A Novel Laser-Based Display and Imaging Combined System

Laser-based displays can offer a wide color gamut to the viewers due to the narrow bandwidth of laser source, which is quite different from traditional display systems. Therefore, the color rendition standards for traditional displays are no longer suitable for laser-based displays. An optimized and practical color reproduction system is proposed and demonstrated for laser-based displays in this paper. Not only the three primary wavelengths are well chosen by considering the luminance efficiency, color gamut, and color matching function, but also the response function of a camera system is optimized for a simplified color reproduction procedure. A remarkably high color rendition quality is presented in our experiment compared with traditional displays.

SECTION 1. Introduction

The concept of using lasers as light sources in display systems can date back to the 1960s [1], but due to the lack of practical laser devices, the laser-based display didn't get much improvement for half century. In recent years, the laser display technology has become a hotspot due to the commercial success of practical laser diode in visible wavelength and diode-pumped-solid-state lasers [2]. Furthermore, with the development of InGaN-based laser diodes towards longer wavelength [3], the commercial applications of laser display have made a great improvement.

The most attractive advantage of laser-based display is the wide color gamut that it can offer to the audience [4]–​[6]. Distinguished from LED or xenon lamp, the color of laser is pure and its chromaticity coordinates lie on the spectrum locus of gamut charts. It can be more than 70% of the total range of colors perceivable by humans due to the monochromaticity of laser source [7]. Except for two-dimensional display, laser is also the ideal light source for holographic three-dimensional display [8]. At the same time, the old standards (e.g., NTSC, EBU) are no longer suitable for laser-based displays. The color rendition quality is not satisfactory if we continuous to use the NTSC signal in a laser-based display directly [9]–​[11]. Some works have been done for achieving high color quality by complex data processing method, which will result in long processing time and complex circuit [12]–​[20].

In this paper, a novel color rendition procedure is presented. The color signals are captured by our modified camera system and then displayed by our modified laser display system directly. No data processing is needed while preserving the image's high quality. Not only the three primary laser sources but also the response function of camera system are optimized in Section 2. The setup of our experiment and the comparison of the color rendition quality are shown in Section 3. Coordinating our specially designed camera system with our modified laser based display system, a simplified and high quality color rendition can be realized.

SECTION 2. Analysis for the Laser Sources

As shown in Fig. 1, the procedure of color reproduction can be divided into two parts: camera system for recording the color data and display system for showing the reproduced color.

Fig. 1.

The procedure of color reproduction can be divided into two parts: camera system for recording and display system for showing.

Generally, we hope the tristimulus values of the reproduced color can be the same as the original one, which is also called as Luther condition [21]. In order to meet this condition, the ideal response function of camera system can be obtained from following equation: $$\begin{equation} \left[ {\begin{array}{c} {\tau _R^o}\\ {\tau _G^o}\\ {\tau _B^o} \end{array}} \right] = {\left[ {\begin{array}{lll} {{X_R}}&{{X_G}}&{{X_B}}\\ {{Y_R}}&{{Y_G}}&{{Y_B}}\\ {{Z_R}}&{{Z_G}}&{{Z_B}} \end{array}} \right]^{ - 1}}*{K^{ - 1}}*\left[ {\begin{array}{l} {{{\bar{x}}_{1931}}}\\ {{{\bar{y}}_{1931}}}\\ {{{\bar{z}}_{1931}}} \end{array}} \right]\tag{1} \end{equation}$$ where ${\tau ^0}$ is the ideal response function of camera, $XY{Z_{RGB}}$ is the coordinate of the three primary colors (RGB) for displaying in the CIE (Commission International de L'Eclairage) 1931 chromaticity plot, ${\overline {xyz} _{1931}}$ is the color matching function in CIE 1931 standards and K represents the nonlinear data processing. It is worth noticing that the ideal response function of camera is decided by the three primary colors and the nonlinear data process. In the following works, the three primary colors will be designed at first. In general, important factors contain the color gamut, the wall plug efficiency (WPE) of laser source and the luminance efficiency. Furthermore, by choosing appropriate wavelengths to optimize the ideal response function of camera system, the influence of the nonlinear data process can also be minimized to achieve a simpler color rendition process.

