The Characterization of the Photometry and Colorimetry of Light Emitting Diodes. Thesis booklet. Péter Csuti

Transcription

1 The Characterization of the Photometry and Colorimetry of Light Emitting Diodes Thesis booklet Péter Csuti Supervisor: Dr. János Schanda Consultant: Dr. Katalin Hangos UNIVERSITY OF PANNONIA FACULTY OF INFORMATION TECHNOLOGY DOCTORAL SCHOOL OF INFORMATION SCIENCE AND TECHNOLOGY 206

2

3 . Overview and goals.. Characterization of the photometry of LEDs Nowadays, light emitting diodes (LEDs) do not count anymore only as the eternal signal lights but those gradually conquer also the fields of general lighting. So, it is important to have a photometry which is iable also when the light of LED based luminaires is measured. A photometer s spectral responsivity match to the V( ) function, which describes the light perception of humans [] and is characterized by the value of the f spectral mismatch index () in percentage as it is published by the CIE (Commission Internationale de l Eclairage) in [2,3]. where: f 0 s ( ) V( ) d 00%, () V ( )d s ( ) s( ) S( ) A V( )d, S( ) s( ) d V ( ), is spectral luminous efficiency function, The A s( ), is the photometer s spectral responsivity function, 2 s ( ), is the photometer s normalized spectral responsivity function, S( ) A, the spectral power distribution of the CIE standard illuminant A [4] as the calibration source of the photometer. s ( ) normalized spectral responsivity function describes the calibrated state of the photometer. The f spectral mismatch index

4 does not give any information about the possible measurement error when measuring a light source with a spectral power distribution (SPD) other than the calibration source CIE standard illuminant A. The f index also does not give information about the direction of the deviation; therefore no error correction can be based on it. The correction is only possible if one knows both, the SPD of the measured light source and the spectral responsivity of the photometer. Those requirements are fulfilled only in really rare cases; usually one might have some information about the colour category of the light sources under test..2. Characterization of the colorimetry of LEDs I ve experienced in previous colour matching experiments using narrow band illuminants (Figure, LED-RGB curve) that a perceptible yellowish-greenish colour shift appeared when matching the chromaticity of a broad band reference (Figure, Halogen curve). This phenomenon happens in cases where the colour matching metameric pairs were set in an objective way by using the colour matching functions of the CIE 93 standard colorimetric observer [5]. Metameric colour stimuli have by definition different SPDs, but the tristimulus values of their colour stimuli are equal. 3

5 Spectral radiance [W/sr/m 2 /nm] 0,020 0,05 0,00 0,005 Figure SPD of metameric colour stimuli The tristimulus values of two visually matched colour stimuli having S (λ) and (λ) 2 S SPDs can be calculated by using the following: X km S( ) x( ), but X X2 X k S ( ) x( ) 2 m 2 which means, the numeric deviation can only be caused by the differences between the human colour perception and the x( ), y( ) and z( ) colour matching functions. Based on this the task is to define the needed changes applied to the x( ), y( )és z( ) functions to satisfy the X X2 equation in case of visually matched colour stimuli..3. Literature review 0, wavelength [nm] Halogén LED-RGB.3.. Photometry of light emitting diodes The photometry of LEDs was a hot topic for the standardization communities for years. In case of white LEDs the f spectral mismatch index can be used to estimate the measurement accuracy of photometers similarly to the broad band illuminants, the incandescent lamps. Due to the fact, that white LEDs have atively broad SPD as 4

