been produced in January 2008. In addition the DE had reviewed the TR from TC1-66 (Indoor daylight illuminants) and the report from R1-39 (Alternative forms of the CIEDE2000 colour difference equation).

7.3 Secretary The main actions of the DS since the last meeting were as follows: · Produce and distribute the Minutes of 2007 Beijing meeting. · Establish a definitive list of country members and agree this list with the CIE CB. · Establish a Division membership list including officers, country members, TC chairs, reporters, liaisons. · Move and maintain the Division website. · Contribute to and distribute the 2008 Activity Report. · Set up the 2008 Stockholm Division meeting. · Deal with approximately 400 emails!

8. TECHNICAL COMMITTEE REPORTS

8.1 Vision: Miyoshi Ayama

TC1-36 Fundamental Chromaticity Diagram with Physiologically Significant Axes: Françoise Viénot FR Part 1 of a TR has been published as CIE Publication 170-1:2006. Part 2 of a TR is being prepared as CIE publication 170-2, which will contain the following chapters: Ch.7. The best spectral luminous efficiency function; Preliminary proposal based on a new experiment: * += mlaV lll )()()(

l l)( and m l)( are the CIE L- and M-cone fundamentals, a is a weighting factor to express the ratio of L and M cones for individual observers. Ch.8. Development of 2-dimensional chromaticity diagrams (x, y) and (l, s) Ch.9. Conclusions and recommendations Ch.10. Tables of data

One member has submitted a manuscript for publication in a journal that proposes an XYZ representation of the cone fundamentals. Once this paper has been published, it will be possible to complete this work.

TC1-37 Supplementary system of Photometry: Ken Sagawa JP A draft TR was prepared at the beginning of June 2008, and will soon be distributed to members and those others concerned (TC1-58) for comments.

TC1-41 Extension of VM(l) Beyond 830nm: Pieter Walraven NL No report. It was agreed that the AD Vision will write to the chairman setting a deadline for completion of this work – if it is still considered necessary.

TC1-42 Colour Appearance in Peripheral Vision: Miyoshi Ayama JP Year Established: 1993 The first draft of a TR will be ready in July 2008: it is 80% complete at the present.

TC1-54 Age-Related Change in Visual Response: Ken Sagawa JP The TR is to be merged into the CIE Guidelines for Accessibility which contains vision data and design guidelines for better visibility and lighting for older persons and persons with disabilities.

TC1-58 Visual Performance in the Mesopic range: Liisa Halonen FL The fourth draft has been distributed to TC members, and was discussed in the TC meeting on June 14 where comments were discussed and actions were explained. Additional references need to be added.

TC1-60 Contrast Sensitivity Function for Detection and Discrimination: Eugene Matinez-Uriegas ES Updated tasks, assignments, and status are as follows:

· Introduction/philosophy of the approach of the technical report. Beau Watson (s), Mark Fairchild (s), Eli Peli (s), Eugenio Martinez-Uriegas Status: Under discussion, 75% complete · Search for data and conditions: Sharon McFadden, Chaker Larabi, David Alleysson (s), Jose M. Artigas (s), Chein-Chung Chen (s) Status: Ongoing, 75% complete · Preparing tables of data and conditions: Sharon McFadden, Lindsay McDonald (s), Hirohisa Yaguchi (s), Sophie Wuerger (s) Status: Ongoing, 50% complete · Combining data sets – analysis and review of data: Victor Klassen, Beau Watson (s), Eli Peli (s), Thom Carney (s) Status: Ongoing, 50% complete · Set up new Web tool for TC1-60 workspace (eRoom was dismantled because of cost cuts). Eugenio Martinez-Uriegas Status: Ongoing, no progress since Dec 2007 · Combine all inputs to update and circulate the TR draft: Eugenio Martinez-Uriegas Status: Ongoing, 50% complete

So far we have got reviews and data updates from Victor Klassen, Chen Chien-Chung, Sophie Werger, and Lindsay MacDonald.

We are currently working on integration of these contributions in the 3rd, and hopefully final, draft of the TC1-60 Technical Report.

TC1-67 The Effects of Dynamic and Stereo Visual Images on Human Health: Hiroyasu Ujike JP A TR is being prepared separately for PSS (Photosensitive seizures), VIMS (Visually Induced Motion Sickness), and VFCS (Visual fatigue caused by stereoscopic images). Contents of the TR for PSS are as follows, 1. Introduction and Background 2. Definition of terms 3. Prevalence and incidence 4. Type of seizure 5. Evaluation methods 6. Visual stimuli 6.1. Flash 6.2. Pattern 6.3. Patho-mechanisms

7. Counter measures 7.1 Reduction of stimulus 7.2 Rationale for the recommendations 7.3. Counter measures: reduction of patients’ sensitivity 8. References

R1-19 Specification on Individual Variation in Heterochromatic Brightness Matching: Hirohisa Yaguchi JP The report has been submitted to the editor and is being reviewed for publication in the CIE collection.

R1-23 Guidelines on Planning a Mesopic Photometry Investigation: Pat Trezona GB No report. It was agreed to disband this reportership: 15 affirmative, 0 negative, 0 abstain.

R1-35 Irregularity ybar10(l): Pieter Walraven NL No report. It was agreed to disband this reportership: 15 affirmative, 0 negative, 0 abstain.

R1-36 Action Spectra for Glare: Judith Fekete HU Introduction One of the special features of human vision is the ability to see well in lighting that ranges from moonlight to bright sunlight: almost a thousand-million-fold change in light level. It is no wonder that the eye remains past comparison by any physical detector of light in yielding appropriate vision over a range that bridges about nine units on a logarithmic scale.

As humans obtain approximately 90% of information by means of visual perception, good visibility is particularly important when driving a motor vehicle. Poor visibility conditions result in a lack of information for drivers.

Night-time driving is a situation that takes place under changing mesopic conditions. Vehicle headlamps illuminate a small part of the visual field, and the headlamps of approaching vehicles modulate the state of adaptation of the driver continuously in a very unpredictable way, as well as causing glare for the driver.

The present report summarizes literature data (in 2007) on the interrelationship between visibility and produced glare.

Papers presented during the past year During the last ten years gas discharge (HID) lamps became popular. These HID headlamps provide visual benefits, chiefly due to their increased light output. Recently, spectrally-tuned high intensity discharge (HID) sources were introduced for use in vehicle forward lighting systems. One study dealt with this subject and has shown that spectrally-tuned HID sources can result in a decreased proportion of short-wavelength content toward oncoming drivers that can reduce discomfort, but will not impact on disability glare equal illuminance at the eye. This study has also shown that spectrally-tuned HID sources can result in increased illumination and increased short-wavelength content that can reliably improve peripheral visibility measured through target detection.

Comfort and visibility characteristics of spectrally-tuned high intensity discharge lighting system J. V. Derlofske, J. D. Bullough, C. Gribbin

Some authors put a question: Are there any correlations between individual brightness efficiency and glare? In their attempt they found that chromatic type observers are likely to be more sensitive to glare than the others.

Study on glare of LED lamp and individual variations of brightness perception Takako Kimura-Minoda, Yudai Fujita, Shinichi Kojima, Miyoshi Ayama 7th International Symposium on Automotive Lighting, Darmstadt, Germany, September 25-26, 2007.

There are very few investigations on the best spectrum for mesopic visibility and even less data on the spectrum of discomfort glare. Some authors have studied these two perceptions and came to the conclusion that it is not enough to optimise for the combined photopic and scotopic luminosity functions but the chromatic components have to be taken into consideration too. They tried to determine the optimum SPD for mesopic conditions providing good visibility and low glare.

Optimizing spectral power distribution of car headlamp lighting J. Fekete, G. Várady, C. Sik-Lányi, J. Schanda CIE 26th Session, Beijing 4-9 July 2007. 56-59. p.

Visibility and glare at mesopic light levels J. Fekete, G. Várady, C. Sik-Lányi, J. Schanda Conf. Illuminat 2007, Cluj-Napoce-Kolozsvár, 31 May – 1 June 2007.

Spectral power distribution optimization of headlights J. Fekete, G. Várady Lux et Color Vespremiensis, Veszprém, 2007. október 19.

Some articles have given a short overview about actual results to mesopic vision and the consequences for new installation of road lighting. In these articles it has been shown that it is indisputable that white light has psychological and physiological benefits and is therefore recommended. On the other hand, one has to realize that a decrease of light level leads to an increase in threshold contrast for foveal objects. So it is necessary to consider a concrete situation and with the help of luminance pictures to decide whether the street illumination should be changed or not.

With this knowledge, and the recommended measurement techniques, a reply can be given to the question as to whether all sodium lamps should be replaced by lamps with white light in public lighting.

Mesopic vision – Should we replace all sodium lamps by lamps with white light in our public lighting? S. Völker In Proceedings CIE Congress 2007, Beijing

Do light sources with a high part of shorter wavelength promote safety? S. Völker In Proceedings ISAL 2007, 24-26. September, Darmstadt

The reporter will be asked to either write the 1st draft of a report, or to continue the review for a further period, and then proceed to the next stage and make a recommendation for further work.

R1-37 Definition of the Visual Field for Conspicuity: Nina Itoh JP The previous studies of various functions of the visual field have been investigated. Those visual fields were analyzed and classified into several functional fields such as: 1. Detection and Perception,

2. Discrimination, 3. Recognition, 4. Performance/Behavior. A part of the revised report was presented at the Division 1 meeting in Beijing, China. Work on the full report is ongoing.

R1-38 Concept and Application of Equivalent Luminance: Yasuhisa Nakano JP This reportership was formed when TC 1-46 was closed during the meeting of Division 1 in León, Spain in May 2005. There has however, been no update since that time although a report had been written. It was agreed to pass the report to TC1-37 for harmonization with and possible inclusion in their technical report and to close this reportership.

R1-40 Scene Dynamic Range: Jack Holm US A preliminary report was submitted to 2007 Activity Report of Division 1. The DS will contact the reporter about future activities.

R1-43 Standard Deviate Observer: Boris Oicherman IL A report is being prepared. The reporter showed the structure of the report which overviews data, methods, and literature concerning observer metamerism. The draft of the report includes the following chapters: 1. Executive summary 2. Variability of colour matching functions 3. Observer metamerism and real-world metamers 4. Optical modelling of observer metamerism

R1-44 Limits of Normal Colour Vision: Sharon McFadden CA A report on the information available to establish the limits of normal colour vision has been prepared and submitted to the AD Vision for submission to the Division 1 members at their 2008 meeting: it is attached to these Minutes as Appendix 2. The recommendations in the report are as follows: · Continue the review for another year to more thoroughly evaluate the available data. · Contact researchers, clinicians and practitioners working in the field of colour vision deficiency to determine if they have relevant unpublished data and if they would be interested in contributing to a Technical Committee on this topic.

R1-45 Luminous Efficiency Functions: Yasuhisa Nakano JP No report. It was agreed to disband this reportership: 15 affirmative, 0 negative, 0 abstain.

8.2 Colour: Ellen Carter

TC1-27 Specification of Colour Appearance for Reflective Media and Self-Luminous Displays: Paula Alessi US Much of the work of this TC was completed several years ago and passed on to Division 8. However, the technical report has just recently been completed and is being checked by one of the Division officers. It will then be sent to the TC members for ballot. It is hoped that it will be published before the next Division meeting.

TC1-44 Practical Sources for Daylight Colorimetry: Robert Hirschler HU The second draft report was ready for the last meeting of the TC (Beijing, 2007). It was 51 pages long with 40 figures and 11 tables, and consisted of the following chapters: 1. Daylight simulator technologies

2. Daylight simulators for visual assessment 3. Daylight simulators in colour measuring instruments 4. Conclusions, recommendations Annexe 1. Standard methods for the evaluation of daylight simulators Annexe 2. Standards for the classification of daylight simulators

It contained the following recommendations:

1. Competing technologies a. Colour matching booths - The current most widely used filtered tungsten + UV and fluorescent lamp technologies already produce acceptable daylight simulators for D50 and D65 (claimed also for D75). - Manufacturers are urged to exploit more the potential of filtered xenon short-arc lamps, because very high quality simulation would be achievable, and the SPD of the source in the booths may be more similar to that used in spectrophotometers. b. Spectrophotometers - For spectrophotometers there is very little ground to recommend any source other than the currently nearly exclusively used pulsed xenon lamps.

2. Methods for the evaluation of daylight simulators The formation of a new TC is recommended - “to intercompare standard and other published methods for the evaluation of daylight simulators, with particular reference to the rendering of (UV or visible excited, white or coloured) fluorescent specimens, and select or develop a method considered adequate in all respects; and - to develop standard methods for the adjustment and verification of the UV/visible balance in colour matching booths and in single-monochromator spectrophotometers.”

3. Recommendation of standard daylight sources - The standardisation of any particular source (as “best representing daylight”) is not recommended.

