Open Access
Add to BibSonomyAbstract
This is a comment on a recently published paper by Lin at el. [Optics Express 22(S4), A1029 (2014)]. The original article suffers three fatal defects as follow: (1) Some peak distortions are not present in the spectral loss simulation due to large wavelength step; (2) The correlations between the modulated deviations (MDs) and the color rendering indices (CRIs) are poor at high CRI region; (3) How to determine the optimal peak wavelengths of multi-peak LED in the guideline for the CRI optimization does not make any statement. There are poor correlations between the MDs and CRIs at high CRI region.
© 2015 Optical Society of America
1. Introduction
The correlations between color rendering indices (CRIs) and the spectral power distributions (SPDs) have been reported by Lin et al [1]. However, in Lin’s article, there are three defects as follow: (1) Some CRI peak distortions are not present in the spectral loss simulation due to large wavelength step; (2) The correlations between the modulated deviation (MD) and CRIs are poor at high CRI region; (3) How to determine the optimal peak wavelengths of multi-hump LED does not give any example in the guideline for the CRI optimization. The following will make a detailed analysis and pointed out some defects.
2. Three defects
2.1 Some CRI peak distortions are not present in the spectral loss simulation
The CRI distortion analysis in the spectral loss simulation with well-width set on 50 nm at correlated color temperatures (CCTs) of 3500 K, 5455 K, and 8500 K is not detailed enough due to the wavelength step of 50 nm, which is shown in Figs. 1(a)-1(c) reported by Lin [1]. If setting the wavelength step of 5 nm, these CRI distortions are also shown in Figs. 1(d)-1(f). From Fig. 1, it is found that some peak distortions are not present for the 50 nm step, and that there are great differences of the CRI distortions between with 50 nm and 5 nm steps. So the conclusions about CRI distortions reported by Lin [1] have to be modified. Figures 1(d)-(f) shows that the CRI distortion patterns at various reference CCTs have the similar structures of themselves. There are four peak distortions for Ra, and three peak distortions for R9. The wavelengths of peak distortions decrease as the CCT increases. The maximum distortion appears in the yellow region, secondly in the red region.
Fig. 1 The CRI distortions in the spectral loss simulation with well-width set on 50 nm and wavelength step set on 50 nm, on (a) 3500 K, (b) 5455 K, and (c) 8500 K respectively reported by Lin et al [1], and the CRI distortions with well-width set on 50 nm and wavelength step set on 5 nm, on (d) 3500 K, (e) 5455 K, and (f) 8500 K respectively. Solid lines highlight Ra (black) and R9 (red) respectively, and other indices are illustrated as dashed lines.
The Ra and R9 distortions with well-widths of 5, 10, 20 and 50 nm at the CCT of 5455 K by using wavelength step of 5 nm, and the reported by Lin [1] are shown in Figs. 2(c)-2(d) and Figs. 2(a)-2(b), respectively. There are great differences of the CRI distortions between with 5 nm and 50 nm steps. So the conclusions about Ra and R9 distortions with various well-widths reported by Lin [1] need to modified. Figures 2(c)-2(d) shows that the Ra and R9 distortion patterns with various well-widths have similar structures of themselves. As the well-width decreases, the distortions in both Ra and R9 are reduced.
Fig. 2 The (a) Ra and (b) R9 distortions with well-widths of 5, 10, 20 and 50 nm by using wavelength step of 50 nm reported by Lin et al [1], as well as (c) Ra and (d) R9 distortions with well-widths of 5, 10, 20 and 50 nm by using wavelength step of 5 nm at the CCT of 5455 K.
2.2 Poor correlations between MDs and CRIs at high CRI region
In order to reveal a high coherence between the modulated deviations (MDs) and its related CRIs, an example of the simulation on a set of 3-peak spectra with Gaussian function are illustrated in Fig*. 5 and Fig*. 6 (* labeled as shown in [1], Lin et al) by continuously changing the intensities of the Gaussian peaks. However, their CCTs will change as the intensities of the Gaussian peaks change, which is shown in Table 1. As pointed out in [1], Lin et al, it is worth conducting comparisons between simulated SPDs (s-SPDs) and reference SPDs (r-SPDs) under the same CCT, in order to discover Ra and R9 sensitivity in different spectral regions. However, the reference CCT of 5455 K was selected in calculating MD between s-SPD and r-SPD reported by Lin [1] although the CCT of s-SPD was different, and the CCT with optimal CRIs (Ra = 95.5, R9 = 84.9) was not 5455K but 5937K. So, the credibility of correlations between MDs and CRIs may be questioned.
Table 1. Parameters of Gaussian Functions in Simulations and the CCT Ranges of s-SPDs*
For the example above, if intensities of the Gaussian peaks are set according to Table 2, the plots of Ra and R9 vs. MD of the simulation on a set of 3-peak spectra by adjusting the intensity of each color component are illustrated in Fig. 3. The results show that correlations between MDs and CRIs by adjusting the intensity are poor at high CRIs region if the reference CCT is unchanged.
Table 2. Peak Wavelengths, FWHMs, and Intensities of Gaussian Functions in Simulations
Fig. 3 The plots of Ra and R9 vs. MD of the simulation on a set of 3-peak spectra by adjusting the intensity of each color component.
