Open Access
Add to BibSonomy
Abstract
A compact near-eye display with a 60° horizontal field of view, wide eye box of 5 mm, and high resolution of 720 p is proposed and developed by combining a transmission hologram that duplicates the beam of a scanning display and a reflection hologram that reflects duplicated beams toward the user’s eye. The feasibility of the proposed near-eye display is demonstrated by examining the specifications and exposure of 24 multiple holograms. A compact NED that can display images with a horizontal FOV of 60° and that has a 6.2 mm × 4.8 mm eye box and 720 pixels vertical resolution is achieved.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1 Introduction
A near-eye display (NED) is an essential device for achieving augmented reality/mixed reality that overlays virtual images on the real world. NEDs must be compact to increase their portability and allow them to be worn for extended periods. They must also have a wide field of view (FOV) so that virtual information can be overlaid anywhere in the FOV. Because the horizontal FOV required to receive valid information is 60° to 90° [1], the horizontal FOV of a NED should be at least 60°. A conventional NED technique uses a concave half-mirror [2–5]. This method can expand the horizontal FOV to 60°; however, there is a tradeoff between increasing the FOV and reducing the NED size. In other words, it is difficult to achieve both a wide FOV and small size.
Although NEDs using a waveguide (exit pupil expansion) can achieve a small size and wide eye box with a relatively thin lens [6–8], the refractive index of the waveguide limits the expansion of the FOV, making it difficult to achieve a horizontal FOV of ≥60°. To widen the FOV, there is a method that divides the FOV and layers the waveguide [9]; however, the lens area in front of the eye becomes thicker. Some previous studies achieved a compact NED with a horizontal FOV of ≥60° [10,11]; however, the eye box was very narrow, and users could not see the images if they moved their eyes slightly.
Investigations for achieving a wide eye box while maintaining a small size and wide FOV can be found in the literature. To enlarge the eye box, one study used a hologram array with the function of an off-axis concave mirror [12], which was as small as 2 mm2. Another study used two holograms with the diffusion function and concave mirror function [13]. This technique expanded the eye box by reflecting the light of pixels diffused by a diffuse hologram onto the concave mirror hologram. Moreover, some studies have used multiple-exposure holograms and a spatial light modulator to display three-dimensional content [14–16]. However, these methods resulted in low resolution.
As mentioned above, no conventional method satisfies all of the following requirements: (i) small size, (ii) horizontal FOV of ≥60°, (iii) wide eye box, and (iv) high resolution. Our study aims to develop a NED optical system that satisfies all of these requirements. In this paper, we propose a method to widen the eye box by creating multiple Maxwellian images with a wide FOV by multiple-exposure holograms. We also experimentally verify the feasibility of the proposed method.
2 Proposed NED
Our proposed method uses a scanning display and two holograms: a transmission hologram and a reflection hologram (Fig. 1(a)). Both holograms are volume holograms. The transmission hologram is a multiple-exposure hologram that duplicates the beam from a scanning display to create several images (Fig. 1 (b)). The reflection hologram reflects each beam duplicated by the transmission hologram toward the user’s eye (Fig. 1(c)). Since the FOV of an image is determined by the numerical aperture of the reflection hologram that diffracts the beam, the FOV is determined by the numerical aperture of the lens used in the exposure of the reflection hologram. Therefore, a large FOV can be achieved by using a large numerical aperture lens when exposing the reflection hologram. Multiple exit pupils are created because the beam from the scanning display is duplicated by the transmission hologram. The multiple exit pupils enlarge the eye box. The image seen through one exit pupil is the Maxwellian image because the beam’s diameter, which corresponds to one pixel of the scanning display, is sufficiently narrow and is not refracted by the eye lens (the beam has no focus point). A NED optical system comprising a thin glass substrate and two holograms on both sides of the substrate can achieve a wide FOV of >60° and an enlarged eye box.
Fig. 1. Proposed near-eye display system. (a) Overview of the near-eye display system. This system uses a scanning display and two holograms: a transmission hologram and a reflection. (b) Transmission hologram. The transmission hologram duplicates the scanning beam. (c) Reflection hologram. The reflection hologram reflects each beam duplicated by the transmission hologram toward the user’s pupil.