Changes suggested for the 3rd Draft:

1. Minor technical details Changes recommended by TC members will be made.

2. Editorial changes 2.1 Several TC members considered the report far too long: the 3rd Draft will be significantly shorter. 2.2 Annexe 1. (Standard methods for the evaluation of daylight simulators) and 2. (Standards for the classification of daylight simulators) may contain too detailed description of the standards themselves and thus may violate copyrights. In the 3rd Draft only a very brief description of the standards will be included.

Schedule for the 3rd Draft Ready to be circulated within the TC: 15 August 2008 Final considerations by TC members: 30 September 2008 Sent to Division Editor: 31 October 2008

TC1-55 Uniform colour space for Industrial Colour-Difference Evaluation: Manuel Melgosa ES Unfortunately, the chairman’s request for existing experimental datasets on color

differences, which was published in Color Research and Application 32, 159 (2007), has not received any response. Recently the chairman has contacted by email different TC members to encourage submissions of new experimental datasets. At the present time it seems that, in addition to the four datasets used in the development of CIEDE2000 (BFD-P, Leeds, RIT-DuPont and Witt), we will be only able to add the dataset obtained by Quiao & Berns at RIT (which was used for the development of the T function in CIEDE2000: CR&A 23, 302-313, 1998), and perhaps only one or two new datasets.

A meeting of CIE TC 1-55 will be held on June 12, 2008, during the Fourth European Conference on Color in Graphics, Imaging and Vision, CGIV 2008, Barcelona (Spain). We hope that at this meeting the available experimental datasets to be used by CIE TC 1-55 will be definitely agreed.

The four experimental datasets employed in the development of CIEDE2000 have been reviewed and, after some minor correction (mainly in the old RIT-DuPont dataset), they are now available to TC members at Dr. Melgosa’s website http://www.ugr.es/~basapplcolor/

Different members have produced significant results related to the goals of our TC. For example (without being exhaustive): · the proposal of the DIN99 color spaces and corresponding color-difference formulas (G. Cui et al. CR&A, 27, 282-290, 2002), · the CAM02 formulas (M.R. Luo et al. CR&A, 31, 320-330, 2006) and another formula based on CIECAM02 (Berns et al, Proc. AIC 2007, pp. 24-28, 2007), · the GP formula based on OSA-UCS color space (R. Huertas et al. JOSA A 23, 2077-2084, 2006), · and the STRESS index, as a new alternative tool to the PF/3 index (P.A. Garcia et al., JOSA A 24, 1823-1829, 2007). Recent analyses carried our by Drs. Oleari, Berns, Huertas, and Melgosa seem to indicate that there are relevant differences amongst the four experimental datasets employed in the development of CIEDE2000. Discussion on these topics and other advances will be held during our forthcoming TC meeting in Terrassa. ## TC1-56 Improved Colour Matching functions: Michael Brill US CIE TC1-56 met on 9 July at the 2007 Beijing CIE meeting (attended by M. Brill, R. Luo, A. Robertson, B. Oicherman, and 21 guests). At that meeting, Boris Oicherman gave a special presentation summarizing his color-matching PhD thesis work at the University of Leeds. Based on the three experiments undertaken on behalf of the objectives of TC 1-56 (By Drs. Oleari, Oicherman, and Nakano), the TC chairman set the objective to complete the final report of the committee by the CIE Mid-Term Meeting of 2009. That report would summarize the data and provide a recommendation. Tentatively, that recommendation was to continue to use Grassmann additivity throughout colorimetry, except at low luminance levels for which mesopic color mechanisms must be considered.

After the Beijing meeting, the committee had a brief but vigorous email discussion about the tentative recommendation. Except at low luminance, the recent publications provide evidence that averaging of intra-observer trials restores additivity and transformability of primaries. However, it remains for CIE TC1-56 to understand the apparent discrepancy of the spectrum locus as derived from Maxwell with maximum-saturation color matches, as reported by Wyszecki and Stiles (W&S) in Section 5.6.6 of Color Science (2nd edition, 1982). Intra-observer data seem to have been averaged there, and the luminances were not low. The only distinctions I see between the maximum saturation and Maxwell procedures are (a) only Maxwell matching keeps the comparison field constant, hence potentially “nulling out” certain visual variations; (b) Maxwell matching at each wavelength involves two matches to

the comparison field, which might compound random error; and (c) Maxwell matching always occurs at the chromaticity of the (white) comparison stimulus, but maximum-saturation matching occurs (as the name would suggest) at maximum saturation. It appears that Grassmann additivity cannot be maintained in the face of the discrepancy. However, Alan Robertson noted that the data in W&S Section 5.6.6 basically came from a single observer (although a very good one), so confirmation might be in order.

[Note: To introduce the mechanics of Maxwell color matching, I attach a tutorial (Appendix 3 to these minutes) deriving equations relating light measurements to Maxwell-match color-matching functions. Toward applicable theory, W&S’s Section 5.3 relaxes Grassmann’s laws via conditions I-III (p. 295) to accommodate Maxwell but not maximum-saturation matching. The attached Appendix 4 provides thoughts toward a front-end visual process satisfying these laws.]

Further email discussion ensued in January-February 2008. Boris Oicherman said that cross-media matching data from his Ph.D. thesis, soon to be available at: http://www.oicherman.com/Boris_Oicherman_Phd_thesis.pdf, confirm failures of Grassmann additivity. Danny Rich’s experience in cross-media color matching confirms that the match is asymmetric and therefore is not logically tied to symmetric-match additivity. Therefore, Boris’s data may suggest (but do not mandate) a conclusion about non-additivity, and we will examine them. Alan Robertson suggested that, to model additivity failures, it may be necessary to pursue the possibility that a color match may be broken by adaptation that uses more than the usual three inputs (e.g., from a remote part of the retina). We are probably not going to arrive at a definitive model in TC1-56, but may do so in a later effort.

Rolf Kuehni emphasized that we must be about the task of writing a report so as to bring the activities of TC1-56 to a graceful conclusion---within a year, per encouragement by CIE Division 1 leaders. We must tell a story that places non- additivity data in perspective, weighted by the relevance of that data to unambiguous interpretation in terms of Grassmann failure.

One open question is to relate the Maxwell-Maximum-Saturation incompatibility reported by Wyszecki & Stiles (Section 5.6.6) to Boris's general statement that observers use more blue light than their CMFs predict when matching broadband light by a narrowband triplet. In addition, we must not forget that two studies (by Nakano and Oleari) have confirmed additivity after all. How can we tell a consistent story with such disparate ingredients?

Another open question is to propose criteria for data and replication to be acceptable results for a recommendation by a CIE committee like TC1-56.

We will continue to discuss these matters, and then distill a consensus position from which the report can emerge.

TC1-57 Standards in Colorimetry: Alan Robertson CA The Standards being written by TC 1-57 will be parts of the S014 series as follows:

CIE S014-3 Colorimetry - Part 3: CIE Tristimulus Values CIE S014-4 Colorimetry - Part 4: CIE 1976 L*a*b* Colour Space CIE S014-5 Colorimetry - Part 5: CIE 1976 L*u*v* Colour Space CIE S014-6 Colorimetry - Part 6: CIE DE2000 Colour Difference Formula

The TC decided to work first on Parts 4 and 5 (CIELAB and CIELUV) because these were expected to be less controversial than Parts 3 and 6.

Part 4 (CIELAB) was the first to be completed and was published in September 2007 as CIE Standard S 014-4/E:2007.

A third draft of Part 5 (CIELUV) was approved by the TC in June 2007 and sent to the Central Bureau for further processing. It was then approved by the Division and the Board of Administration in January 2008. There were no negative votes but there was a comment as follows: “do not want to hinder the publication of this series of standards, but from my understanding what a CIE standard should be, the L*, u*,v* part does not come up to the mark. This part is not used in the industry anymore, and should not be endorsed in the form of a standard. Only the u',v' part is of value.” As this issue had already been discussed and resolved within the TC, no action was taken. The latest draft (DS 014-5.2/E:2008) has now been sent to National Committees for comments with a deadline of 5 August 2008.

A second draft of Part 3 (Tristimulus values) was sent to the TC in January 2008. Comments were generally favourable but a third draft will be needed. Most of the changes relate to abridged methods for 10- and 20-nm intervals where care must be taken to follow the recommendations of CIE Publication 15:2004 and not to pre-empt the work of TC 1-71 (Tristimulus integration). The new draft has been delayed to allow time for discussions within TC 1-71 but the way is now clear to write it very soon. This third draft will refer to the weighting-function method for 10- and 20-nm intervals but will state clearly that this method produces results that may differ from the standard method (1-nm interval, 1-nm bandwidth) and that (for the time being), it is the user's responsibility to determine suitability for their purpose. We will refer to published work (ASTM, Li, Ohno etc) but only as "informative" references, not as "normative" references. Hopefully, the work of TC 1-71 will pave the way for a second edition of the Standard in the future which will deal with "abridged" input data. The related work of TC 2-60 (Effect of instrumental bandpass function and measurement interval on spectral quantities) is also being monitored to be sure that the Standard does not conflict with potential recommendations of that TC.

The final work of the TC will be Part 6 (CIEDE2000). A first draft will be produced within a few months.

TC1-61 Categorical Colour Identification: Taiichiro Ishida JP This TC met in Stockholm. We reviewed the last meeting in Beijing where data were presented using the CIECAM02 colour appearance model as opposed to the previous use of CIELAB and Munsell space. It was agreed that work would start on a technical report describing the application of colour categorisation for surface colours at different levels of illumination (1000 lx and 1 lx), using both Munsell and CIECM02 to present the data.

The proposed table of contents of the technical report is: Chap 1. Introduction Chap 2. Psychophysical studies on color categorization (Review) 2-1 Studies on color category 2-2 Major parameters Chap 3. Categorical colors at different levels of illuminance (Kyoto Univ.) 3-1 Methods 3-2 Results 3-3 Data comparison Chap 4. Categorical colors in color spaces 4-1 Munsell color space 4-2 CIECAM 02 Chap 5. Summary Appendix table of the data

The psychophysical studies on color category are:

No. authors (year) method category color sample test size surround light level light source subject 200W Boynton, monolexemic summarized by 3.8 x 3.8 cm (≈ 40 cd/m2 at N5 1 424, OSA ≈ N5 photoflood 7 Olson (1987) color naming 11 basic colors 4 deg) bkgnd lamp, 3200K Uchikawa, 200W monolexemic summarized by 3.8 x 3.8 cm (≈ 40 cd/m2 at N5 2 Boynton 424, OSA ≈ N5 photoflood 10 color naming 11 basic colors 4 deg) bkgnd (1987) lamp, 3200K Sturges, approximated monolexemic summarized by 5.2 x 3.3 cm 3 Whitfield 446, Munsell ≈ N7 1000 lx CIE Standard 20 color naming 11 basic colors (65cm) (1995) Illuminant D65 photopic, 1.4, ≈N6 separated tungsten lamp Middleton, categorical R, Or, Y, G, B, 64 (V≈6) 0.44, 0.14, 4 2 deg by dark with filter, ≈ Mayo (1952) color naming P Musell 0.044, 0.014, 3 (1) surroud 6500K 0.005 lx high color Ishida et al. selecting 11 basic colors 5 x 5 cm (≈4 1000, 10, 1, 0.1, rendering type 5 8, Munsell Black 4 (1995) color chips + YG, BG deg) 0.03, 0.01 lx fluorescent lamp ≈ 5300K high color categorical 11 basic colors 3 x 3 cm (≈ 1000, 10, 1, 0.1 rendering type 6 Ishida (1999) 156, Munsell ≈ N5 15 color naming + YG, BG 2.5 deg) lx fluorescent lamp ≈ 5300K Segawa, fluorescent categorical 7 Uchikawa 11 basic colors 205, OSA 4 deg ≈ N5 2000, 5, 0.1 lx lamp, D65 3 color naming (1998) type presented on dark, gray CRT, 145 for Shinoda et al. categorical 5 x 5 cm (≈ (aperture 40, 30, 20, 10, 8 11 basic colors each luminance --- 3 (1993) color naming 2deg) mode, surface 5, 2 cd/m2 level on mode) average D65, EX-L, EX-N, W, R100, G100, Yaguchi et al. categorical 9 11 basic colors 292, Munsell 5.5 x 7.0 cm N5 1000 lx G55R45, 4 (1999) color naming G80R20, H, HF, MHL, NH, NX, IL Segawa et al. categorical fluorescent 10 11 basic colors 205, OSA 4 deg N5 2000, 5, 0.5 lx 3 (1998) color naming lamp high color Ryuchi, Ishida categorical 11 basic colors 1000, 10, 1, 0.1 rendering type 11 76, Munsell 4, 2, 1, 0.5 deg N5 5 (2000) color naming + YG, BG lx fluorescent lamp ≈ 5300K 44(18- fluorescent grouping 26) 12 Sagawa (2000) 286, Munsell 4.4 x 6.1 deg Black 500, 0.5 lx lamp similar color 45(60- Tc=4000K 76)

We then discussed the contents of the technical report and saw some examples of how the data from different researchers could be combined onto the same map. It was suggested that we add a new Chapter 5 Summary to the technical report. The other suggestions were: to use H instead of h in data descriptions and to show the boundary of each category. It was noted that CIECAM02 could not be used for the 0.1 lx data.

