FEATURE. Appropriate Color-rendering Indices and Their Recommended Values for White LED Lighting in UHDTV Program Production

Transcription

1 Appropriate Color-rendering Indices and Their Recommended Values for White LED Lighting in UHDTV Program Production Tetsuya Hayashida We selected appropriate color-rendering indices and determined their recommended values for white LED lighting in ultrahigh-definition television (UHDTV) program production. Since UHDTV cameras have more accurate color reproduction capability than HDTV cameras, the color rendering properties of the lighting are more critical. By evaluating images taken under white LEDs with different color rendering indices, we found that the combination of Ra and R9 is the most appropriate color-rendering index and that the values Ra 9 and R9 8 are to be recommended for white LED lighting in UHDTV program production.. Introduction The UHDTV color gamut specified in Recommendation ITU-R BT. (Rec. ) ) is wider than the HDTV color gamut specified by Recommendation ITU-R BT.79 (Rec. 79) ). UHDTV cameras have more accurate color reproduction than HDTV cameras ). Because of this accurate color reproduction, higher color-rendering properties are required in lighting for UHDTV production. A transition in lighting from incandescent lamps and fluorescent lamps to white LED lighting is progressing concurrently with the transition of the television system from HDTV to UHDTV. Although white LED lighting is expected to be implemented in television studios and sports facilities, the appropriate color-rendering indices for evaluating white LED lighting in UHDTV production are still unknown. The color-rendering index (CRI), which is defined by the International Commission on Illumination (CIE), is a quantitative measure of the ability of a light source to reveal the colors of various objects in comparison to a blackbody radiator or daylight ). The standard CRI values (R -R 8 ) are derived from CIE 97 test-color samples consisting of eight low-saturation colors and six special CRI values (R 9 -R ) derived from the realistic colors red, yellow, green, blue, Caucasian skin color, and green foliage, respectively. In addition, Japanese Industrial Standards (JIS) has added the skin color of Japanese women as R ), but since this addition is not internationally recognized, it is not taken up in this paper. The general CRI value (CIE R a ) is derived as the arithmetic mean of the eight standard CRI values (R -R 8 ). Standard organizations such as JIS 6) and sports organizations 7) 8) have specified recommended R a values in broadcasting. These values have come to be used in HDTV production. Although R a is widely used internationally for evaluating illuminants, it is reported that highly saturated colors cannot be properly evaluated in terms of R a under white LED lighting 9) ) since there are no high-saturation colors included in test-color samples R -R 8. To overcome the defect of R a and to evaluate white LED lighting properly, numerous new methods for evaluating color rendition have been proposed as a replacement for R a )-), but none have yet been standardized. Therefore, we conducted a subjective evaluation to derive appropriate color-rendering indices and their recommended values for white LED lighting in UHDTV production. In this study, we evaluated the internationally recognized R a and special color-rendering indices R 9 -R in consideration of the continuity with the conventional R a. Furthermore, while various types of white LED lighting exist, we focus on the white LED lighting that is based on a combination of a blue LED and phosphors and is expected to be widely used for television studios and sports facilities.. Subjective evaluation to select appropriate color-rendering indices and determine recommended values for UHDTV production. Method () Participants Thirty-four nonexpert participants (8 females and 6 males; age range = - years; mean age = 6. years) with normal vision (in compliance with Recommendation ITU-R BT.) ) participated in the subjective evaluation experiment.

2 () Apparatus To create different lighting conditions, we used a special lighting booth consisting of LEDs with different peak wavelengths. Three levels of correlated color temperature (CCT) * were employed:, K (studio lighting),,6 K (stadium lighting), and 6, K (reference white light of a television system). For each CCT, we configured five lighting conditions, including four simulated white LED lighting conditions and one reference lighting condition, with different CRI values. Four simulated white LED lights were selected considering their popularity and commercial availability. The spectral power distributions of these lighting conditions are shown in Fig.. To shoot the materials, we used a prototype UHDTV camera ) ) developed at NHK STRL. A linear matrix optimized under simulated daylight at a CCT of,6 K was used. The white balance of the camera was adjusted for each lighting condition using a color chart. The opto-electronic transfer function (OETF) defined by Rec. 79 was applied for gamma correction in the camera. For this subjective evaluation, we used a -inch K laser-backlight LCD covering 98% of the Rec. color gamut 6). * The color temperature of blackbody radiation that appears to have a color nearest that of the target light source. The unit of CCT is degrees Kelvin (K), the thermodynamic temperature. () Materials Two groups of objects - Sports and Flowers - were prepared. Examples of evaluation images of these two groups and their respective chromaticity distributions are shown in Fig.. The Sports group represented outdoor sports scenes and included uniforms, an artificial lawn, athletic tracks, national flags, and a soccer ball. The Flowers group comprised a wide range of colors and included floral arrangements, fruits, toy blocks, wine glasses, and bottles. Each of these groups also included color charts that were used for adjusting the white balance. The color charts included in the Sports group were evaluated separately as an object group called Charts. () Procedure The subjective evaluation experiment was conducted in dim surroundings. The display brightness was adjusted so that the white patch on the ColorChecker Classic chart in the image was cd/m when taking the patch to have a video level of %. The luminance of the wall behind the monitor was set to a value in the range of - cd/m. This setup was used to ensure that the ratio of the peak luminance of the screen image to that of the monitor background remained within a range of approximately -%. The viewing distance was set to,9 mm, which is three times the screen height. In the subjective evaluation, the reference image was Radiation intensity (a.u.) Radiation intensity (a.u.) Radiation intensity (a.u.) Wavelength (nm) Wavelength (nm) Wavelength (nm) (a) CCT=,K (b) CCT=,6K (c) CCT=6,K a.u. arbitrary unit i: Efficacy-oriented LED ii: High-saturation color rendering LED iii: High R a value and low R 9 value LED v: Reference lighting iv: High-fidelity color rendering LED Figure : Spectral power distributions for lighting conditions