In the conventional eye box expansion method using multiple-exposure holograms [14–16], the optical axes of multiple converging lights, which are the object lights for recording multiple-exposure holograms, are not parallel, resulting in the display of multiple shifted images if two-dimensional images are displayed (Fig. 2 (a)). In this study, the eye box is expanded by aligning the optical axes of the object lights (converging lights) recording the reflection holograms so that beams corresponding to the same pixel are parallel (Fig. 2 (b)). The following sections discuss the fabrication of the NED optical system described above as well as the verification of the feasibility of the proposed method.
Fig. 2. Reconstructed image when the hologram is irradiated with two-dimensional (2D) image light. (a) Eye box expansion using conventional multiple exposures. The images appear to be shifted because light comes from different directions. (b) Eye box expansion proposed in this study by multiple exposures of the reflection hologram so that the convergent light reflected in the direction of the eye is aligned in parallel.
3 Specifications of NED and holograms
Figure 3 presents an overview of the developed NED. The horizontal FOV is 60°, and the distance between the reflection hologram and the eye is set to 15 mm so that the NED can be used as a regular pair of glasses. The beam emitted from the scanning display is incident on the transmission hologram at the same angle regardless of the scan angle using a lens. Collimated beams (beam from the scanning display) incident on the transmission hologram are incident at an angle of 30° to the hologram surface to avoid hitting the user’s face.
Fig. 3. Near-eye display developed in this study. The horizontal FOV is 60°, and the distance between the reflection hologram and the eye is set to 15 mm. The reflection holograms are multiply exposed. One of the reflection holograms to be multiply exposed has a size of 17.5 mm × 10 mm. The beam emitted from the scanning display is incident on the transmission hologram at the same angle regardless of the scan angle using a lens. The beam from the scanning display incident on the transmission hologram is incident at an angle of 30°. Quantitative specifications of the image pitch, eye box size, recording angle for multiple exposures, and parallelism of reflected beams are examined.
The reflection holograms are multiply exposed, where only one of the many beams duplicated by the transmission hologram diffracts (reflects) independently due to the angular selectivity of the volume hologram. One of the reflection holograms to be multiply exposed has a size of 17.5 mm × 10 mm, and multiple-exposure holograms are created by shifting the 17.5 mm × 10 mm hologram in parallel at a 1-mm pitch on one sheet of the hologram material.
3.1 Pitch of Maxwell image
It is desirable that two or more image beams always enter the pupil so that multiple image beams reflected by the reflection holograms appear to be connected. Therefore, the pitch for shifting the reflection holograms (image pitch) should be 1 mm.
3.2 Eye box
The targeted eye box size is 5 mm (width) 3 mm (height) based on human physiological data. This eye box size is the size at which all 60° image beams enter the pupil when the FOV is 40° such that humans can see with only eye movements without moving their head [17]and the pupil size is 2 mm (in bright areas).
3.3 Parallelism of reflected beams
In the proposed method, the beams of pixels displaying the same image must all be parallel. In other words, all beams reflected in the direction of the eye by the reflection hologram must be parallel. Since the resolution of the user’s eye with a visual acuity of 1.0 is 1 arcmin, the deviation from parallel should be less than 1 arcmin (Fig. 4).
Fig. 4. Relation between the observed virtual image and the deviation from parallel of beams corresponding to the same pixel. If the beams are not parallel, users see multiple shifted images. The deviation from parallel should be less than 1 arcmin.
3.4 Recording angle for multiple exposure
To achieve an eye box size of 5 mm (horizontal) × 3 mm (vertical), a hologram is created that reconstructs 6 horizontal × 4 vertical = 24 images with a 1-mm pitch. The NED is an optical system with 24 exit pupils.
To investigate the angle at which each of the 24 multiple-exposure reflection holograms diffracts independently without responding to any light other than the reference light incident at the time of exposure, the angular selectivity was calculated based on the coupled wave theory proposed by Kogelnik [18]. The hologram material used was Covestro HX200 photopolymer (thickness = 16 µm, refractive index modulation Δn = 0.03) [19].