The work plan is to write a draft of the technical report and circulate it via email to the TC members. The documents (presentation, TR draft, and data) will be uploaded on to the chairman’s website: http://www.users.kudpc.kyoto-u.ac.jp/~f52168/. To open the files enter the password “colourcat”.

TC1-63 Validity of the Range of CIEDE2000: Klaus Richter DE The last meeting of CIE TC1-63 was here in Stockholm on June 14, 2008.

Activity of CIE TC1-63 The committee decided in 2005 to produce test charts with large equal differences in

CIELAB. The CIELAB differences of adjacent and separated patches of the test charts * are mostly in the range ∆E ab = 10 to 40 with a maximum of 140. The same original test charts have been available for free. Only relative colour differences on a relative visual scale of adjacent and separated colours have been studied. One example test chart is shown in Appendix 5 of these minutes. At least four member countries, DE, GB, ES and CZ, agreed in 2005 to produce visual data and four sets of result are now available. Three other members (Schanda, Nakano, Yaguchi) received the test samples but have not submitted experimental results up to now.

Results from different countries for large colour differences: At the moment, only qualitative statements for the results of the four countries were given. For large colour differences (∆E* > 10) the correlation between visual results and CIELAB is better than compared to CIEDE2000 according to 3 countries (CZ, DE, ES). The results of GB have shown that for large colour differences there is only better performance for CIELAB if the samples are divided in two groups with 10 < ∆E* < 40 and 40 < ∆E* < 140. There are publications from Kittelmann (2005), Vik (2007), Melgosa (2007, 2008) and Luo (2007) with the experimental results from the four countries DE, CZ, ES, and GB. R Luo and P Kittelmann gave reports at the TC meeting about their recent results for large, medium and threshold data. According to R Luo CIEDE2000 performs best for the overall region.

According to the terms of reference, TC1-63 should investigate the application of the CIE DE2000 equation at threshold, up to CIELAB colour differences greater than 5. Therefore new experiments at threshold were started three years ago. P Kittelmann has produced many experimental results at threshold as a function of many parameters, including ﬁeld size and sample distance; these data will be available at the end of 2008 when his PhD thesis is to be published. A special procedure of adding coloured light to one half of the sample allows Kittelmann to avoid the hairline effect at the border of adjacent colours. According to Klaus Witt, 20% of the observers report a colour differences if the samples are identical. This produces problems for the evaluation of the experimental data and this effect is excluded by the Kittelmann procedure (no gloss differences and no hairline geometric difference).

The table below shows some preliminary results. If the weighting factors in the red-green and yellow-blue visual processes are optimized then the Stress values S100 (normalized to 100) increase to a large degree (from 60 to 80). If this optimization is used then at threshold CIELAB performs better compared to CIEDE2000.

In the literature, an effect known as small-field (yellow-blue) tritanopia is described. At about 4 minutes (0.06°) it is not possible to distinguish colours in the yellow-blue direction of equal luminance. The reason is that the (blue) S cones are rare in the retina, so the spatial discrimination in the yellow-blue direction is limited. One paper shows a reduction of the yellow-blue contribution to the colour difference by a factor 2 at 30 minutes (0.5°) compared to 2°.

In diploma work, S Lander (2008) has studied visually a 9 step colour series between white and yellow and the complementary series between black and violet-blue. The colour difference between two adjacent steps was in the region of approximately ∆E* = 8. All steps appear as a continuous series and the single steps disappear at about 1° viewing angle. This indicates that we must study the colour difference as a function of viewing angle.

CIEDE2000 was developed for a viewing angle of 10°. For image application, the viewing angles are often less than 2°. The experimental results of Lander indicate that the yellow-blue discrimination may be reduced by a factor 2 or more for 2° compared to 10°.

For a better correlation between visual results and calculations there are already two weighting factors in the CIEDE2000 formula. In the following table, two weighting factors, α and b for red-green and yellow blue, are added in the CIELAB formula. This produces an increase in the Stress index S100 from about 60 to 80. This indicates a better description for threshold data of adjacent colours.

Correlation of experimental data and colour difference formula by index S100 Calculation with standard parameters Calculation with optimized parameters Stress index Name and value of Stress index Name and value of Formula S100,s parameters S100,o parameters CIELAB 55 a = 1 ß = 1 80 a = 0,52 ß = 0,15 CMC 57 l = 1 c = 1 71 l = 0,42 c = 2,42 CIE94 59 KC = 1 KH = 1 71 KC = 4,43 KH = 2,03 CIEDE2000 61 KC = 1 KH = 1 74 KC = 2,95 KH = 3,18 DIN99 68 kE = 1 kCH = 1 77 kE = 1,76 kCH = 1,95 DIN99o 59 kE = 1 kCH = 1 75 kE = 0,78 kCH = 3,44 LABJNS 60 a0 = 1 b0 = 1,8 81 a0 = 2,52 b0 = 0,61

TC1-64 Terminology for Vision, Colour and Appearance: Sharon McFadden CA The latest version of the ILV has been distributed for BA ballot. The updated ILV will be very different in appearance from the current version. The numbering system used in the existing version will no longer exist; all the terms will be listed alphabetically. Moreover, in addition to a paper version, the ILV will appear on the CIE website. Given the new formats and the desire to keep the ILV more current than in the past, the BA hopes to develop a new process for reviewing existing terms and adding new terms. To be prepared for this process, the members of TC1-64 are currently reviewing a list of new terms for possible inclusion in the next update to the ILV. The list includes terms proposed during the Division 1 ballot of the current vocabulary and new terms which have appeared in recent Division 1 technical reports. If anyone else is interested in reviewing these terms or would like to submit additional terms for inclusion in the ILV, please contact the Chair.

TC1-66 Indoor Daylight: Janos Schanda HU I am glad to inform you that finally we got to the stage that our Technical Report has been sent to Division 1 and the Board of Administration for ballot. Votes and comments are due by 2008-08-05.

TC1-68 Effect of Stimulus Size on Colour Appearance: Pete Bodrogi HU It was suggested to change the Terms of Reference from: “To compare the appearance of small (2°) and large (>20°) uniform stimuli on a neutral background.” to: “To compare the colour appearance of small (2°) and large (>20°) uniform stimuli on a neutral background or with no background.”

It was argued however, that “colour” was implied by the title of the TC and the application to include no background was beyond the scope of the present TC. Thus no changes were approved.

Large Self-luminous colours • Draft 1 of the Technical Report has been prepared and discussed with TC members. • 1 paper appeared in CR&A: G. Kutas, P. Bodrogi, Color Appearance of a Large Homogenous Visual Field, Volume 33, Issue 1, Pages 45-54, February 2008.

• 1 further paper was accepted in CR&A: Kaida Xiao, et al., Colour Appearance Prediction for Room Colours.

Illustration 1: Large Surface Colours

Table of Contents of the Technical Report 1. Introduction 2. Concept of Equivalent Colour to Describe the Appearance of Large (>20°) Uniform Stimuli 3. Visual Experiments on the Colour Size Effect 3.1 Brightness and stimulus size…. 3.2..... 3.3..... 3.4 Xiao et al. (reflecting surfaces, room colours) 3.5 Kutas et al. (self-luminant colours) 4. Mathematical Models for the Colour Size Effect 4.1 Reflecting surfaces (room colours, Xiao et al.) 4.2 Self-luminant stimuli (large displays, Kutas et al.) 5. Recommendations of the Technical Committee 5.1 Concept of equivalent colour 5.2 Use of mathematical models (two models for 4.1 and 4.2)

TC1-69 Colour Rendering of White Light Sources: Wendy Davis US This committee has been maintaining its timeline.

Since the last meeting, but prior to this meeting: · Last fall, Members intending to conduct vision experiments relevant to the work of the Committee completed experiment questionnaires, detailing their plans. These were submitted to the Chair and distributed to the entire Committee. Fifteen experiments are planned, by six different research groups. · An Experiment Guidelines Subgroup was formed, but low participation made it unable to issue guidelines.

At the Stockholm meeting held the previous day: · Twelve active Members were in attendance, as well as many observers. Since the Committee is still in the process of conducting experiments (less than half-way through this task) those conducting experiments gave updates on their work. · University of Leeds: Ronnier Luo reported on an experiment conducted by one of his students. Observers estimated color differences of samples under different illuminants with a grey scale method. The predictions from different color difference formulae and chromatic adaptation transforms were compared to the visual data. · University of Pannonia: Janos Schanda reported on the status of experiments planned in his laboratory. Work is underway, and sources and other hardware are being developed, but it is too early to present results. · An experiment performed in collaboration with GE Lumination was reported by

Emil Radkov. Several different 6500 K sources were experimentally tested for perceived color difference, chromatic discrimination, preferences, and harmony. · NIST: Wendy Davis reported that a new experiment has been added that will explore whether the Hunt Effect is apparent between sunlight conditions and indoor artificial light conditions. Other experiments have been slowed due to hardware setbacks, but are still planned to be completed. · Philips Lighting: Peter van der Burgt reported on an experiment. Under several light sources observers rated color differences and preferences of 24 colored objects. · LASH ENTPE: Sophie Boissard reported on an experiment. A pair of sources illuminated identical fruits and vegetables and observers selected the source that produced the most beautiful rendering, the most natural appearance, and the most suitable illumination for selecting fruits and vegetables. The other planned experiments are still being developed, as the experimenter is unsure of certain details of the methods. · Osram Sylvania: Wendy Davis reported that this work was slowed by the birth of a baby and questions of the stability of the light sources (resolved). The experiment will investigate the trade-off in perceived object vividness of artificial food place-settings between illumination level and rendered object chroma.

Other work was presented from those who had not submitted experiment plans: · CRC: Françoise Viénot reported on previous research on color rendering, which included a discrimination task, colorfulness ratings, and hue naming. A new experiment is being conducted in which Kruithof’s findings are being re-evaluated. Performance, color perception, and subject feelings are all being tested. · Osvaldo da Pos: Pilot studies are being conducted to investigate how judgments of light sources depend on information given to observers. Color matching and semantic scales are being used in this experiment. · Laboratorium voor Lichttechnologie: This group is developing an LED lighting system to produce uniform illumination of 1000 lx at a surface 40 cm in diameter located 80 cm from the source.

TC1-70 Metameric Samples for Indoor Daylight Evaluation: Balázs Kranicz HU Survey of references Many color experts have shown a special interest in the question of how daylight distributions could be replaced by artificial sources. If a lamp is designed for this special purpose it is of high importance how the quality of the lamp can be determined or expressed.

CIE Technical Committee 1-51 published a method that fulfilled the requirements.1 As the present TC deals with metameric samples for indoor daylight in this report only the visible range will be discussed. A later report should discuss the question whether fluorescent samples also have to be defined.

The essential part of this publication1 is that five color samples are given by their spectral reflectance functions. These are called standard samples. For each standard sample a metameric pair is defined numerically for D55, D65 and D75. (Data are given with 3 decimal values.) Then one has to calculate the average color difference in CIELAB or CIELUV under the daylight simulator to be tested and based on the average color difference value the quality ranking can be achieved.

Later, CIE prepared another publication on this topic which contained metameric samples also for D50.2

The wavelength interval in reference 1 and 2 was from 400 nm to 700 nm with a step of 5 nm. As for colorimetric purposes the wavelength interval 380 nm to 780 nm is

preferred, McCamy completed the standard samples and their metameric pairs for the latter larger interval.3 Based on these data a new international standard was published for the classification of daylight simulators.4

Indoor daylight CIE TC 1-66 is about to publish a Report with the title ‘Indoor Daylight Illuminants’.5 The members of this committee studied the spectral effect of several types of glass sheets and defined the indoor version of D50 and D65 that are referred as ID50 and ID65 respectively.

It is obvious that if efforts were made to manufacture artificial daylight simulators, then efforts will be made to manufacture lamps realizing indoor daylight distributions.

TC 1-70 has been formed to design metameric samples for indoor daylight illuminants, i.e. for ID50 and ID65.

Metameric samples for ID50 and ID65 Studying the standard and metameric samples designed for daylight simulators, given in the international standard4, we can obtain 3 requirements: · The spectral range should be the interval from 380 nm to 780 nm. · The wavelength step should be 5 nm. · The values of the spectral reflectance functions at each wavelength value should be given by 3 decimal values. It seems to be a good idea to keep these standard samples, so the task was to derive new metameric pairs for the standard samples, taking the previous requirements into account. A further requirement was that the shape of the spectral reflectance functions of the new metameric samples be as similar to those used for D65 and D50 simulators as possible.