6 well. In such cases the value of f does corate well with the measurement error: in case the f index has a higher value the expected error will also be higher. CIE in the publication No 27 [6] states that the photometric measurements of white LEDs should be made with photometers having f value less than 3 %. In case of colour LEDs the f value does not corate well with the expected measurement error. There are cases when two photometers having equally low (e.g. f < 2 %) spectral mismatch values respond significantly different to the same illumination caused by a colour LED. The cause of this phenomenon is that the photometer s spectral responsivity ( s () ) could contain a range where the deviation from the V ( ) spectral luminous efficiency function is significant. This deviation has a atively small effect in case of a broad band calibration source (like the CIE standard illuminant A), but in case the colour LED s 20 nm 40 nm wide SPD to be measured falls into such a range then we will receive a measurement result differing significantly from the real photometric value [7]. The question was investigated in detail by Young et al. [8], but they didn t give a solution to the problem. Later Y. Ohno and K. Murray did work out a recommendation what they called f,led index [9]. Both indices are based on the f spectral mismatch function, but they corate with the expected measurement error only in case when mathematical model based LED SPDs are used Colorimetry of light emitting diodes The fundaments of standardized colorimetry were laid down in the last years of the 920s when the first standardized colour matching function (CMF) formulation was accepted by the CIE in 93 [0] 5

7 based on the visual experiments of Wright [,2] and Guild [3]. Later Vos recommended corrected CMFs [4] after Smith and Pokorny cone responsivity spectra were accepted [5] by the vision scientists. By the end of the 950s there was considerable evidence that colour perception depends on the viewing angle. CIE created the 0 colorimetric system [6] which ensured a better match between the visual observations and the objective measurements. Guild s and Wright s results already showed the dispersion of the measured curves. A similar phenomena could be observed by the measurements of Stiles and Guild. Naturally every observer is able to repeat the measurements only with a specific dispersion; however among observers whom have normal colour eyesight, categorized by other techniques, character dependent variances can be also found. Sarkar analysed the observer s data of Stiles and Burch and come to a conclusion that the observers could be divided into eight big groups [7]. Unfortunately the genotype assignment isn t available but based on the work of Neitz and Neitz the physiological reason of the smaller spectral response differences became clear [8]. The TC-36 technical committee of the CIE published their experimental LMS-CMFs evant for the eye s cornea, which I also used to evaluate my visual experiments..4. Aim of my research Based on the above facts it was important to investigate this topic because there are some signs that in the near future there will be a big demand for iable photometry and colorimetry. The aim of my work was on the one hand that the photometers should be easier selected and on the other hand to explain the colour matching problems experienced by using tuneable light sources with single colour LEDs 6

8 and to give a solution to those problems. By the help of my visual experiments I was looking for the answer how to define more precisely the chromaticities of visually matched metameric stimuli and how to decrease further the earlier experienced perceptual differences. 7

9 2. Synthesis of the results 2.. Summary of my work In the field of the photometry of LEDs I have defined the real photometric error indicator (2), where among the input parameters are the and the s( ) ative spectral responsivity of the observed photometer, S( ) LED spectral power distribution of the tested LED. In this way the comparison of photometers with different spectral responsivities are possible, with the help of either real or modelled LED spectra. With the real photometric error indicator I expanded the partial mismatch index according to (3); wherewith an expected measure error caused by the photometers s( ) spectral responsivity mismatch to the V ( ) function can be forecast in four (blue, green, yellow and red) wavelength ranges. According to my visual experiments in the field of colorimetry I appointed that when observers visually collate different filtered lights of filament lamps with the combination of certain narrow spectrum LEDs, this accordance is not correctly described by the colour matching functions of the CIE 93 standard colorimetric observer. This phenomenon can be noticed by observing the chromaticities seen on the u, v chromaticity diagram of Figure 2. The so called fundamental colour matching functions x F, y F and z F describe the accordance with significantly smaller errors; those colour matching functions are genereated by matrix transformation from the l( m( ) and s( ) cone responsivity curves and the results were published by the technical committee TC-36 of the CIE. During my research I developed the colour matching function system x MF, 8