2.1 Red and Blue

For the consideration of portability and minimizing the system's complexity, the laser diode is thought to be the perfect light source for laser display. In this paper, we hope that the display system combined with imaging system can realize a wider color gamut, a higher luminance efficiency, and a better color rendition at the same time. We are more concerned about the function of the whole system, not merely the display system. Thus, the three primary wavelengths will not be designed deliberately to compare with the existing ones, such as the Rec. 2020. Considering the development levels of laser sources, products are varied and mature enough for the red and blue LDs. While for the green laser, few products could meet the requirements and their performances have little difference [22]–​[24]. We firstly optimize the red and blue laser sources and consider more about luminance efficiency and color gamut. The problem is how to choose the appropriate wavelengths.

As shown in Fig. 2(a), the black triangle is a typical color gamut for Rec. 2020, which is formed by the three primary wavelengths of 467 nm, 532 nm, and 630 nm. The color gamut increase when the wavelength of primary red source moves towards the longer or the wavelength of primary blue source moves towards the shorter. The color gamut used here is defined as: ${\rm{Color}}\;{\rm{Gamut}}\;{\rm{Area}} = {{\rm{A}}_{{\rm{display}}}}/{{\rm{A}}_{{\rm{standard}}}}$ [25]. But since the sensitivity of human eye varies with wavelength (shown in Fig. 2(b)), the blue and red wavelengths tuning with a bigger slope will influence the efficiency more greatly than the green wavelengths tuning with a smaller slope. The wider color gamut will result in a lower luminance efficiency. There is trade-off between the color gamut and the luminance efficiency. Furthermore, we need to pay attention to the wall plug efficiency (WPE) of laser diode [26]. As the technologies of red and blue laser diode get mature, the WPE of different commercial laser diodes need to be considered. As discussed in paper [27] , the suitable wavelength of the laser diodes for the red sources are from 635 nm to 645 nm and for the blue sources are from 440 nm to 450 nm. Considering the real working state of the light source of laser display system, such as the working life, the stability, and the utilization efficiency, all LDs in this experiment are set to work at typical parameters. Here we define an efficiency ${L_{st}}$, which means the luminance when the electric input power is 1 watt. The efficiency ${L_{st}}$ and associated color gamut of some typical products are shown in Table 1 and 2. The color gamut is calculated when the other two sources are fixed. The color gamut coverage rate is defined as the ratio of the color gamut formed by three monochromatic laser sources to the whole color gamut formed by the chromaticity plot.

Fig. 2.

There is a trade-off between the color gamut and the photopic sensitivities. (a) The color gamut contained in the triangle formed by three monochromatic laser sources will increase with a longer red wavelength or a shorter blue wavelength. (b) The photopic sensitivities change greatly with the wavelengths of blue and red laser sources. Either a longer red wavelength or a shorter blue wavelength will drastically decrease the photopic sensitivities.

Table 1 The parameters of red laser diodes commercially used on the market (The color gamut is calculated when the green source is fixed to 535 nm and the blue source is fixed to 450 nm)

Table 2 The parameters of blue laser diodes commercially used on the market (The color gamut is calculated when the green source is fixed to 535 nm and the red source is fixed to 635 nm)

As the results shown in Table 1 and 2, the color gamut varies little compared with the variation of ${L_{st}}$. Taking both factors into consideration, we can find that HL63193MG of 638 nm wavelength and PL TB450B of 450 nm wavelength are the best choices for red and blue laser sources respectively.

2.2 Green

As shown in equation (1), the ideal response function of camera system is affected by the coordinates of the three primary RGB colors. In traditional display systems, the primary RGB sources are phosphor which will result in the negative value of the response function curve. Since the response function of real camera system can only be positive, a nonlinear data processing K needs to be applied for high quality color rendition as shown in equation (1). This method requires time for processing, which is not friendly to high-speed photography and will increase the complexity of camera system. In laser-based displays, we want to develop a color rendition procedure without the nonlinear data processing since the response function of camera system has the potential to get a minimized negative area. In this section, we will focus on choosing an appropriate wavelength for green laser diode by optimizing the ideal response function of camera system and the color gamut.