The technical details of how the new metameric samples have been derived will be discussed and published later. Now it is only stated that the task could be handled with a quite usual kind of least square method. However the requirement that the spectral reflectance functions had to be given by 3 decimal values involved some combinatoric tricks. The results are shown in Figs. 1 and 2. CIELAB color differences between standard samples and the new metameric pairs derived for ID65 and ID50 are summarized in Table I. Table I. Color differences between the standard samples and the new metameric pairs under indoor daylight illuminants Pair #1 Pair #2 Pair #3 Pair #4 Pair #5 Average

* ID65 DEab 10, 0.005 0.005 0.008 0.007 0.008 0.007

* ID50 DEab 10, 0.002 0.005 0.009 0.005 0.008 0.006

References 1. CIE 51-1981. A Method for Assessing the Quality of Daylight Simulators for Colorimetry. 2. CIE 51.2-1999. A Method for Assessing the Quality of Daylight Simulators for Colorimetry. 3. McCamy C S: New Metamers for Assessing the Visible Spectra of Daylight Simulators and a Method of Evaluating them. Color Research and Application, Oct 1999, vol. 24, no. 5, p. 322-330. 4. ISO 23603 – CIE S 012 E. Standard method of assessing the spectral quality of daylight simulators for visual appraisal and measurement of colour. 5. CIE 18x:2008. Indoor Daylight Illuminants. DRAFT.

CIE metamers used for D65 simulators and metameric samples designed for ID65

1.0

0.9

0.8

0.7

0.6 CIE 0.5 MP_ID65 0.4

0.3 Spectral reflectance factor Spectral reflectance

0.2

0.1

0.0 380 420 460 500 540 580 620 660 700 740 780 Wavelength, nm

Figure 1. Spectral reflectance functions of the metameric pairs used for D65 simulators (CIE) and of the new metameric pairs derived for indoor D65 (MP_ID65)

CIE metamers used for D50 simulators and metameric samples designed for ID50

1.0

0.9

0.8

0.7

0.6 CIE 0.5 MP_ID50 0.4

0.3 Spectral reflectance factor Spectral reflectance

0.2

0.1

0.0 380 420 460 500 540 580 620 660 700 740 780 Wavelength, nm

Figure 2. Spectral reflectance functions of the metameric pairs used for D50 simulators (CIE) and of the new metameric pairs derived for indoor D50 (MP_ID50)

TC1-71 Tristimulus Integration : Changjun Li CN During this year there has been email correspondence within the technical committee. There have been generally two voices.

Voice 1 says that for accurately computing the CIE tristimulus values and for the best agreement among laboratories, we need a unified method. This would involve: 1. studying the effects of bandpass and scanning interval on object colour measurements; 2. recommending a standard method for computing weighting factors for calculating CIE tristimulus values for reflectance measured at intervals larger than 1 nm;

3. quantifying the accuracy using the equations:

n n n Dl )( Dl )( Dl )( = å , RWX jjX = å , RWY jjY = å , RWZ jjZ j=0 j=0 j=0

with Dl = 5, 10, and 20 nm 4. recommending sets of weighting factors under some standard illuminants such as A, C, D50, D65, D75, F2, F7, and F11 with two standard observers, and at 10 nm and 20 nm intervals.

Possible methods for Voice 1 involve using: · ASTM Table 5, · ASTM Table 6, · Venable method, · Optimum weights by Li, Luo and Rigg, · Least square weights by Wang, Li and Luo, · Direct selection method, · Oleari ‘s local power expansion method

Voice 2 says that we do not need to standardise any method. But the measured bandpass error to the measured reflectance must be corrected. Once we have the bandpass-error-corrected reflectance, we can compute the tristimulus values. This sounds good. Currently CIE TC2-60 is working on this approach. But problems with this approach are: a) we have to have, or be able to estimate, the instrumental function, b) how to compute the tristimulus values with the bandpass-corrected reflectance since it is in general that the bandpass-corrected reflectance function is at intervals larger than 1 nm. For problem (a), it is difficult: manufacturers never provide the instrumental functions. CIE TC2-60 has suggested a method for calculating the instrumental functions, however, it seems that companies cannot afford to have the extra instrument to do the measurement. Suppose we know the instrumental function, then the measured reflectance can be corrected using the formulae suggested by CIE TC2-60. Since in general the measurement is at an interval larger than 1 nm, ASTM Table 5 can be used for the computation of the tristimulus values. In addition, the bandpass-corrected reflectance can be interpolated into 1 nm data, and then the CIE 1 nm formula can be used. The latter method is denoted as “CIE-R”.

To summarize: possible methods for Voices 1 and 2 are: · ASTM Table 5, · ASTM Table 6, · Venable method, · Optimum weights by Li, Luo and Rigg, · Least square weights by Wang, Li and Luo, · Direct selection method, · Oleari ‘s local power expansion, · ‘CIE-R’

The agreed comparison procedure is to: 1. Obtain 1 nm data for standard samples. 2. Use 1 nm xbar, ybar, zbar CMFs, and 1 nm for SPD (using interpolation if needed). 3. Compute ground-truth XYZ values using 1 nm summations (1 nm Riemann sums). 4. ‘Measure’ reflectance from 1 nm standard at Δλ nm interval using an assumed triangular instrument function.

5. Using the candidate method to derive a weight set or applying bandpass correction to the measured reflectance; 6. Compute XYZ values using the candidate method; 7. Compare the XYZ values at Step 6 with the ground-truth XYZ values at Step 3.

TC1-72 Measurement of Appearance Network : MapNet : Mike Pointer GB MApNet is now active, currently with 64 members! There are 8 Subject Groups:

1. Physical aspects of To understand what happens to light incident on a appearance surface, and to include interference, diffraction, absorption and scattering at surfaces and in the bulk, leading to effects that are responsible for colour, gloss, translucency and texture. 2. Non-imaging To measure light, including BSDF (including BRDF appearance metrology and BTDF), gloss, geometry, reflection, translucency, physical texture, distinctness-of-image, etc. 3. Imaging appearance To measure spatial information, including images, metrology focussing on the object, the illuminant, visual rendering, the reproduction of appearance, etc. 4. Gloss To investigate visual correlates that interpret the physical stimulus in terms of gloss. 5. Colour To investigate visual correlates that interpret the physical stimulus in terms of colour. 6. Translucency To define the relevant measurements and the contextual conditions which determine specific impressions of transparent/translucent objects. 7. Texture To investigate visual correlates that interpret the physical stimulus in terms of texture. 8. Total appearance To consider the interpretation of visual aspects of objects and scenes.

The Technical Leaders are: 1. Joanne Zwinkels National Research Council Canada, Ottawa, Canada 2. Frédéric Leloup KaHo Sint-Lieven University College, Gent, Belgium 3. Marina Bloj Division of Optometry, Bradford Optometry, Colour and Lighting Laboratory, University of Bradford, UK 4. Gaël Obein LNE-INM/Cnam - Institut National de Métrologie 5. Changjun Li* Dept of Colour Science, University of Leeds, UK Anya Hurlbert University of Newcastle, UK 6. Osvaldo da Pos Dipartimento di Psicologia Generale, University of Padua 7. Mike Chantler Texture Laboratory, School of Mathematical and Computer Sciences, Heriot Watt University, UK 8. Not yet appointed

*Changjun Li is only contributing to Group 5 within the remit that he has as Reporter R1-42 on extensions of the CIECAM02 colour appearance model.

Achievements · Network established · Subject areas decided · Technical Leaders appointed · Details of appropriate meetings have been circulated · A database of relevant published work (>1300 references) has been donated by one network member · Offers have been received to host the next CIE Expert Symposium on Appearance

· Consideration is being given to establish at least two new Technical Committees.

Next activities – by year end · Major reports from Technical Leaders · Set location of next Expert Symposium and appoint organising committee

R1-32 Emotional Aspects of Colour: Gunilla Derefeldt SE Guinlla Derefeldt sent her apologies for not being able to attend the Division meeting and asked that the reportership continue for one more year.

R1-39 Alternative Forms of the CIEDE2000 Colour-Difference Equations: Mike Pointer GB R1-39 is problematic in that the paper to which it refers is still not published in a recognised and available Journal. It was proposed and agreed that the report go forward for the Division and BA ballot and then be published in the CIE Collection.

R1-41 Adaptation Transforms: Boris Oicherman IL Formed in 2007 in Beijing, this reportership is one year old. Boris Oicherman reported that he had not been able to work on this report and would not be able to do so in the foreseeable future. Therefore he recommended closing the reportership and this was agreed: 12 affirmative, 0 negative, 0 abstain.

R1-42 Extensions of CIECAM02: Changjun Li CN This TC was also formed at the CIE 26th Session in Beijing, July 2007 and had an open meeting during the Color & Imaging Conference, CIC15, held in Albuquerque, November 2007.

Problems with CIECAM02: · The colours that are causing an issue are near the spectral locus. Some consider these colours to have little practical importance. The correction needed is to clarify the mathematics. It is possible that these colours are important for next generation displays. · Part of the discussion in this TC will be to define the applicability of the CIECAM02 model (e.g., profile connection space (PCS), colour difference, spectral colours). · The McCann data set was discarded in making the current CIECAM02. It does not fit well to the overall model. Perhaps that dataset should be an indicator of the correction that is needed. · There is debate regarding whether narrow band imaging is intended to be included in CIECAM02. This TC would appreciate contributions of datasets of narrow band colours. · What is needed is a canonical closed form method to handle all PCS CIELAB and CIEXYZ colours in colour management systems. · Will the CIECAM02 correction be a mathematical modification within the current structure, or will the structure need to be changed to address narrow band colour? · Perhaps the approach should be for a short-term fix (tweak the mathematics) but also to look at future needs for spectral imaging. · Tweaking the mathematics (CAT) will not address the issues of dealing with the full range of PCS colours. · The major problem with CIECAM02 (02) is that its domain is smaller than the ICC PCS. Therefore, what is absolutely needed is guidance for what the engineer using the system should do when encoding data. Perhaps there are two opposite solutions: The first is to acknowledge the smaller domain and tell how to encode when you are outside it. Second is to enlarge the CIECAM02 domain to fill the entire ICC PCS. The second option is preferred. Then there was some discussion about the current domain being illuminant dependent, but the HPE triangle … … · If CIECAM02 were improved to cover the full domain of the PCS would that be enough? No, because of the issues of change of illuminant. The first thing to do is

to check for which illuminants CIECAM02 works.

Consensus items agreed: · We should expand the domain. · We should not break (or change) anything that is currently working well. · First, we should check for which illuminants CIECAM02 works, and at what illuminance levels it works. · We should make an action item to specify the domain that is desired.

Extensions to the CIECAM02: · CIC15 paper by Chengyang Fu et al: Quantifying colour appearance for unrelated colour under photopic and mesopic vision, pages 319-324. This study investigates the colour appearance for unrelated colours with two sizes (0.5° and 10°) under photopic and mesopic vision. The same test colours with different field sizes were assessed under different luminance levels with a reference white ranging from 60 to 0.1 cd×m-2. Eight phases of psychophysical experiments were conducted to obtain visual data assessed by a panel of nine observers. The experimental results were used to test two colour appearance models: CAM97u and CIECAM02. · The CIECAM02 model was revised for predicting unrelated colours. It was found that CAM97u predicted brightness visual results more accurately than the CIECAM02 and the opposite was found for predicting colourfulness visual results. For predicting hue, both CAM97u and CIECAM02 gave satisfactory predictions. · This paper provides a starting point for extending CIECAM02 to mesopic vision and to predict the colour appearance for unrelated colours.

Activity of Working Group within IEC: · This group (TC 110 / WG 5) is to standardise the CIECAM02 for evaluating the quality of OLEDs based on two papers: CIE 26 Session, Beijing: A Study on the Method of Evaluating Image Quality with CIECAM02; IMID2007: A Study on the Evaluation Method of Perceptual Contrast with CIECAM02 · Because CIECAM02 is under revision with TC8-11, it is not ready to be standardised yet. If it is standardised, it should be by CIE rather than IEC. · I was informed of this situation by Paula J. Alessi and she has written to the Vice President Technical, Janos Schanda of the CIE. I do not know the result up to now.

9. LIAISON REPORTS

9.1 L1-1 AIC: Paula Alessi The next meetings of the AIC were as follows:

· Stockholm, Sweden: Colour: Effects and Affects, 15-18 June 2008. See: www.aic2008.org · Sydney, Australia: 11th AIC Congress, 27 September – 2 October 2009. See: www.aic2009.org · Mar del Plata, Argentina: Color and Food: From Production to Consumption, October 2010. · Zurich, Switzerland: The Staging of Colour - Real and Virtual Environments, June 2011. For further details see: www.aic-color.org/congr.htm

9.2 L1-2 CCPR (Consultative Committee for Photometry and Radiometry): Michael Stock

The CCPR and its working groups met at the BIPM on 21-22 June 2007. Most of the detailed technical discussions were held in the working group meetings.

The key comparison working group (WG-KC) initiates key comparisons in the field of radiometry and photometry. These are the technical foundation for the mutual recognition of national measurement standards and of calibration and measurement certificates issued by national metrology institutes in the framework of the CIPM MRA (Mutual Recognition Arrangement, www.bipm.org/en/cipm-mra).

A standing agenda item of the meetings is an update on the status of the ongoing key comparisons, carried out by the CCPR and the Regional Metrology Organizations (RMOs). Most key comparisons of the first round are now finished and preparations for the second round started. It is planned to start in 2009 with a new comparison of spectral regular transmittance, to be followed by comparisons of luminous responsivity and luminous intensity and flux. The results of the key comparisons can be found in the key comparison data base kcdb.bipm.org.