10 y MF and z MF by expanding the matrix-transformation and its input parameters. This system describes the colour matching set by the observers with error smaller than Δ(u'v') = 0,006. In the subscript of the functions the MF abbreviation refers to Modified Fundamental New scientific results The new scientific results of my research are summarized in the following Theses. Thesis No. and Thesis No.2 are about the characterization of the photometry of light emitting diodes, while Thesis No.3 and Thesis No.4 can be formulated according to the results of the characterization of the colorimetry of light emitting diodes. Thesis : The real photometric error index (2) can be used to compare photometers with known [P], [P2], [P3], [P4] s( ) spectral responsivities. PE s( ), ( ) s S LED V S LED S LED 780 (, ) ( ) d ( ) ( ) d 380 V( ) S( ) d LED (2) where, V ( ), is spectral luminous efficiency function, s( ), the spectral responsivity function of the photometer, s (, ), the normalized spectral responsivity function of a photometer calibrated with a standard LED ( S( ) ) having dominant wavelength, S( ) LED, spectral power distribution of the LED to be measured. 9

11 Thesis 2: By using the partial spectral mismatch indices in (3) the value of the real photometric error can be estimated. [P], [P2], [P3], [P4] 780 s 380,PAR ( ) 780 f (, ) V( ) d 380 V ( )d, (3) where s (, ) s( ) S( ) V( ) d S( ) s( ) d, and V ( ), is the spectral luminous efficiency function, s( ), is the spectral responsivity function of the photometer, s (, ), the normalized spectral responsivity function of a photometer calibrated with a standard LED ( S( ) ) having dominant wavelength, S( ), spectral power distribution of the standard LED having dominant wavelength. 0

12 Thesis 3: The CIE 93 standard colour matching functions do not describe the colour matching correctly if certain narrow band LEDs and broad band colour stimuli are matched. (Figure 2). [P5], [P6], [P7] 0,6 v' 0,5 #3 #5 #4 #6 # #9 #7 #8 Figure 2 Experimental colour stimuli calculated with CIE 93 2 CMFs. The empty circles show the chromaticity of the visual observations, red triangles show the chromaticity of the references Thesis 4: The modified (4) fundamental colour matching functions describe the colour matching set by the observers with smaller error than Δ(u'v') = 0,006 irrespectively of the peak wavelength of the primaries. xmf (, , , l( ymf ( 0, ,4496 0, m( zmf ( 0, , , s( ) where 0,4 #2 RGB LED Visual average Broad-band reference 0,3 0,0 0, 0,2 0,3 0,4 0,5 u' i s i 6 (4) s( ) ( ) because of the shift of the function values, the subscript MF refers to Modified Fundamentals. [P5], [P6], [P7]

13 3. Further research possibilities The photometry of light emitting diodes is still a hot topic at professional forums. In 206 various LED standard manufacturer offer their products with different designs. Those products are most commonly used in measurement setups of integrating spheres where besides the single colour LED standards also exists multi-colour LED standard products in compact systems. Today s developments aiming more and more onto spectrally resolved measurement devices thanks to their prosperous price. The research topics spread from estimating the spectral power distribution of the LEDs under test to enhance further the spectral matching of the photometers spectral responsivity to the V ( ) function. The spreading of the LED based displays proves the actuality of further research of the colorimetric description. Those are used also in special applications simultaneously with older technologies like cathode ray tube displays in the printing industry or in broadcast studios. An interesting research topic is to find the variance minima of the responsivity of the human perception by choosing the primaries of a display. Another possible research direction is the individual setting of multi channel tuneable luminaires to a given white reference stimulus, which is mostly needed in the human centric general lighting. 2