For each of the 24 reflection holograms, the exposure angle was determined so that the diffraction efficiency of the light coming from 23 directions other than the reference light was determined to be 0.5 or less when the diffraction efficiency at the Bragg match angle (reference light angle = reconstruction light angle) was set to 1.
Table 1 presents the angles of the reference lights of the reflection hologram (object lights of the transmission hologram). The individual circles in Fig. 5 denote the center locations of each of the 24 multiplexing reflection holograms. Based on the angles in Table 1, the glass thickness between the transmission and reflection holograms was designed to be 1.1 mm so that the pitch of the image was 1 mm.
Table 1. Exposure Angle of the Object Lights of the Transmission Hologram and the Reference Lights of the Reflection Hologram
Fig. 5. Center position and angle of 24 reflection holograms. The same letters (A-E) represent the same elevation angle. The numbers (1-8) are the number of holograms exposed at the same elevation angle (different azimuth angles).
4 Overview of hologram exposure
A transmission hologram duplicates the beam of a scanning display incident on it in multiple directions. The transmission hologram is recorded by exposure to a reference light (collimated light), corresponding to a beam from the scanning display aligned in parallel by a lens, and to multiple object lights (collimated lights), corresponding to beams in multiple directions that are duplicated. The optical axes of all collimated lights intersect at a point on the hologram surface. The reference light is incident at an angle of 30° to the hologram surface (Fig. 6(a), (b)).
Fig. 6. Recording and reconstruction of the transmission hologram and the reflection holograms. (a) Recording of the transmission hologram. The transmission hologram is recorded by exposure to a reference light and multiple object lights. (b) Reconstruction of the transmission hologram. A transmission hologram duplicates the beam of a scanning display incident on it in multiple directions. (c) Recording of the reflection hologram. The lens for the object light moves in parallel to match the position of the reference light. (d) Reconstruction of the reflection hologram.
Reflection holograms are exposed at multiple angles, where one hologram is recorded and then another hologram is recorded at a different angle of incidence of the reference light on one sheet of hologram material. The reference lights of the reflection holograms are the same collimated lights as the object lights of the transmission hologram, and the object lights are convergent. The lens for the object light of the reflection hologram moves in parallel to match the position of the reference light (object light from the transmission hologram) (Fig. 6 (c), (d)).
After recording the transmission and reflection holograms, they are affixed to both sides of a thin sheet of glass, and a beam from a scanning display is incident on the holograms to enable image observation (Fig. 7).
Fig. 7. Combined transmission and reflection holograms. Users can observe virtual image when beam from a scanning display is incident on the holograms.
5 Ghost phenomenon
If the diffraction efficiency of the reflection hologram is low, depending on the angle of the beam coming from the transmission hologram, the beam that is not diffracted by the reflection hologram propagates through the glass by total reflection and diffracts again in another location in the reflection hologram. This rediffracted beam causes the ghost phenomenon because a beam from a different direction also appears in addition to the beam in the correct direction (Fig. 8 (a)).
Fig. 8. Ghost phenomenon problem. (a) Mechanism of ghost phenomenon. The beam that is not diffracted by the reflection hologram propagates through the glass by total reflection and diffracts again in another location in the reflection hologram. (b) Louver with alternately layering a light-absorbing layer and a resin with a refractive index equivalent to that of the hologram. The louver absorbs beams not diffracted by the reflection hologram. The louver was attached to the back of the reflection hologram with respect to the user’s eye position.
A louver that absorbs beams coming from the angles presented in Table 1 and allows a view of the real-world scene was designed by using the optical simulation software Zemax. As illustrated in Fig. 8(b), the louver was created by alternately layering a light-absorbing layer and a resin with a refractive index equivalent to that of the hologram. The louver was attached to the back of the reflection hologram with respect to the user’s eye position. The results of the Zemax calculation demonstrated that when the thickness of the absorbing layer was 50 µm, all beams from the transmission hologram coming from multiple angles could be absorbed if the louver thickness was at least 4/3 times thicker than the pitch of the absorbing layer.