The working group on calibration and measurement capabilities (WG-CMC) coordinates the international review of the declared calibration and measurement capabilities (CMCs) of national metrology institutes. The outcome of this review process is a list of internationally recognized CMCs, which are listed in the key comparison database of the BIPM (kcdb.bipm.org/AppendixC). For radiometry and photometry, calibrations of 37 countries are covered, the total number of recognized calibrations in this field is larger than 900. At its last meeting the working group has decided on a procedure of how to maintain the CMCs if new key comparison results become available.

The working group on strategic planning (WG-SP) will advise the CCPR on future directions and will monitor developments with respect to possible future modifications of the SI system of units. The working group concluded that the impact of the proposed redefinition of the kilogram on the realization of the candela would be insignificant. The working group has prepared a text on the realization of the definition of the candela, which is published in the appendix 2 of the new SI brochure. This appendix is only published in electronic form. See http://www.bipm.org/en/si/si_brochure/appendix2/

On 2 April 2007 the CIPM and the CIE have signed an agreement on reciprocal representation and exchange of information. The CCPR has proposed to grant observer status to the CIE and the WMO. The proposal was approved by the CIPM in October 2007.

The working groups of the CCPR will meet during the NewRad conference in Korea in October 2008. The next regular CCPR meeting is planned for September 2009. General information on the work of the CCPR can be found on www.bipm.org/en/committees/cc/ccpr

9.3 L1-3 ISO/TC6/WG3: Paper, Board and Pulp – Optical Properties: Joanne Zwinkels CA

The following recent activities may be of interest to the CIE:

The next meeting of TC6/WG3 is scheduled in conjunction with the ISO TC6 plenary meetings in Seoul, Korea, June 6-13, 2008.

There has been a change in the Convenorship of WG3 due to the recent resignation of Robert Wood, CAN. Dr. Byron Jordan, CAN has been nominated to assume the Chairmanship of ISO TC6 and Dr. Joanne Zwinkels, CAN has been nominated to replace Dr. Jordan as the Convenor of WG3.

One standard is being balloted as a CD: CD 11476: Determination of whiteness, C/2° (indoor illumination conditions). It is proposed by one of the members to continue to carry out the UV adjustment corresponding to the CIE illuminant C condition but to change from C/2° to D65/10° colorimetric evaluation. This will cause a significant change in the indoor whiteness scale but this change is argued on the following points: the 10° observer is more relevant; CIE whiteness was originally developed only for D65; and there should be consistency in the whiteness measurements with the related ISO 11475: Determination of CIE whiteness, D65/10° (outdoor daylight conditions).

There is one standard being balloted as an FDIS: FDIS 9416: Determination of light scattering and absorption coefficients using Kubelka-Munk theory. At the DIS ballot stage, there were questions about the origin of the precision statement data that are presented in the method. Either the method will proceed to the FDIS level and publication with a round robin comparison being conducted at the time of the next review, or the review process will now start all over as the method is at the end of its time limits.

There is one new work item that is being balloted: a revision of ISO 2469:2007: Measurement of diffuse radiance factor to include the following statement: “A filter or other means shall be provided to ensure that the ultraviolet intensity is negligible for wavelengths shorter than 300 nm.” This revision is proposed to improve inter- instrument agreement due to the fact that the instruments used to measure the fluorescent radiance of fluorescent papers have different relative amounts of UVA and UVB and it has been shown that the fluorescence excited in these papers is due to a combination of UVA and UVB. Thus, for a one-point UVA adjustment procedure, as in ISO 2469:2007, to be valid, the instrument source should produce negligible UVB intensity. This revised procedure will also give closer agreement with CIE standard illumination conditions which specify a zero value below 300 nm.

There is considerable interest in WG3 on the recommendations of CIE TC 1-66 (Indoor Daylight) and two WG3 members are also members of this TC (Zwinkels and Jordan). The draft recommendations of this TC for spectral distributions for indoor D65 (ID65) and indoor D50 (ID50) are being analyzed by WG3 to determine if one of these indoor illuminants is similar to the Illuminant C presently used by WG3.

9.4 L1-4 ISO/TC38/SC1: Textile: Colour Fastness & Measurement M Ronnier Luo

CMCCON2 has passed National Vote to become the new ISO standard entitled Textiles – tests for colour fastness – Part J05: Method for the instrumental assessment of the colour inconstancy of a specimen with change in illuminant (CMCCON02) as part of the 105 series for assessing colour constancy (ISO 105-J05:2007).

A new method for assessing colour difference using dyers’ terms (brightness, depth and hue) has been proposed as a New Work Item (NWI).

A NWI has been recommended to establish an inter-instrument calibration procedure for the UV content of light sources of spectrophotometers used for the measurement of white or coloured materials containing fluorescent whitening agents.

A method for assessing colourfastness grades by a digital imaging technique is making good progress to become ISO standard ISO/CD 105-A11. The method includes an illumination cabinet for capturing images of objects and two formulae for converting colour measurement data to colour fastness grades for colour change and staining respectively. A ring test including labs has been extended to 7 countries. The

data have been analysed showing how accurate the instrumental method is in predicting visual results.

The numerical definition of standard depths is proposed as a NWI.

9.5 L1-5 ISO/TC42: Photography: Jack Holm US

WG18 Edition 2 of ISO 12234-1, the digital camera standard that specifies metadata and digital camera removable memory, directory, and file formats (pointing to Exif and TIFF/EP) has been published. Edition 2 is only slightly revised from Edition 1 (primarily a pointer to DCF was added), but work has begun on Edition 3. It was previously agreed that the Edition 3 revision could begin when Edition 2 was published. Edition 3 is planned to include a significant updating of the common metadata including metadata persistence rules, and the addition of a pointer to JPEG 2000. Work is ongoing and a draft is scheduled for circulation prior to the next WG18 meeting in Cologne (October 2008).

The ISO 12234-2 (TIFF/EP format) revision has begun. TIFF/EP is the base format for most camera raw images. Many users request more standardization in this area, but there are concerns from digital camera makers about how to proceed while protecting vendor differentiation opportunities. There were extended discussions at the Lausanne TC42 meeting (June 2007) and the San Jose WG18 meeting (January 2008), and it was agreed that somewhat processed (e.g. defect and noise removal, aberration correction, dark current and flat field correction) camera raw and/or scene-referred TIFF/EP format profiles could be developed. The method of supporting demosaiced camera RGB and scene-referred images using ICC profiles was presented in Lausanne. In San Jose, additional metadata items for color processing were proposed, and it was agreed to work on creating optional pre-processing metadata for defect removal, etc. Liaison communications are ongoing with JTC1 SC29 concerning JPEG XR, which also uses a TIFF type file format. The ad-hoc group for this work was expanded and more discussions are planned for Cologne.

Revision and maintenance work continues on ISO 12231 (Vocabulary), ISO 12233 (Camera resolution), ISO 14524 (Camera OECF), ISO 15739 (Camera noise & dynamic range), ISO 16067 (Scanner OECF & noise), ISO 21550 (Scanner dynamic range), ISO 20462 (Subjective image quality evaluation).

ISO 12232 (Digital camera ISO speed and exposure index) Edition 2 was published recently; it includes significant technical changes and additions from Edition 1.

JWG21 A complete revision of the ISO 5 (Densitometry) series is in progress. JWG21 met in Paris (April 2008) to address the comments received on the NP/CD ballots, and DIS drafts are being prepared. The primary objective of the revision is to better specify the calculation of density from spectral measurements, and add some graphic arts features like polarization.

JWG23 ISO 22028-1:2004 was reconfirmed (for five more years) at the Lausanne plenary with a technical corrigendum to correct two typos. There are ongoing discussions about extending Annex B tables to include new standard color encodings as they are developed.

The ICC has requested a revision of ISO 22028-2 (ROMM RGB) addressing the ICC Perceptual Intent Reference Medium (PRMG) gamut. “The PRMG should be explicitly mentioned; a diagram of the encoding range, legal encoding range, and fuzzy

reference medium gamut should be provided along with explanatory material. Describe the difference between an available encoding range, a legal encoding range and encoded values within a reference medium gamut.”

At the JWG23 meeting in San Jose, there was discussion of the need for a new scene-referred color encoding standard for digital camera captured images. Existing scene-referred encodings include RIMM and ERIMM RGB in ISO 22028-3 and scRGB and scRGB-nl in IEC 61966-2-2. It was suggested that the RIMM encoding has too little above adopted white headroom and the ERIMM encoding too much. The scRGB headroom is about right but the use of the 709 primaries requires the use of negative values or offsets that are not well supported by many popular imaging software applications. An ad-hoc group has been formed to study the issues and make a recommendation.

A NP has been approved to standardize ECI RGB as ISO 22028-4. The NP draft appears to be in good shape and this standard is likely to proceed quickly.

JWG24 The comments on the ISO 3664 (Viewing Conditions) Edition 3 CD ballot have all been resolved and the DIS is currently out for vote. Significant changes include tighter UV tolerances for print viewing and display viewing tolerances that are linked to display brightness.

9.6 L1-6 ISO/TC130: Graphic Technology: Danny Rich US

During the past year, this ISO technical committee has been very busy working on standards that reference CIE recommendations and standards. Of particular interest to Division 1 are the following activities.

TC130 Graphic Arts and TC42 Photography have two joint working groups. One is involved in revising all of the standards on the measurement of reflection and transmission density (ISO 5-1, 5-2, 5-3 and 5-4). The second is involved in revising ISO 3664, the standard for sources for viewing graphic arts materials for color and tone. The latter is still based on CIE D50 but has been wrestling with the problem that the CIE has not yet issued a methodology for assessing the ultraviolet range of this illuminant. Publication 51 added visual range metamers for D50 but did not extend the original work from D65 to D50 for the UV (300nm to 400nm) spectral region. This is becoming a critical issue as more substrates for printing, especially digital print, are being supplied with increasing amount of fluorescent whitening agents. There are two approaches to resolve this problem. In 3664 on viewing sources and in 13655 on instrumental measurements, three sources have been defined. Source 1 will be an exact simulation of CIE D50 without any definition of how to assess “exactness”. Source 2 will be a source that is a category B or better simulation of D50 in the visible, using the methods of Publication 51 but includes a sharp-cut filter that blocks all radiation below 400nm. For spectral measurements, this requirement is relaxed and any spectral source can be used as long as it has no UV radiance.

ISO TC6/WG3 made a proposal to TC130 that they adopt a new illuminant for the graphic arts. In the past 2 years, TC6 has been active in developing new standards that use an illuminant which they have named, “Indoor Daylight”. This illuminant was constructed by taking the historical CIE illuminant C and extrapolating the short wavelength end of the spectrum down to 300nm. Since the graphic arts has standardized on CIE D50 but has no definition of the UV portion, TC6/WG3 suggests that TC130 adopt D50 for wavelengths above 400nm and “Indoor Daylight” below 400nm. It is highly desirable to have a CIE prescription for simulation of D50 below 400nm.

In ISO standards that involve tolerances on the color of reproductions are beginning to either replace or supplement the traditional CIELAB differences with CIEDE2000 tolerances. Experience is showing that the newer equation is giving reliable results for visual correlation.

9.7 L1-7 ISO/IEC JTC1/SC28 Office Equipment: Klaus Richter DE This liaison report is attached to these Minutes as Appendix 6

9.8 L1-8 ISO/TC159/WG2 Ergonomics: Ken Sagawa JP This liaison report is attached to these Minutes as Appendix 7.

9.9 L1-9 International Association of Lighthouse Authorities: Ian Tutt GB No report submitted.

10 NEW WORK ITEMS The following were approved:

10.1 Technical Committee: Real Colour Gamut (C) Terms of Reference: To recommend a gamut representative of real (non-fluorescent) surface colours and defined by associated spectral reflectance data. Chairman: Changjun Li CN

The establishment of this TC was approved: 11 affirmative, 0 negative, 2 abstain.

10.2 Reporter: Evaluation of Whiteness (C) Terms of Reference: To review the current status of whiteness measurement and recommend future requirements. Joanne Zwinkels CA has agreed to be this reporter.

The establishment of this reportership was requested by Joanne Zwinkels as part of her liaison with ISO/TC6/WG3 – see above. The WG had met during the week prior to this Division meeting and a need for clarification of, and perhaps change in, the use the CIE Whiteness formula had become apparent.

The establishment of this Reporter was approved: 12 affirmative, 0 negative, 0 abstain.

10.3 Reporter: Hue Angles of Elementary Colours (C) Terms of Reference: To review the current literature on elementary (unique) hues for potential imaging applications. Thorstein Seim NO agreed to be this reporter.

The establishment of this Reportership was requested as part of the liaison with ISO/JTC1/SC28 Office Equipment – see report above.

The establishment of this Reporter was approved: 12 affirmative, 1 negative, 0 abstain.