In fact, for the LED with 3-peak spectra, the relative intensity of each Gaussian peak at the specified CCT can be determined by color mixture equation [2], assuming the peak wavelength and FWHM of each color component has been given. Thus, CRIs have also been determined. For the example above, Both Ra and R9 at CCTs of 2700 K to 6500 K on the Planckian or daylight locus are list in Table 3. The highest Ra is 85.9 at CCT of 4998 K, and the highest R9 is 95.6 at CCT of 5971 K. So it is not necessary to intensity optimization.
Table 3. Both Ra and R9 at CCTs of 2700 K to 6500 K on the Planckian Locus for the LED with 3-peak Spectra with the Peak Wavelengths of 455 nm, 555 nm and 650 nm, and the FWHMs of 30 nm, 100 and 30 nm, Respectively
2.3 No optimal peak wavelengths in the guideline for the CRI optimization
It is important to optimize the peak wavelength rather than the intensity of each color component for multi-peak LED with the specified CCT. How to determine the optimal wavelengths of multi-peak LED does not give any example in the guideline for the CRI optimization reported by Lin [1]. In order to evaluate the correlations between Ra and R9 with their respective MDs in case of changing the peak wavelengths of the Gaussian functions at the specified CCT, an example of the simulation on a set of 3-peak spectra at the CCT of 5455 K by adjusting the peak wavelength of each color component are illustrated in Fig. 4. The peak wavelengths and FWHMs of the Gaussian functions are listed in Table 4. The results show that there is a poor coherence between MDs and its related CRIs at high CRIs region, especially for R9.
Fig. 4 The plots of Ra and R9 vs. MD of the simulation on a set of 3-peak spectra at the CCT of 5455 K by adjusting the peak wavelength of each color component.
Table 4. Peak Wavelengths and FWHMs of Gaussian Functions in Simulations
Two examples of the simulation on a set of SPDs with 3 and 4 peaks on a series of specific wavelengths and with fixed FWHMs reported by Lin [1] are shown in Table 4. It is found that the distances from the Planckian locus on the CIE 1960 uv chromaticity diagram (Duv) of the SPD with 4 peaks is + 0.0136, which is out of the range of the chromaticity tolerance quadrangles of white-light sources [3, 4]. So the SPD of with 4 peaks do not satisfy the requirements recommended for general lighting with solid state lighting products. So the example of the SPD with 4 peaks is not appropriate. The peak wavelengths of two examples are not best. The optimal peak wavelength and relative intensity of each color component as well as their performance of the SPDs with 3 peaks at CCT = 5937 K and with 4 peaks at CCT = 6338 K can be obtained by nonlinear program for maximizing CRIs, which are list in Table 5. The difference between the optimal parameters and the reported by Lin [1] is very significant. This also reveals a poor coherence between MDs and its related CRIs.
Table 5. Optimal Peak Wavelength and Relative Intensity of Each Color Component as well as Their Performance of SPDs with 3 Peaks at CCT = 5937 K and with 4 Peaks at CCT = 6338 K
It has been widely accepted that CRI cannot characterize color quality of LED sources, especially for color-mixed LED [5]. One problem with the CRI is that it can give fairly high scores to sources that render some saturated object colors very poorly [6, 7]. In particular, the report from CIE Technical Committee TC 1-62 “Color rendering of white LED light sources” [5] summarizes several problems of the CRI when applied to white LED sources. The CRI score does not correlate well with visual evaluation in many cases. One of reasons was assumed to be the different order of magnitude of the color differences occurring if the reflecting samples are illuminated by a white LED light source and by other light sources, due to the peculiar spectral power distributions of the white LED light sources “interacting” with the spectral reflectance of the test-color samples. This is especially noticeable for the case of test-color sample No. 9 of the CIE method which is a strong red test-color sample. Therefore, LEDs with higher CRI values do not always have good color quality.
3. Conclusions
According to the above analysis, we draw the conclusion that there are poor correlations between the MDs and CRIs at high CRI region. The correlation between color rendition and SPD need to be further in-depth analysis and research.
Acknowledgments
This work was supported by Shanghai Science and Technology Committee (No. 13111104200).
References and links
1. Y. Lin, Z. H. Deng, Z. Q. Guo, Z. G. Liu, H. Lan, Y. J. Lu, and Y. G. Cao, “Study on the correlations between color rendering indices and the spectral power distribution,” Opt. Express 22(S4Suppl 4), A1029–A1039 (2014). [CrossRef] [PubMed]
2. I. Moreno and U. Contreras, “Color distribution from multicolor LED arrays,” Opt. Express 15(6), 3607–3618 (2007). [CrossRef] [PubMed]
3. ENERGY STAR for SSL Luminaries ver. 1.1, 2008.
4. American National Standard, Specifications for the Chromaticity of Solid State Lighting Products (ANSI C78.377), NEMA, 2008.
5. “TC 1-62: Color rendering of white LED light sources,” CIE 177:2007 (International Commission on Illumination [CIE], Vienna, Austria, 2007).
6. N. Narendran and L. Deng, “Color rendering properties of LED light sources,” Proc. SPIE 4776, 61–67 (2002). [CrossRef]
7. N. Sándor and J. Schanda, “Visual colour rendering based on colour difference evaluations,” Lighting Res. Tech. 38(3), 225–239 (2006). [CrossRef]
