Research on Color Quality Evaluation Method of White LED

1 Color rendering index and color quality evaluation of light source

Nowadays, people's requirements for indoor lighting quality are getting higher and higher, and lighting evaluation has also evolved from the past with illuminance as the main indicator to comprehensive evaluation based on comfort. The color quality of the light source is one of the important indicators that determine the indoor lighting effect.

1. 1 Defects in color rendering index (CRI)

In the past few decades, the Color Rendering Index (CRI) developed by the International Institute of Illumination (CIE) is the most commonly used index for evaluating color quality. However, with the emergence of new light sources and deep research, the existing problems have become increasingly prominent [1]. According to CIE's definition of colorrendering, it refers to the effect of a light source on the color table of an object compared to the color table under the reference illuminator [2]. The index used for evaluation is the color rendering index CRI (Color Rendering Index).

Researchers such as the United States [3,4], Hungary [5-8], and France [9-11] have found that CRI has many problems when it is used to evaluate the color rendering of light sources, especially LEDs, such as color space. Uniform, small number of color samples and low saturation. Moreover, there is a problem in evaluating the color quality by the color rendering of the light source. According to the definition of color rendering, it evaluates the color fidelity of the light source under the reference illumination body, so the color rendering index will decrease regardless of the direction in which the color table of the object deviates. However, in practice, if the light source increases the color saturation, the visual clarity and the brightness can be improved. Therefore, it is not comprehensive to use the color rendering index to judge the color quality of the light source [12]. Based on the above considerations, CIE clearly stated in the 2007 technical report [13]: The current color rendering index CRI cannot effectively reflect the color rendering of white light illumination sources including white LEDs.

1. 2 The formation of a new light source color quality evaluation system

In order to correct some problems of CRI, WendyDavis and Yoshi Ohno [14,15] of NIST in the United States developed a new evaluation system CQS (Color Quality Scale). However, the system is mainly based on mathematical calculations, and its reliability still needs to be verified by more visual experiments. For example, Nicola Pousset et al. [16] compared the experimental results of color preference with the CQS values ​​of each light source, and found that the CQS value is not satisfactory in reflecting the true color quality.

To fully evaluate the color quality of a light source, many factors need to be considered, such as color reproducibility, huediscrimination, vividness, preference for color, and visual clarity. , visual comfort and more. So far, there has not been a unified light source color quality evaluation system in the world, but one thing is certain: the color quality of the light source is determined by the human visual experience. Therefore, we need to determine the evaluation system based on a large number of visual experiments.

The main methods used in the current research can be summarized as follows: Keep the brightness of the desktop illumination or the line of sight the same; and ensure that each light source to be tested has a close color temperature or color coordinate; the visual tasks performed by the subjects are mainly: Color difference evaluation; subjective scoring of preference, vividness, coordination, etc.; color discrimination experiment to examine color discrimination. However, most of the current studies are the comparison between the visual experiment results and the color rendering index CRI, and have obtained clear conclusions. The verification of the CQS system only contains color preference, and does not contain other aspects such as color reduction, so it needs to be improved. In addition, most of the visual experiments are aimed at European populations, and the results may not be applicable to Asians due to differences in color preferences between the East and West. Therefore, in this experiment, the CRI and CQS of the light source are calculated, and the characteristics of color reduction and color preference are considered based on the visual experiment, in order to improve the reliability verification of the CQS system and evaluate the color quality. Visual experiments complement Asian samples.

2 Experimental methods

Five light sources were selected for the experiment, including two halogen lamps, two LEDs, and one compact fluorescent lamp (CFL), all of which are high-end products on the market, with halogen lamps as reference sources at the same color temperature. The basic information of the light source is shown in Table 1. According to the color temperature, the five kinds of light sources are divided into two groups of 6500K group and 5000K group, and the visual experiment is performed separately. First, the relative spectral power distribution of each light source was collected. See Figure 1 and Figure 2. The spectrum acquisition instrument is the STC3000 spectrometer, and then the CRI and CQS indices are calculated based on the spectrum.

Visual experiments were performed in a dark room and made into two separate compartments, as shown in Figure 3. The top of the compartment is equipped with a light source that does not produce glare and is evenly illuminated on the table. The average illumination of the desktop in both compartments is 500lx.

The experiment used the method of within-subjects design to select 10 subjects, 5 males and 5 females, aged 20 to 24 years old, all with normal vision or corrected visual acuity, and screened for colorless blindness. Wait for other eye diseases. Using the reversal experimental condition design, each subject participated in three sets of experimental tasks under all experimental conditions. The first item is the score of the color difference. The MCC color card used in the experiment is a standard method specified by the CIE to compare the color difference between the light source to be tested and the reference source. During the experiment, the MCC color card was placed on the desktop of the two compartments. The light source to be tested (CFL or LED) was placed on the side of the compartment, and the halogen lamp of the same color temperature was placed as the reference light source on the other side, so that the subjects observed the two sets of color cards. 24 pairs of color samples, and score the color difference of each pair. The rule of scoring is to first classify the chromatic aberration into VS (very small), S (small, small), M (medium, medium), L (large), VL (very large). There are five broad categories (Table 2), followed by five levels of subdivision, and finally expressed in the form of VS4, L2. In order to facilitate statistical data, the statistics are converted to the number of the last line 1 to 25.

All three experiments were evaluations of preference. Simulated flowers and fruits are placed in the compartments for easy observation and judgment. The second is a separate rating for each light source, ranging from 0 (very dislike) to 100 (very like), and from 1 (very dislike) to 5 (very like). The third item is a pairwise comparison, which allows the subject to select a preferred source number.