There was some discussion as to whether a reportership should be formed to assess the intra- and inter-observer variability of unique hue judgements, based on some recent work. It was agreed that the researcher should be encouraged to publish the work and then the above reporter can then consider it.

Secretary’s Note: The numbers for these new Technical Committees and Reporters will be assigned by the CIE Central Bureau after they have been approved by the Board of Administration.

11. NEXT MEETING Possible venues for the next meeting of the Division in 2009 are: Budapest, Hungary as part of the CIE Mid-Term meeting May 2009. Sydney, Australia: in conjunction with the AIC Congress in September / October 2009. A vote showed that that country members present were evenly spilt and it will thus be necessary to ballot all country members by email.

Secretary’s Note: The result of an email ballot that ended on 18 July 2008 showed that the preferred meeting location is Budapest, Hungary. Thus the next meeting of CIE Division 1 will be held in Budapest, Hungary in association with the CIE Midterm Meeting and associated conference in May-June 2009.

13. ANY OTHER BUSINESS None.

CLOSE OF MEETING The Director thanked everyone for attending, and the hosts for their organisation of the practical aspects of the meeting.

Dr Michael R Pointer Secretary – CIE Division 1 August 2008

APPENDIX 1 CIE DIVISION 1 VISION AND COLOUR

First Meeting of the Luo Term

15 June 2008

AGENDA

Opening Session

9:00 – 9:30 1. Opening and welcome by Director, Ronnier Luo

2. Apologies for absence

3. Membership

4. Attendance

5. Approval of agenda

6. Approval of minutes of Beijing meeting

7. Matters arising from those minutes

8. Report from the Director: Ronnier Luo

9. Report from the Editor: John Setchell

10. Report from the Secretary: Mike Pointer

Business Session

9:30 – 11:45 11. Report of the Vision Section: Miyoshi Ayama

11:45 – 12:45 12. Report of the Colour Section: Ellen Carter

12:45 – 13:30 Lunch break

13:30 – 14:30 13. Report of the Colour Section: continued

14:30 – 15:00 14. Liaison reports

15:00 – 15:30 15. New work items

15:30 – 16:00 16. Any other business: Location of next meeting

16:00 17. Close of meeting There will be a break for coffee/tea during the morning and afternoon sessions. Mike Pointer Division Secretary

APPENDIX 2: Summary Report on R1-44: Limits of Normal Colour Vision

Reporter: S. McFadden (CA)

Terms of Reference: To review the literature to see what information is available to establish the limits of normal colour vision.

Even a cursory review of the literature indicates that the measurement of colour vision has been and continues to be extensively researched. For example, a recent review of clinical colour vision tests (Dain, 2004) cites 192 papers. Further evidence can be found in the report published by Division 4, CIE 143-2001(Commission Internationale de l'Éclairage, 2001), on colour vision standards for transport. The aim of the report was to encourage international harmonization in colour vision requirements in maritime, air, rail and road transport, and the use of valid methods for the assessment of colour vision. It specified the colour vision requirements necessary to ensure safe and reliable recognition of coloured signal lights and other colour coded visual information. In addition to documenting the importance of good colour vision for people involved in the transportation industry, it recommends tests procedures that should be used in assessing colour vision and provides detailed information on the recommended colour vision tests.

The impetus for this Reportership was a paper presented by Barbur, Rodriguez-Carmona, and Harlow (2006) at the ISCC/CIE Expert Symposium on 75 Years of the CIE Standard Colorimetric Observer. The paper documented their efforts to establish a better test for assessing colour deficiency based on the “the statistical limits of colour discrimination in ‘normal’ trichromats”. Thus, the goal of a TC arising out of this Reportership would not be to recommend the most suitable tests, but would be to link the assessment of colour deficiency to the measurement of normal colour vision. Hopefully this would lead to not only improved assessment of the severity of colour vision loss, but also better characterization of normal colour vision capabilities.

Given the size of the literature of colour vision testing, both normal and abnormal, it has not been possible to conduct a detailed review. Most of the information presented below is based on Boynton’s chapter on colour deficiency (Boynton, 1979) and recent articles on clinical (Dain, 2004) and occupational testing (Squires, Rodriguez-Carmona, Evans, & Barbur, 2005). Although there have been innumerable papers on the measurement of both normal and defective colour vision since its publication, Boynton’s book still provides a reasonable summary of our basic knowledge in this area. The more recent articles confirm that a standard agreed upon method for assessing colour vision based on the limits of normal colour vision that will meet most users’ requirements does not currently exist. Possible reasons for the failure to satisfactorily define the limits of normal colour vision are:

· Most colour vision tests are aimed at characterizing either the type or severity of colour deficiency. In contrast, most research is aimed at understanding or defining the “standard colour observer”.

· Colour vision is multi-faceted. It can be assessed in terms of colour matching, colour discrimination or colour appearance.

· Colour vision is not one-dimensional. It varies along at least three dimensions, red-green, blue-yellow, and brightness or lightness. However, most tests focus on red-green deficiency.

These issues are explored further below.

Research on normal colour vision focuses on assessing colour matching, colour discrimination and colour appearance. Performance on each of these factors is usually assessed over the three dimensional colour space defined by the red-green, blue-yellow, and brightness or

lightness axis. Each factor has been defined in some detail as a function of a wide range of variables including ambient illumination, stimulus size, and stimulus source (i.e. reflective versus emissive). In all cases, the goal is to establish the nature of the colour perception of the colour normal observer. The CIE has produces multiple documents that summarize and capture the available knowledge on the characteristics of the normal colour observer from the 1931 Standard Colorimetric Observer up to the latest colour appearance model (Commission Internationale de l'Éclairage, 2004; Hunt, 2006). From this research it is possible to establish the variability in colour perception. However, since usually only a small number of “colour normal” observers are used in these studies, it is not possible to derive the limits for normal colour matching, colour discrimination, and colour appearance. Moreover, the methods used in these studies require complex, expensive equipment and extensive testing. Thus, they are not suitable for assessing large numbers of people. The same methods have been used with colour defective individuals, but in most cases, the observers have clearly defined types of defective vision. The data have been collected primarily to better understand the basis of normal colour vision (Boynton, 1979).

In contrast, research on colour deficiency has focused primarily on assessing the presence of a colour deficiency, the diagnosis of the type and severity of that deficiency, and/or the significance of the colour vision deficiency (Dain, 2004). The first two are reflected in the proposed CIE definition of defective colour vision:

Anomaly of vision in which there is a reduced ability to discriminate between some or all colours

NOTE 1 Anomalous trichromatism: Form of trichromatism in which colour discrimination is less than normal.

NOTE 2 Deutan: Adjective denoting deuteranopia or deuteranomalous.

NOTE 3 Deuteranomaly: Defective colour vision in which discrimination of the reddish and greenish contents of colours is reduced, without any colours appearing abnormally dim.

NOTE 4 Deuteranopia: Defective colour vision in which discrimination of the reddish and greenish contents of colours is absent, without any colours appearing abnormally dim.

NOTE 5 Dichromatism: Defective colour vision in which all colours can be matched using additive mixtures of only two matching stimuli.

NOTE 6 Monochromatism: Defective colour vision in which all colours can be matched using only a single matching stimulus.

NOTE 7 Protan: Adjective denoting protanopia or protanomaly.

NOTE 8 Protanomaly: Defective colour vision in which discrimination of the reddish and greenish contents of colours is reduced, with reddish colours appearing abnormally dim.

NOTE 9 Protanopia: Defective colour vision in which discrimination of the reddish and greenish contents of colours is absent, with reddish colours appearing abnormally dim.

NOTE 10 Tritanomaly: Defective colour vision in which discrimination of the bluish and yellowish contents of colours is reduced.

NOTE 11 Tritanopia: Defective colour vision in which discrimination of the bluish and yellowish contents of colours is absent.

A large number of colour vision tests have been developed and validated. Each test has its strengths and limitations (Dain, 2004), but none of them completely discriminate the colour normal from colour defective individual. False negatives and false positives are common. The most commonly used test for screening is the Ishihara pseudoisochromatic plates. It will reliably detect people with even a relatively mild red-green colour defect although it has been shown to pass people with very mild deficiencies and fail some people that would be classified as colour normal. In their recent evaluation of colour vision standards employed by the international aviation community, Squires, Rodriguez-Carmona, Evans, and Barbur (2005) found that all 55 colour deficient participants made multiple errors on the Ishihara test, but in a few cases the number of errors was less than 5. At the same time 7 of the 24 colour normals tested made between 1 and 3 errors. Although widely accepted as a screening test, the Ishihara plates are not suitable for assessing the severity of colour deficiency or even discriminating between protans and deutans. Moreover, they provide no information on people’s ability to do standard colour perception tasks of colour discrimination, colour naming, or colour matching. There does not appear to be any relationship between performance on the Ishihara and normal colour discrimination.

For diagnosing the type and extent of red-green deficiency, the preferred clinical test is the Nagel anomaloscope (Dain, 2004). However, it is relatively expensive and difficult to use. It uses a colour matching task to assess and classify the type and degree of red-green colour deficiency. The subject looks through a viewfinder to see a disk split horizontally into two half fields. The top half of the disk has a red and green light that can be mixed in different proportions by turning a scaled knob to create a range of colors from 0 (green) to 73 (red). The bottom half of the disk is illuminated with a spectral yellow light. The brightness of this light can be altered by turning another scaled knob. The subject matches the appearance of the two half fields in both color and brightness by altering the red/green mixture ratio and the brightness of the yellow field. Multiple studies (see Boynton (1979) for a summary) have shown that it is possible to clearly differentiate the large majority of colour normals from protanomalous and deuteranomalous individuals. Moreover, the anomaloscope is generally regarded as very accurate in distinguishing the protanomalous from the deuteranomalous observer, although the relationship between the parameters of the match and the subject’s overall colour discrimination sensitivity is known to be generally poor (Hurvich, 1972; Wright, 1946). A narrow red / green matching range is often taken to indicate high chromatic sensitivity, typical of normal trichromatic vision. The variability of the matching range within normal trichromats is large and this makes direct comparison with colour deficient observers more difficult. Recent studies have shown that subjects with reduced, red-green chromatic sensitivity, typical of mild deuteranomaly, can pass the Nagel test with midpoint and matching range established for normal trichromats (J. L. Barbur et al., 2008). Such experimental findings illustrate that although the anomaloscope produces accurate results, unusual combination of parameters can produce erroneous matches. “Even the anomaloscope, therefore is not by itself sufficient to grade the anomaly with certainty” (Wright, 1946).

Another common limitation with many colour vision tests, including the two described above, is that they only assess red-green deficiency. Although genetic blue-yellow deficiencies are relatively rare, this type of deficiency is more common with acquired colour deficiencies and age related reduction in colour perception due to yellowing of the lens. An anomaloscope for assessing blue-yellow deficiency has been developed, but so far it has not proven very successful (Dain, 2004). Any tests purported to describe the limits of colour vision would have to describe those limits along all three dimension of colour vision.

The most popular test for assessing the degree and type of colour discrimination deficiency is probably the Farnsworth-Munsell 100 hue test. It is composed of 85 of the100 Munsell hues that cover the whole range of hues available in the Munsell system. According to Dain (2004), Farnsworth removed 15 hues to make the perceptual spacing more equal. However, the colour difference of the remaining hues is not constant and there is a somewhat systematic change in CIE L*. The remaining hues are divided into four sets. Each set is composed of 2 fixed (one at each end) and 22 movable caps. Subjects must arrange the movable hues so

that they provide a progression between the two fixed hues. The pattern and degree of errors indicates the type and extent of colour deficiency. The test is fairly difficult and even most colour normals make a few errors. In fact, the test is often used to screen for superior colour discrimination. However, it is not clear if superior performance is entirely due to superior ability (Dain, 2004). One drawback to the 100 hue test is that it differs substantively from the colour discrimination tasks used in the study of normal colour vision. Typically, colour discrimination tasks try to determine the minimal separation, in the CIE colour space, at which subjects can reliably discriminate two colours (MacAdam, 1942). Thus, it would be very difficult to use this test to directly establish the limits of normal colour discrimination unless data comparing the different approaches exist.

The test developed by Barbur and his colleagues (John L. Barbur et al., 2006; Rodriguez-Carmona, Harlow, Walker, & Barbur, 2005), on the other hand does build on traditional colour discrimination tasks. Using a four alternative forced choice procedure, subjects indicate the direction of motion of a colour defined stimulus. To detect the motion the subject must be able to discriminate the coloured stimulus from its achromatic background. The luminance of the stimulus and its background randomly vary over time and space to prevent the use of luminance information in the stimulus. Their results to date indicate that the test clearly separates people classified as colour normals on a battery of existing colour vision tests from people who have failed one or more of the existing commonly used tests. Moreover, the test allows accurate classification of deutan- and protan-like deficiencies and ranking of colour discrimination capability along the green-red and yellow-blue dimension. The test is currently being used in a large trial involving other conventional tests and discrimination of signal lights in the aviation environment as a precursor to recommending it for use by the UK and US civil aviation authorities. As with any new test, it will be important to see if similar results are found when it is used by clinicians and other testing agencies.