The optimization is done by traversal method. The wavelengths of green lasers vary from 515 nm to 535 nm while the red source is fixed to 638 nm and the blue source is fixed to 450 nm as the results mentioned above. We compare the negative area of the ideal response function. The best wavelength of green laser is selected when the negative area reaches to its minimization. As shown in Fig. 3, 525 nm is the best choice for green laser source. The curve in Fig. 3(a) shows that not only the response function is optimized, but also the coverage ratio of color gamut remains more than 70% when the wavelength is 525 nm.

Fig. 3.

(a) The negative area of the ideal response function reaches to its minimization when the wavelength of green laser is 525 nm while the coverage ratio of color gamut remains more than 70%. (b) The ideal response function of camera system after optimizing the wavelength of green laser.

As can be saw in Table 1 and 2 , different blue and red LDs will affect the color gamut and efficiency greatly. At the same time, as can be saw in Fig. 3(a), the green LDs will influence the negative area more greatly.

SECTION 3. Experiments and Results

3.1 Set Up of the System

In order to meet the requirement of the optimized response function of camera system, a monochromatic CMOS (DCC1545M, THORLABS) and specially-made color filter (supplied by ANDOVER CORP.) are used to compose the camera system. The response function $\tau$ of our camera is defined as: $$\begin{equation} \tau (\lambda) = T(\lambda)S(\lambda)\tag{2} \end{equation}$$ where T is transmittance of the color filters and S is the sensitivity to different wavelength of the CMOS. The comparison of the response function of our real camera and the ideal one is shown in Fig. 4. Noticing that the curves have been normalized for comparison.

Fig. 4.

Comparison between the real and ideal response function of camera system.

As the results in Section 2, the wavelengths of the three primary colors (RGB) are chosen as 638 nm (HL63193MG, OCLARO), 525 nm (NDG7K75T, NICHIA) and 450 nm (PL TB450B, OSRAM). The normalized spectrum of our laser sources is shown in Fig. 5. To eliminate the speckle in the laser-based display system, a dielectric elastomeric actuator (DEA) is used to reduce the spatial coherence of laser sources [28].

Fig. 5.

The spectrum of three primary laser sources for our laser-based display system.

3.2 Color Rendition Quality

The laser based projector contains the chosen laser sources, speckle eliminating component, collimating lens, an LCoS (SVGA, 800*600) and projection lens. There are three steps to test the color rendition quality.

First step: Adjusting the output power of each laser source.

The output of each primary laser source is decided by the display white point. As usual, the white point is D65 defined by CIE (chromaticity coordinates x = 0.31271 y = 0.32902). The ratio needs to be fine-tuning by white point matching. The white color chart is provided by PANTONE. The color data of white is captured by our optimized camera. Then the laser based projector we built will be used to display the white image. We use a spectrometer (USB2000+, Ocean Optics) to measure the displayed color. The ratio of each laser power is then carefully adjusted to make the reproduced white color same as D65 in the chromaticity coordinates. At last, the ratio is: $$\begin{equation} {P_R}:{P_G}:{P_B} = 1:0.35:0.24\tag{3} \end{equation}$$

Second step: Measuring the original color data.

We choose seven colors randomly from the color charts as shown in Fig. 6 . The seven colors should be away from each other in the chromaticity chart. To avoid the influence of other colors when measuring the parameters of one color. The colors are measured one by one by using the pantone plus series. The coordinates of these original colors are measured in a standard light environment. The standard source here we use is D65 source as defined by CIE standards.

Fig. 6.

The colors used for testing. The seven colors should be away from each other in the chromaticity chart.

Third step: Measuring the color rendition quality.

The digital data of the colors in step 2 will be taken by the optimized camera and input into our laser display system without any data processing. The coordinates of these displayed colors are measured by the spectrometer then. Furthermore, we make a control group represents the traditional displays. The camera is 5D II (CANON) and the display is a DLP projector (X1220H, ACER). The main parameters are: the light out is 2700 lumens; the resolution is 1024 × 768; and the contrast ratio is 3000:1.