The thickness of the absorbing layer (louvers) may narrow the vertical FOV. In addition, if the pitch is narrow, multiple diffraction occurs and the real-world scene (the outside view) becomes a multi-image. However, the louver is designed so that the FOV and diffraction are not affected, and the louver, which has a thickness of 0.94 mm and an absorption layer pitch of 0.7 mm, is constructed. Using louvers, the outside view appears 40% darker (Fig. 9).
6 Experiment
Based on the specifications obtained in Section 3, we created a NED optical system and verified the feasibility of the proposed method through experiments. Figure 10 presents the exposure system for the transmission and reflection holograms. One optical exposure system was constructed because the object lights of the transmission hologram and the reference lights of the reflection hologram were the same. As the first step, monochromatic green (532 nm) holograms were exposed using an Azur 5W continuous wave fiber laser (line width < 200 kHz).
The transmission hologram was exposed by simultaneously irradiating a collimated reference light with an intensity of 2.16 mW/cm2 and 24 collimated object lights with an intensity of 0.09 mW/cm2 each for 4 s without pre-exposure. The reflection holograms were created by stacking three hologram materials with eight multiple-exposure reflection holograms on one hologram material. This was because it is difficult to expose 24 multiple-exposure holograms on a single Covestro material, and furthermore, 24 multiple exposures on a single hologram material result in low diffraction efficiency. The recording sensitivity characteristics of the hologram material were obtained through a preliminary experiment (Fig. 11), and based on these characteristics, the exposure schedule for eight multiplexes was adjusted through several exposure experiments (Table 2). The room temperature was set to 25°C ± 0.2°C, and the exposure system was covered during exposure to avoid the influence of air vibration.
Fig. 11. Recording sensitivity characteristics of the hologram material. Pre-exposure is performed until the sensitivity curve rises. The cumulative diffraction efficiency is divided into eight equal parts so that eight multiplex holograms (eight reflection holograms) have the same diffraction efficiency, the required exposure energy(laser power × time) is calculated from the graph, and the schedule is determined.
Table 2. Exposure Schedule
7 Results
After the exposure experiment, the angular selectivity of the reflection holograms was determined. In addition, the exposed transmission and reflection holograms were combined, and the virtual image observed with the NED system was evaluated.
7.1 Angular selectivity of reflection holograms
The exposed reflection holograms were irradiated one at a time by 24 reference lights in a phase conjugate configuration. The angular selectivity was determined by observing the number of object lights reconstructed for a single reference light. The reconstructed object lights were converged using a lens and captured with a camera (Fig. 12(a),(b)). As a result, there were holograms in which some object lights were reconstructed for a single reference light (Fig. 12(c),(d)).
Fig. 12. Angular selectivity of reflection holograms. (a) Method of verifying the angular selectivity. The object light of the reflection hologram reconstructed by phase conjugation is focused by a lens and captured by a camera. The number of spots reconstructed by one reference light is counted. (b) Reconstructed object light spots reconstructed when 24 reference lights are irradiated onto the reflection hologram at once. (c) Reconstructed light spot when the reflection hologram B1 is irradiated by the reference light (no problem with angular selectivity). (d)Reconstructed light spots when the reflection hologram B5 is irradiated by the reference light (problem with angular selectivity).
7.2 Evaluation of NED system
The virtual image was evaluated by irradiating scanning display images onto a combination of exposed transmission holograms, reflection holograms, and louvers. Figure 13(a) presents the evaluation system. A MicroVision ShowwX+ HDMI projector (848 × 480 pixel resolution, 530 nm center wavelength) was used as the scan display, while a Ultimems scan display (720 p resolution, 519 nm center wavelength) was used for resolution evaluation only. A See3CAM_130 camera (1080 p resolution, 60.2° horizontal viewing angle) manufactured by e-con Systems India Pvt Ltd was used to capture the virtual image.The scanning display image was projected onto the combined holograms, and 24 exit pupils could be observed (Fig. 13(b)). The virtual images that can be observed from each of the 24 exit pupils are shown in Fig. 13(c). The ghost phenomenon was eliminated by the louver (Fig. 13(d),(e)).