3 y Rec. Rec. 79 Sports group: (a) Evaluation image x Sports group: (b) chromaticity distribution y Rec. Rec. 79 Flower group: (c) Evaluation image x Flower group: (d) chromaticity distribution Figure : Example of evaluation images and their chromaticity distributions. captured under simulated daylight conditions and four comparative images were captured under the four white-ledlighting conditions having different CRI values for each CCT. The reference image and one of the four comparative images captured under the same CCT lighting conditions were displayed side by side, divided by a gray border. The participant was able to move the gray border horizontally using a dial (Fig. ). The participants evaluated each comparative image using the degradation category rating (DCR) method with a five-stage degradation scale, as shown in Table. In this method, a score of. is considered to be the acceptability limit, and this value was used to derive the recommended values described in the next section 7) 8). Since the four comparative images were evaluated for each of the three groups (i.e., Sports, Flowers, and Charts ) at three CCTs (,,,6, and 6, K), a total of 6 Table : Five-stage degradation scale of DCR method Figure : Photograph of subjective evaluation. A participant evaluates a reference image (left side) and a comparative image (right side) displayed side by side by controlling the gray border. Score

4 trials per participant were performed. The order of presenting images to a participant was randomized to eliminate any order effect. * An index indicating the strength of the linear relationship between two variables. The correlation coefficient is a dimensionless quantity with a value between - and. Values closer to indicate a stronger relationship between the variables.. Results and discussion Figure shows the degradation mean opinion scores (s) obtained by the DCR method with radar charts for each CCT. For all CCTs and object groups, it can be seen that s exceed the. acceptability limit for (iv) high-fidelity color-rendering LED and (ii) high-saturation LED lighting conditions. Figures and 6 show the distributions of the correlation of with R a and R 9 -R, respectively, for each lighting condition. The correlation coefficient * r is shown on the upper left side of each graph. These results show that the correlation between and R 9 is the highest. This experiment confirmed that R 9 correlates most strongly with the and that choosing R 9 as an additional index is valid. Considering the compatibility with the existing index of R a, it is practically appropriate to use the combination of R a and R 9. Although it is not a perfect solution, it seems to be reasonable to use R 9 to complement R a until a new more appropriate index is standardized to prevent further mismatch with the conventional R a. Therefore, we derived the recommended value for the combination of R a and R 9 by applying the acceptability limit of.. To obtain the recommended value of each index, we plotted the distributions of each index value and the s with 9% confidence intervals *, as shown in Fig. 7. In this fig- 7 r = Ra Figure : Distribution of correlation between and general CRI value of R a. The markers denote the four white LED lighting conditions, three CCTs, and three objects shown in Fig.. Sports Sports Sports Charts Flowers Charts Flowers Charts (a) CCT=,K (b) CCT=,6K (c) CCT=6,K Flowers i: Efficacy-oriented LED ii: High-saturation color rendering LED iii: High R a value and low R 9 value LED iv: High-fidelity color rendering LED Figure : Radar charts showing for the color differences in UHDTV imagery obtained in subjective evaluation for each CCT of (a), K, (b),6 K, and (c) 6, K. Marker shapes denote the three objects. Marker sizes denote the three CCTs. 6