The third dimension along which colour vision can be assessed is colour appearance or colour naming. Typically, colour appearance is assessed using some form of colour naming task (Boynton & Olson, 1987). Alternatively, people are asked to describe a colour in terms of the amount of red and green, or blue and yellow it contains (Ayama & Sakurai, 1999). Unlike colour matching and colour discrimination, there is considerable variability in colour naming tasks especially if other than basic colour names are allowed. Thus, clinical tests to assess defective colour naming ability are usually limited to a small number of basic colours – red, green, blue, yellow, and white. The most common clinical test is the lantern test. There have been several instantiations, but in most cases subjects are shown a series of coloured lights or pairs of coloured lights and required to name the colour of each light. The lights are designed to simulate signal lights and the colours are usually restricted to red, green, white, and in some case blue and yellow. In their recent evaluation, Squires et al. (2005) measured the performance of normal and colour defective individuals on three different lantern tests. Their results highlight the problems of trying to assess normal colour naming capability. To start, the chromaticity coordinates of the lights varied across tests and the luminances of the lights varied both within and across tests. In some cases, the chromaticity coordinates were outside the limits specified by CIE S004/E-2001 (Commission Internationale de l'Eclairage, 2001) for the colours of signal lights. With only one of the lamps, did all 24 colour normal subjects meet the pass criterion. Twelve failed a second lamp and 3 failed the third lamp. In most cases, the white light was incorrectly given a non-white label or vice versa. On the other hand, between 3 and 8 deuteranomalous subjects passed at least one of the lantern tests with at least one subject passing all three. Thus, it would appear that establishing the limits of normal colour naming is not currently possible. This probably reflects not only the limitations of the tests, but also the limitations of our knowledge of colour naming. When TC1-61 publishes its report, we may have a better understanding of “normal” colour naming performance.

Recommendations

Based on the brief review of the literature carried out to date, it is not clear if a Division 1 Technical Committee to define the limits of colour vision is warranted at this point. Certainly, the large current literature on colour vision testing and recent advances in understanding the limits of normal chromatic discrimination suggest that the topic continues to be of both fundamental and occupational interest. Moreover, the results of studies such as those conducted by Squires et al. (2005) suggest that the current tests and standards are inadequate for reliably and consistently describing the limits of normal colour vision. There does appear to be sufficient data to define the limits for normal colour discrimination. On the other hand, it is not clear if there are sufficient data that the final report would provide substantively more than CIE 143-2001(Commission Internationale de l'Éclairage, 2001). This is because very little of the research to date appears to:

· systematically evaluate the colour matching, colour discrimination or colour naming performance of a wide range of colour deficient individuals;

· systematically measure the performance of a large number of colour normal individuals on tests of colour deficiency; or

· integrate the data from studies of colour discrimination etc. with the data from studies of colour deficiency.

However, a more detailed review of the literature would be necessary to confirm the above.

Thus, it is recommended that the Reportership be continued for another year to allow a more definitive review.

As part of this review, it is recommended that the Reporter approach other researchers, clinicians, and practitioners involved in colour vision testing to determine if

· they would be interested in participating in a TC on this topic;

· they have unpublished data that could be used to help define the limits of normal colour vision.

Acknowledgement

I would like to thank Dr John Barbur for his contributions to this report. His knowledge and extensive research in this area greatly improved its accuracy and completeness.

References Ayama, M., & Sakurai, M. (1999). Effects of achromatic surround on color appearance in the peripheral retina. In Proceedings of the CIE 24th Session (Vol. 1, pp. 120-122). Vienna, Austria: Commission Internationale de L'Eclairage. Barbur, J. L., Rodriguez-Carmona, M., & Harlow, A. (2006). Establishing the statistical limits of "normal" chromatic sensitivity. In Proceedings of the ISCC/CIE Expert Symposium '06: 75 Years of the CIE Standard Colorimetric Observer (CIE x030:2006, pp. 168-171). Vienna, Austria: CIE Central Bureau. Barbur, J. L., Rodriguez-Carmona, M., Harlow, J. A., Mancuso, K., Neitz, J., & NET, M. (2008). A study of unusual Rayleigh matches in deutan deficiency. Visual Neuroscience, 25, 1-10. Boynton, R. M. (1979). Variations and defects in human colour vision. In Human Color Vision (pp. 337-389). New York: Holt, Rinehart and Winston. Boynton, R. M., & Olson, C. X. (1987). Locating basic colors in the OSA space. Colour Research and Application, 12(2), 94-105. Commission Internationale de l'Eclairage. (2001). Colours of Light Signals (CIE S004/E-2001).

Vienna, Austria: CIE Central Bureau. Commission Internationale de l'Éclairage. (2001). International recommendations for colour vision requirements for transport (CIE 143-2001). Vienna, Austria: CIE Central Bureau. Commission Internationale de l'Éclairage. (2004). A colour appearance model for colour management systems: CIECAM02 (CIE 159:2004). Vienna, Austria: CIE Central Bureau. Dain, S. J. (2004). Clinical colour vision tests. Clinical and Experimental Optometry, 87, 276-293. Hunt, R. W. G. (2006). CIE Colour appearance models, their past and future. In Proceedings of the ISCC/CIE Expert Symposium '06: 75 Years of the CIE Standard Colorimetric Observer (CIE x030:2006, pp. 85-90). Vienna, Austria: CIE Central Bureau. Hurvich, L. M. (1972). Color vision deficiencies. In D. Jameson & L. M. Hurvich (Eds.), Handbook of Sensory Physiology. Vol. VII/4. Visual Psychophysics (pp. 582-624). Berlin: Springer-Verlag. MacAdam, D. L. (1942). Visual sensitivities to color differences in daylight. Journal of the Optical Society of America, 32(5), 247-274. Rodriguez-Carmona, M. L., Harlow, A. J., Walker, G., & Barbur, J. L. (2005). The variability of normal trichromatic vision and the establishment of the "normal" range. In Proceedings of the 10th Congress of the International Colour Association (pp. 979-982). Granada, Spain. Squires, T. J., Rodriguez-Carmona, M., Evans, A. D. B., & Barbur, J. L. (2005). Color vision tests for aviation: comparison of the anomaloscope and three lantern types. Aviation, Space and Environmental Medicine, 76, 421-429. Wright, W. D. (1946). Researches on normal and defective colour vision London: Henry Kimpton.

APPENDIX 3: Color-Matching Functions Using Maxwell’s Method: A Tutorial

Michael H. Brill, 16 Jan 2008, revised 27, 28 Jan 2008

I think no tutorial has yet been written to understand how the color-matching functions for Maxwell color matching are related to measured quantities---although general descriptions have been provided by Wyszecki and Stiles, Fairchild, and Thornton. Therefore, here is a tutorial to fit that need.

Imagine a trichromatic visual system in which Grassmann’s laws work and give a convex spectrum locus with the usual line of purples (in chromaticity). Posit a white comparison stimulus C with tristimulus vector W, and three monochromatic primaries at wavelengths λr > λg > λb. Divide the spectrum locus into three compact, disjoint regions: (i) From 0 to λ’r (the complementary wavelength to λr with respect to W), (ii) from λ’r to λ’b (the complementary wavelength to λb with respect to W), and (iii) from λ’b to infinity. For any variable wavelength λ, the Maxwell method finds the match to C (always the same C) composed of power Pλ(λ) of the variable wavelength and power at two of the three primary wavelengths. The problem addressed by this memo is to express the color-matching functions rbar(λ), gbar(λ), and bbar(λ) for these primaries in terms of the quantities obtained during Maxwell matches as described above.

In region (i): The two participating primaries are r and g, with respective powers Pr(λ) and Pg(λ). The matching equations are as follows:

Pλ(λ) rbar(λ) + Pr(λ) rbar(λr) + Pg(λ) rbar(λg) = Wr Pλ(λ) gbar(λ) + Pr(λ) gbar(λr) + Pg(λ) gbar(λg) = Wg (A1) Pλ(λ) bbar(λ) + Pr(λ) bbar(λr) + Pg(λ) bbar(λg) = Wb.

Here, Pr(λ) and Pg(λ) represent, respectively, the power in the red or green primary in the color mixture involving variable-wavelength λ that matches W. Similarly, Pλ(λ) represents, for the same color-mixture, the power at the variable wavelength λ.

We don’t know the components of W because we don’t know the color-matching functions yet, but we do know that λb lies in region (i) and therefore has the same matching equations as any other wavelength in region (i):

Pb(λb) rbar(λb) + Pr(λb) rbar(λr) + Pg(λb) rbar(λg) = Wr Pb(λb) gbar(λb) + Pr(λb) gbar(λr) + Pg(λb) gbar(λg) = Wg (A2) Pb(λb) bbar(λb) + Pr(λb) bbar(λr) + Pg(λb) bbar(λg) = Wb.

Also, the following constraints on well-normalized color-matching functions apply:

rbar(λr) = gbar(λg) = bbar(λb) = 1 rbar(λg) = rbar(λb) = bbar(λg) = bbar(λr) = gbar(λr) = gbar(λb) = 0. (A3)

Now we equate the left-hand sides of Eq. A1 with the left-hand sides of Eq. A2 using the constraints of Eq. A3, thereby eliminating W. In the process:

Pλ(λ) rbar(λ) + Pr(λ) = Pr(λb) Pλ(λ) gbar(λ) + Pg(λ) = Pg(λb) (A4) Pλ(λ) bbar(λ) = Pb(λb).

Solving Eq. A4 for rbar, gbar, and bbar gives

rbar(λ) = [Pr(λb) - Pr(λ)] / Pλ(λ) (A5a) gbar(λ) = [Pg(λb) - Pg(λ)] / Pλ(λ) (A5b)

bbar(λ) = Pb(λb) / Pλ(λ). (A5c)

In region (ii): The two participating primaries are r and b, with respective powers Pr(λ) and Pb(λ). The matching equations are as follows:

Pλ(λ) rbar(λ) + Pr(λ) rbar(λr) + Pb(λ) rbar(λb) = Wr Pλ(λ) gbar(λ) + Pr(λ) gbar(λr) + Pb(λ) gbar(λb) = Wg (A6) Pλ(λ) bbar(λ) + Pr(λ) bbar(λr) + Pb(λ) bbar(λb) = Wb.

We know that λg lies in region (ii) and therefore has the same matching equations as any other wavelength in region (ii):

Pg(λg) rbar(λg) + Pr(λg) rbar(λr) + Pb(λg) rbar(λb) = Wr Pb(λg) gbar(λg) + Pr(λg) gbar(λr) + Pb(λg) gbar(λb) = Wg (A7) Pb(λg) bbar(λg) + Pr(λg) bbar(λr) + Pb(λg) bbar(λb) = Wb.

As in Region (i), we equate the left-hand sides of Eq. A6 with the left-hand sides of Eq. A7 using the constraints of Eq. A3, and then divide by Pλ(λ):

rbar(λ) = [Pr(λg) - Pr(λ)] / Pλ(λ) (A8a) bbar(λ) = [Pb(λg) – Pb(λ)] / Pλ(λ) (A8b) gbar(λ) = Pg(λg) / Pλ(λ). (A8c)

In region (iii): The two participating primaries are g and b, with respective powers Pg(λ) and Pb(λ). A procedure analogous to the above yields:

gbar(λ) = [Pg(λr) – Pg(λ)] / Pλ(λ) (A9a) bbar(λ) = [Pb(λr) – Pb(λ)] / Pλ(λ) (A9b) rbar(λ) = Pr(λr) / Pλ(λ). (A9c)

Note that Eqs. (A5), (A8), and (A9) are represented in more compact and less explicit form by Wyszecki&Stiles Eq. 1(5.3.2):

rbar(λ) = [Pr - Pr(λ)] / Pλ(λ) gbar(λ) = [Pg - Pg(λ)] / Pλ(λ) bbar(λ) = [Pb – Pb(λ)] / Pλ(λ)

with the proviso that the power of one of the primaries [Pb(λ), Pg(λ), or Pr(λ)] is zero in the respective three regions.

Region Boundaries: The only step left is to decide for any wavelength λ which of the three regions it’s in. One answer is to process the wavelengths in increasing order. Automatically you start in region (i). When only λr is needed with λ to match C, you are transitioning from region (i) into region (ii). When only λb is needed with λ to match C, you are transitioning from region (ii) into region (iii).

APPENDIX 4: Visual Realization of Wyszecki & Stiles Section 5.3.2 Conditions

Michael H. Brill, 8 May 2008, revised 13 May 2008

In search of what could visually be responsible for the Maxwell/Max-Saturation discrepancy, one can begin by consulting W&S Section 5.3.2 (p. 295), wherein a nonlinear visual model is held to satisfy three conditions (given a chosen comparison “white” stimulus C):

Condition I: A set of three primary wavelengths exists such that C is matched by a unique recipe of positive radiant powers at those primary wavelengths.

Condition II: Given a test wavelength and a set of primary wavelengths for which Condition I is satisfied, C matches a unique recipe of positive radiant powers of the test wavelength and at most two of the primary wavelengths.