In order to compare the color rendition quality of the two groups, color difference ${\rm{\delta }}$ is used [29] . The CIE 1976 UCS is a uniform color space. ${\rm{\delta }}$ is defined as the geometrical distance of two colors in the CIE 1976 UCS chromaticity chart. $$\begin{equation} {\rm{\delta }} = {\left({\delta _L^2 + \delta _C^2 + \delta _H^2} \right)^{1/2}}\tag{4} \end{equation}$$

Where ${\delta _L},$ ${\delta _C}$ and ${\delta _H}$ represent the differences of lightness, chroma, and hue of two colors respectively. This is an intuitive and simple way to calculate the color difference. In Table 3, ${\delta _{laser}}$ is the color difference between the original and the reproduced colors by using laser-based display while ${\delta _{tr}}$ is by using the traditional display.

Table 3 The comparison of color difference between the optimized laser-based display and the traditional display. (Color 1 is the first color bar from the right in Fig. 6, and then color 2 is the second one and so on)

The results in Table 3 clearly shows a remarkable high color rendition quality. Averagely, a 50% reduction of color difference can be realized by using our designed system. Compared with the traditional displays, not only the supporting color gamut of laser display is much bigger, but also the color rendition quality can reach a higher level. What’ more, compared with Rec. 2020, our system can realize a wider color gamut, due to the area of white triangle formed by the three primary wavelengths are bigger as shown in Fig. 2(a). The color difference will also be smaller, because our initial purpose is to abandon the gamut mapping. It is worth noticing that we apply no data processing method in our color rendition procedure, which means it's suitable for high-speed photography in special conditions.

SECTION 4. Conclusions

In this paper, both the three primary laser wavelengths and the response function of camera system are optimized for a higher color rendition quality in laser-based displays. The wavelengths of red and blue laser are chosen from commercial products by considering both the luminance efficiency and color gamut. The comparison shows that the best choice for the red wavelength is 638 nm (HL63193MG, OCLARO) and for the blue wavelength is 450 nm (PL TB450B, OSRAM). Furthermore, the wavelength of green laser is optimized for minimizing the negative area of the ideal response function of camera system. By choosing 525 nm as the wavelength of green laser, the negative area of ideal response function reaches to its minimization. We use specially-made color filters and monochromatic CMOS to make the real camera response function the same as the ideal one as far as possible.

Our system have a big change in the image capture and display pipeline, which could bring some unique advantages over traditional display systems. Firstly, without any data processing, the signals captured by our optimized camera can be inputted to our modified laser-based display and displayed directly. Normally, the process of displaying includes: capturing the original image, encoding with pre-defined color profiles, gamut mapping (optional, depending on whether the color profiles is compatible with the display system) and displaying. Using our method, just two steps are needed: capturing and displaying, which can greatly simplify the process. Neither computing nor data processing is needed, which means high-speed photography and live broadcast with high quality signals could be achieved. Secondly, our method can reduce the color difference and preserve the high quality of color signals. The system has an even better performance in color rendition due to the color gamut mapping is abandoned. Thirdly, our method can avoid the confusion caused by the disunity of standards. The basic progress from imaging to display include: capturing the color signals, encoding the colors with pre-defined color profiles (sRGB, Rec. 709, Display P3 and Rec. 2020 etc.), converting the color signals according to the characteristics of display system, and displaying. It is worth to notice that, except for time consuming, both the encoding and converting have various standards, which have not yet formed a unified opinion. This may cause some confusing results in color management, such as “the definition of color gamut is sometimes confusing and misleading [25]. The disputes brought by various display standards and gamut mapping methods could be handled well in our method. In conclusion, our system can simplify the process from capturing the color signals to displaying greatly and avoid the confusion caused by the disunity of standards. The color difference is much less and a remarkable high color rendition quality is realized compared with the traditional display. We think this result provides new thought for real color rendition in laser-based display systems.