5 ure, the lighting conditions satisfying the acceptability limit including their 9% confidence intervals are indicated by red lines. It is sufficient to decide the recommended value so that only the lighting conditions indicated by one of the red lines are classified. For the combination of R a and R 9, R a 9 and R 9 8 are considered appropriate. Although there is a distribution that does not satisfy the acceptability limit with the condition of R a 9 alone, adding the condition of R 9 8 causes the acceptability limit to be satisfied. This condition is consistent with the conventional practice of determining the recommended CRI values in increments 8). * A range of values within which the mean value of data targeted by statistical analysis has a 9% probability of being included. r =.9 r =.89 r =.67 r = R 9 R R r =.6 r = R R R 9 Figure 6: Distributions of correlation between and special color-rendering indices R 9 -R. The markers denote the four white LED lighting conditions, three CCTs, and three objects shown in Fig.. r =.767 r = R a 6 R 9 8 Figure 7: Distributions of correlation of with general CRI value of R a and special color-rendering index R 9. The markers denote the four white LED lighting conditions, three CCTs, and three objects shown in Fig.. 7

6 . Conclusions In this study, we conducted a subjective evaluation to derive the appropriate color-rendering indices and their recommended values required for white LED lighting in UHDTV program production. From the correlation analysis between the and the general CRI value of R a and special color-rendering indices R 9 -R for each lighting type, the highest correlation was achieved by R 9. Taking these analysis results and continuity with R a into account, the combination of R a and R 9 was found to be practically appropriate and recommended values of R a 9 and R 9 8 were obtained. On the basis of the above results, appropriate colorrendering indices and their recommended values for white LED lighting in UHDTV program production have been specified in the ARIB Technical Report TR-B 9). References ) Rec. ITU-R BT.-, Parameter Values for UHDTV Systems for Production and International Programme Exchange () ) Rec. ITU-R BT.79-6, Parameter Values for the HDTV Standards for Production and International Programme Exchange () ) K. Masaoka, Y. Nishida, T. Soeno, T. Yamashita, M. Sugawara and A. Saita: Designing Camera Spectral Sensitivities for UHDTV, SMPTE Motion Imaging Journal, Vol., No. 8, pp. 6- () ) CIE Publication., Method of Measuring and Specifying Colour Rendering of Light Sources, International Commission on Illumination (99) ) JIS Z 876:99, Method of Specifying Colour Rendering Properties of Light Sources, Japanese Standards Association (99) 6) JIS Z 97:8, Recommendation for Sports Lighting, Japanese Standards Association (8) 7) FIBA Central Board: Official Basketball Rules -Basketball Equipment, International Basketball Federation () 8) IAAF Track and Field Facilities Manual Editorial Board: Technical Services, IAAF Track and Field Facilities Manual 8 Edition, International Association of Athletics Federations () 9) K. Smet, W. R. Ryckaert, M. R. Pointer, G. Deconinck and P. Hanselaer: Correlation between Color Quality Metric Predictions and Visual Appreciation of Light Sources, Opt. Express, 9, pp () ) A. David, P. T. Fini, K. W. Houser, Y. Ohno, M. P. Royer, K. A. Smet, M. Wei and L. Whitehead: Development of the IES Method for Evaluating the Color Rendition of Light Sources, Opt. Express,, pp () ) W. Davis and Y. Ohno: Color Quality Scale, Opt. Engineering, 9(), 6 () ) Tech, Method for the Assessment of the Colorimetric Properties of Luminaires-The Television Lighting Consistency Index (TLCI-), FTV-LED () ) J. Holm, T. Maier, P. Debevec, C. LeGendre, J. Pines, J. Erland, G. Joblove, S. Dyer, B. Sloan, J. di Gennaro and D. Sherlocky: A Cinematographic Spectral Similarity Index, SMPTE 6 Annual Technical Conference (6) ) Rec. ITU-R BT.-, Methodology for the Subjective Assessment of the Quality of Television Pictures () ) T. Yamashita, R. Funatsu, T. Yanagi, K. Mitani, Y. Nojiri and T. Yoshida: A Camera System Using Three -m-pixel CMOS Image Sensors for UHDTV, SMPTE Motion Imaging Journal,, pp. - () 6) S. Maeda, N. Okimoto, H. Yasui, A. Heishi, E. Niikura, K. Minami, Y. Nishida and Y. Kusakabe: RGB Laser-backlight LCD, Proceedings of the ITE Annual Convention, A- () 7) T. Hasegawa, M. Inoue and T. Mitsuhashi: Subjective Assessment of TV Picture Quality and Data Processing, Journal of Institute of Television Engineers of Japan, 7, pp. - (98) 8) J. Okamoto and T. Hayashi: Latest Trends in Image Media Quality Assessment Technologies, IEICE Fundamentals Review, 6, pp () 9) Association of Radio Industries and Businesses (ARIB), Color Rendering Indexes and Recommended Values of White LED Illuminant for UHDTV Programme Production, ARIB TR-B (6) (in Japanese) 8