14 4. Publications ated to the Theses [P] Csuti P, Kránicz B: Description of a partial f error index recommended for LED photometry, Light & Engineering, Vol.4, No.., ISSN (Thesis and Thesis 2) [P2] Csuti P, Kránicz B, Schanda J, Comparison of the goodness of fit of photometers to the V(λ) function using real LED spectra, CIE Expert Symp. on LED Light Sources: Physical measurement and visual and photobiological assessment, Tokyo, Japan, June (Thesis and Thesis 2) [P3] Csuti P, Schanda Gy, Schanda J, Decreasing the uncertainty of LED photometric and colorimetric measurements, Lighting in the XXI Century CIE Divisional and Technical Committees Meetings, León, Spain, 2-2. May (Thesis and Thesis 2) [P4] Csuti Péter, Világítástechnikai évkönyv , Világítás és ember című fejezetben a LED-es fényforrások fotometriai és villamos tulajdonságainak meghatározása című rész. pp , HU ISSN (Thesis and Thesis 2) [P5] Csuti P., N. Vidovszky Á., Schanda J., On the Application of Modern Light Sources with emphasis on home lighting, Przeglad Elektrotechniczny, Vol 2008, No. 8, p 84-88, PL ISSN (Thesis 3 and Thesis 4), Impact factor: 0,242 (2008) [P6] Csuti, P; Schanda, J (200), A Better Description of Metameric Experience of LED-Clusters, Light & Engineering, Vol. 8, No., p 44-50, 200, ISSN (Thesis 3 and Thesis 4), Impact factor: 0,036 (200) [P7] Csuti P, Schanda J: Colour matching experiments with RGB-LEDs, Color Research and Application, April Vol.33, Issue 2, p 08-2 (Thesis 3 and Thesis 4), Impact factor:,0 (2008) 3

15 5. References Gibson, K.S., The ative Visibility Function, Recueil des Travaux, 6th Session 924, Commission internationale de l'eclairage proceedings, 926. Cambridge University Press, Cambridge. pp Comission Internationale de l Eclairage, Methods of characterizing the performance of radiometers and photometers. Publication CIE Comission Internationale de l Eclairage, Methods of characterizing illuminance meters and luminance meters: Performance, characteristics and specifications. Publication CIE Publication CIE International Organization for Standardisation, Comission Internationale de l Eclairage, Standard Illuminants for Colorimetry, ISO 664-2:2007(E)/CIE S 04-2/E:2006, Joint ISO/CIE Standard 5 Comission Internationale de l Eclairage, Proc. of the 8th Session of CIE, Cambridge, 9-29, (93) 6 Comission Internationale de l Eclairage, Measurement of LEDs, Publ. CIE27:2007, ISBN Suzuki, K., Kohmoto, K., Nakagawa, Y., Kondoh, H., Bandou, K., Oba, H., Yagi, T., Okazaki, J., Yamada, K.: Round robin LED photometry test in Japan. CIE Symp. 200 LED Measurement CIE x pp. - 8 Young, R., Muray, K., Jones, C.F., Quantifying Photometric Spectral Mismatch Uncertainties in LED Measurements, Proc 2 nd CIE Expert Symp. on LED Measurement, Standard methods for specifying and measuring LED and LED cluster characteristics, Gaithersburg, USA, May 200. pp CIE x Comission Internationale de l Eclairage, Measurement of LEDs, Publ. CIE27.2 (Revision of CIE27-997) Draft No. 4, Dec Comission Internationale de l Eclairage, Proc. of the 8th Session of CIE, Cambridge, 9-29, (93) Wright, W.D., A re-determination of the trichromatic coefficients of the spectral colours, Trans. Opt. Soc. London 30, 4-64 (928-29) 2 Wright, W.D., A re-determination of the mixture curves of the spectrum, Trans. Opt. Soc. London 3, 20-2 (929-30) 3 Guild, J., The colorimetric properties of the spectrum. Philos. Trans. Roy. Soc. London, Ser. A 230, (93) 4 Vos, J.J., (978) Colorimetric and Photometric Properties of a 2 Fundamental Observer, Col. Res. and Appl., 3, No. 3, pp Smith, V.C., Pokorny, J, (975) Spectral sensitivity of the foveal cone photopigments between 400 nm and 500 nm, Vision Research, 5, pp Comission Internationale de l Eclairage Proceedings (964) Vienna Session, 963, Vol. B, pp (Committee Report E-.4.), Bureau Central de la CIE, Paris 7 Sarkar, A., Blondé, L. (203), Colorimetric Observer Categories and Their Applications in Color and Vision Sciences, Proceedings of the CIE Centenary Conference, April 203, Paris 8 Neitz, M, Neitz, J., Molecular genetics of color vision and color vision defects, Arch. Ophthalmol. 8, (2000). 4