Condition IIIa: If a spectrum that matches C is augmented by positive radiant power at some or all visible wavelengths, it no longer matches C.

Condition IIIb: If spectra P and Q both match C, then R = aP + (1-a) Q matches C (where 0 £ a £ 1). The same holds when R, P, and Q are permuted.

A three-dimensional Grassmann structure clearly satisfies Conditions I-III, but can we find a different model that would accommodate the discrepancy between Maxwell and Maximum-Saturation matches? What nonlinear visual model satisfies Conditions I-III? The most challenging of these axioms is Condition IIIb. If a match is represented by equality via a nonlinear function f (probably a vector-valued function) of the spectrum, then for what f does f[P] = f[Q] = f[C] imply that f[aP + (1-a) Q] = f[C] ? I have tried summing a power function of the SPD wavelength by wavelength, and also Palmer’s mesopic luminance model, and neither satisfies Condition IIIb.

One class of models consistent with Condition IIIb is the following: Let the rod (R) and cone (L, M, S) tetrastimulus values comprise the 4-vector p = (L, M, S, R), and the tristimulus values comprise 3-vector x. Then, given p, Condition IIIb says that 3-vector x that is input to color vision must satisfy

p . f(x) = q(x), (B1) where q(x) is a scalar function of x, f is a 4-dimensional column-vector-valued function, and . is a dot product (or, equivalently, matrix multiplication). Clearly if two vectors p satisfy Eq. B1 with the same x, then any convex combination of those p vectors satisfy B1 with the same x. These two values of p and their convex combination comprise metamers in the Maxwell-match space. Equation B1 mathematically defines a ruled 3-surface in p-space. Changing x is analogous to changing the level of water in which the ruled surface is immersed; by the ruled-surface property, the intersection of the planar water surface and the surface defined by Eq. 1 is always a straight line. Of course, the three degrees of freedom in x imply that the orientation of the analogue water surface has three degrees of freedom.

To render Eq. B1 into a specific model, one must be able to predict x from any input vector p. This implies three equations with the form of Eq. B1:

. p fi(x) = qi(x), (B2) where i = 1, 2, 3.

Although Eq. B2 is abstract, one particular case is quite familiar: Let

. fi(x) = p ai – xi b, (B3a)

qi(x) = ci – xi d , (B3b)

where ai and b are constant column-4-vectors, and ci and d are constant scalars. Equations B2 and B3 imply a projective transformation from p to x:

. . xi = (p ai + ci ) /(p b + d) (B4)

. for i = 1, 2, 3. The center of projection p0 is somewhere (unspecified) on the plane p b + d = 0 in p-space; all points p that lie on a line through p0 are metameric (give the same point in x space). The origin of p-space maps to the point xi = ci /d .

In general, the model in Eq. B4 (designed to satisfy Condition IIIb) is inconsistent with Condition IIIa: Every point x on or in the spectrum locus (a substantial part of positive x-space) must arise from a point p in positive p-space. Condition IIIa demands that any positive excursion of p must change p’s image x. But that would mean the line in p-space from p0 to p must not be aligned in the direction of entirely positive (or entirely negative) p, so as to avoid positive-p excursion being along a line of metamers.

However, Wyszecki and Stiles stipulate that the conditions must apply only for matches to a single comparison stimulus C, whose p-vector I now call pC. For matches that depart from pC, the vector x can approach a classical three-dimensional Grassmann structure. Here is a model that passes all the W&S Conditions, given a unique comparison stimulus C: 3 . . xi = β (p ai + ci ) /(p b + d) + (1 – β) å Aij pj , (B5) j=1 2 where β = k/[k + |p – pC| ], k is a positive constant, and Aij are components of a constant 3x3 matrix from LMS to x. Equation B5 is a convex combination of the perspective model of Eq. B4 and the 3D Grassmann model, the former given complete weight for Maxwell matches to C, and the latter given preponderant weight for maximum-saturation matches.

This is just an illustration of a line of reasoning. Other constraints will have to be examined before such a model can be deemed “C-worthy.” Further exploration is needed to find a front-end visual model consistent with W&S Conditions I-III, and also consistent with visual experiments that are not Maxwell matching.

APPENDIX 5: TC1-63 Test Chart

APPENDIX 6: L1-7 ISO/IEC JTC1/SC28 Office Equipment: Klaus Richter DE PowerPoint Report follows

APPENDIX 7: L1-8 ISO/TC159/WG2 Ergonomics: Ken Sagawa JP PowerPoint Report follows

Liaison Report from ISO/IEC JTC1/SC28 to CIE Div. 1

Liaison Report of SC28 to the CIE Division 1 meeting, June 15, 2008 ISO/IEC JTC1/SC28 Plenary Meeting with 60 participants last week in Germany 13 Standard projects under Work in ISO/IEC JTC1/SC28 in the areas: Image Quality, Productivity, Yield

Relations to the work of SC28: ISO/IEC SC28 uses the CIELAB standard in many standard documents, for example in ISO/IEC 15775, ISO/IEC TR 19797, ISO/IEC TR 24705. The colour data rgb used in these documents have a device dependent relation to CIELAB. For download adresses of some of the public documents see: http://www.ps.bam.de/33872E

There are at present 10 NWI listed by AWG/PWG5 “Office Colour Work Group”, for example no. 9 (doc. j28n1194) “Colour space standards for offices” Ð provide a device independent common colour space for office equipment.

Liaison Report from ISO/IEC JTC1/SC28 to CIE Div. 1

122

See for similar files: http://www.ps.bam.de/De39/; www.ps.bam.de/De.HTM Achromatic colours Chromatic elementary colours Chromatic device colours BAM material: code=rha4ta application for measurement of printer or monitor systems BAM registration: 20071001-De39/10L/L39E00NP.PS/.PDF "Neither-nor"-colours Television (TV), Print (PR) Photography (PH) five achromatic colours: four chromatic elementary colours: six chromatic device colours: N Black (french noir) R Red C Cyan blue D Dark grey neither yellowish nor blueish M Magenta red G Green Z Central grey neither yellowish nor blueish Y Yellow H Light grey B Blue O Orange red W White neither greenish nor reddish L Leaf green J Yellow (french jaune) V Violet blue neither greenish nor reddish

De390−3X

Input Output Input and output media and applications Technical Report Input media Output media Application (TR) or Standard −− −−Basis ISO/IEC TR 24705

analog analog ISO/IEC-test chart (hardcopy) Hardcopy Copier ISO/IEC 15775

analog digital ISO/IEC-test chart (hardcopy) File Scanner ISO/IEC TR 24705

 Hardcopy Printer ISO/IEC TR 24705 digital analog ISO/IEC-test chart (file) Softcopy Monitor ISO/IEC TR 24705

De390−7N

Liaison Report from ISO/IEC JTC1/SC28 to CIE Div. 1

122 See for similar files: http://www.ps.bam.de/De42/; www.ps.bam.de/De.HTM BAM material: code=rha4ta application for measurement of printer or monitor systems BAM registration: 20071001-De42/10L/L42E00NA.PS /.TXT

Application of colour in daily life and in Information Technology (IT): Design, architecture, art, industrial products Information technology of printers

Measured for CIE standard illuminant D65 Measured for CIE "other" illuminant D50 colour order system: name and coordinates Device system name and coordinates: RAL Design System (CIELAB): Printer system (illuminant D50): LCH*, lightness, chroma, hue cmy, content of "cyan", "magenta", "yellow" Munsell Colour System: Display system (standard illuminant D65): VCH*, lightness (Value), Chroma, Hue rgb/sRGB, content of "red", "green", "blue" Natural Colour System (NCS): IT colour coordinates: undefined for users! ncu*: blackness, chroma, elementary hue text DIN 33872: Relation to colour order systems! For CIELAB: Definition of relative coordinates lab* with linear relation to LAB* CIELAB: LAB* : lightness, red-green and jellow-blue chroma; LCH* : lightness, chroma, hue Definition of relative coordinates similar to coordinates of the colour order system NCS lch*: relative lightness l*, relative chroma c*, relative hue h* = hab/360 of CIELAB tce*: triangle lightness t*, relative chroma c*, relative elementary hue e* ncu*, icu*: relative blackness n* or brilliantness i*, relative chroma c*, elementary hue text u*

De420−7N

Liaison Report from ISO/IEC JTC1/SC28 to CIE Div. 1

122 Input: Colorimetric Offset Reflective System ORS18a Output: Colorimetric Offset Reflective System ORS18a See for similar files: http://www.ps.bam.de/De15/; www.ps.bam.de/De.HTM BAM material: code=rha4ta application for output of monitor, data projector, or printer systems BAM registration: 20080301-De15/10L/L15e00NP.PS /.PDF with rgb data of the ORS18a; adapted (a) CIELAB data with hue number ORS18a; adapted (a) CIELAB data L*=L*a a*a b*a C*ab,a h*ab,a L*=L*a a*a b*a C*ab,a h*ab,a four elementary hues b*a n= 00 to 19 b*a OMa 47.94 65.39 50.52 82.63 38 OMa 47.94 65.39 50.52 82.63 38 1 0 0 = Red R YMa 90.37 −10.26 91.75 92.32 96 00 = Red R YMa 90.37 −10.26 91.75 92.32 96 1 1 0 = Yellow J LMa 50.9 −62.83 34.96 71.91 151 05 = Yellow J LMa 50.9 −62.83 34.96 71.91 151 a*a a*a CMa 58.62 −30.34 −45.01 54.3 236 CMa 58.62 −30.34 −45.01 54.3 236 0 1 0 = Green G VMa 25.72 31.1 −44.4 54.22 305 10 = Green G VMa 25.72 31.1 −44.4 54.22 305 0 0 1 = Blue B MMa 48.13 75.28 −8.36 75.74 354 15 = Blue B MMa 48.13 75.28 −8.36 75.74 354 NMa 18.01 0.0 0.0 0.0 0 NMa 18.01 0.0 0.0 0.0 0 WMa 95.41 0.0 0.0 0.0 0 WMa 95.41 0.0 0.0 0.0 0 RCIE 39.92 58.66 26.98 64.57 25 RCIE 39.92 58.66 26.98 64.57 25 JCIE 81.26 −2.16 67.76 67.79 92 JCIE 81.26 −2.16 67.76 67.79 92 GCIE 52.23 −42.25 11.76 43.87 164 GCIE 52.23 −42.25 11.76 43.87 164 BCIE 30.57 1.15 −46.84 46.86 271 BCIE 30.57 1.15 −46.84 46.86 271

1 1 0 05 06 04

Yellow J 07 Yellow J 03 greenish redish greenish redish

08 02

09 01 yellowish yellowish yellowish yellowish

0 1 0 Green G Red R 1 0 0 10 Green G Red R 00

bluish bluish bluish bluish 11 19

12 18 greenish redish greenish redish Blue B 13 Blue B 17

14 16 0 0 1 15

De150−7N, 20 step hue circle with elementary colours R, J, G, B (left) 20 step hue circle with elementary colours R, J, G, B (right) CIE Division 1 2008 Stockholm Liaison with ISO/TC159 “ Ergonomics” K. Sagawa

TC159/WG2 “Ergonomics for people with special requirements”

ISO/TR22411 1st version to be published soon 2nd version of the TR is being edited (for more informantion)

TC159/SC5/WG5 “Physical environment for people with special requirements”

WI 24502 Age-related luminance contrast (changed from “Age-related luminance” ) CIE Division 1 2008 Stockholm Age-related luminance contrast K. Sagawa

700 700 å å ll VLVL )(2)(1 aeae )(,)(, D-D llll 400 400 C a)( = 700 å lVL )(2 ae )(, Dll 400 0.5 L1 0.0 10 years old

efficiency -0.5

-1.0

-1.5

-2.0 L2

-2.5 70 years old -3.0 L1e,l, L2e,l : spectral radiance

Log relative luminous luminous Log relative -3.5 350 400 450 500 550 600 650 700 750 V(l)(a): spectral luminous efficiency Wavelength (nm) of age (group) a. CIE Division 1 2008 Stockholm Age-related relative luminance; L(a) K. Sagawa

CIED1 2008 Stockholm not accepted ! L(a) =∫Le,lV(l)(a)dl

) 0.0020 Age-related luminance for observer of 20s 2 -2 -1 Characters: Lc= S Le,l(c)V(l)(25)Dl = 0.0155 Wm sr 0.0015 -2 -1 Background: Lb= S Le,l(b)V(l)(25)Dl = 0.0076 Wm sr

Luminance contrast 0.0010 = (L -L )/(L +L ) =

2 c b c b 34 %

L (c) Age-related luminance for observer of 70s 0.0005 e,l Characters: L = S L (c)V(l) Dl = 0.0092 Wm-2sr-1 2 L (b) c e,l (75) e,l -2 -1 2 Background: Lb= S Le,l(b)V(l)(75)Dl = 0.0082 Wm sr

Spectral radiance (watt/m sr (watt/m radiance Spectral 0.0000 400 450 500 550 600 650 700 Luminance contrast

Wavelength (nm) = (Lc-Lb)/(Lc+Lb) = 6 %