Fig. 13. Evaluation of NED system (a) Evaluation system. (b) Reconstructed light when the scanning display beam is incident on the hologram. 24 exit pupils could be observed. (c) The virtual images that can be observed from each of the 24 exit pupils (d)Observed image without louver (e) Observed image with louver (effect of the louver). The ghost phenomenon was eliminated by the louver.
Next, we discuss the NED optical system size, FOV, eye box size, and resolution. We also evaluate the image position shift and ANSI contrast.
7.2.1 Size of NED optical system
The thickness of the optical system, which is a combination of the reflection and transmission holograms, louver, and reinforcing glass, was approximately 4.1 mm(Fig. 14). All components are laminated together. Two pieces of 1-mm glass for reinforcement were used for handling during the experiment; however, the thickness of this glass can be reduced. The length and width of the glass can be reduced to the hologram exposure area (22.5 mm × 13 mm) by cutting the glass.
Fig. 14. Configuration of near-eye display optical system. The reflection and transmission holograms, louver, and reinforcing glass are combined.
7.2.2 FOV
A camera with a horizontal viewing angle of 60.2° was placed at the eye relief position to capture a virtual image (Fig. 15 (a)). Although distortion was observed due to the effect of the lens of the evaluation system, the virtual image was displayed over the entire viewing angle of the camera, indicating that a horizontal FOV of 60° was achieved. Figure 15(b) illustrates a virtual image of Nreal Light augmented reality glasses. (diagonal FOV of 52° (published), horizontal FOV of 45° (calculated) [2]) with the same camera under the same condition. The camera was placed in the eye relief position, but the same image can be seen wherever the camera is placed within the eye box area (Fig. 16). If the camera is placed is outside the eye box area, part of the image will be vignetting and the user cannot observe the entire image.
Fig. 15. Observed virtual image (a) Virtual image of proposed near-eye display system. A horizontal FOV of 60° was achieved. (b)Virtual image of Nreal Light augmented reality glasses (diagonal FOV of 52° (published) and horizontal FOV of 45° (calculated) [2]).
7.2.3 Eye box size
The size of the eye box was measured by a camera placed at the eye relief location. The area where the virtual image could be observed at a FOV of 60° was measured by moving the camera vertically and horizontally. The measured size was 6.2 mm (horizontal) 4.8 mm (vertical). The target size of 5 mm (horizontal) 3 mm (vertical) was thus achieved. The reason why the eye box was larger than designed is that the size of each exit pupil was large.
7.2.4 Resolution
The resolution of the MicroVision ShowwX+ HDMI projector was 848 × 480 pixels, which is not sufficient for large-screen NEDs; therefore, we switched to a Ultimems projector with a resolution of 1280 × 720 pixels to evaluate the resolution. The measured center wavelength of this projector was 519 nm. The incident light of the scanning display image was changed to 23° (Fig. 17(a)) because the wavelength of the scanning display beam for reconstruction was significantly different from the exposure wavelength. The exposure angle was 30°. For evaluation, a resolution chart with black-and-white line pairs (Fig. 17(b)) was used to determine the number of black-and-white line pairs that could be resolved.
Fig. 17. Resolution evaluation (a) System for evaluating the resolution of the proposed near-eye display. The scanning display was switched to a Ultimems projector with a resolution of 720p(1,280 × 720 pixels) to evaluate the resolution. The incident light of the scanning display image was changed to 23° because the reconstruction wavelength significantly differed from the exposure wavelength. When verifying the vertical and horizontal resolutions, the distance d between the lens and hologram was adjusted to the distance with the highest resolving power. (b) Resolution chart for evaluating vertical resolution.
The image evaluation optics used in this experiment and the off-axis concave mirror (exposed hologram with collimated and convergent lights) produce astigmatism when collimated light is incident on them. Therefore, when verifying the vertical and horizontal resolution, respectively, the distance d between the lens and the hologram was adjusted to the distance with the highest resolving power.
The results are presented in Fig. 18. Due to poor quality holograms, it was not possible to evaluate the resolution of the entire screen. When evaluating the area where the image was visible, 360 horizontal line pairs (vertical resolution of 720 pixels) and 240 vertical line pairs (horizontal resolution of 480 pixels) were resolved. The horizontal resolution was 320 pixels (160 line pairs) when the vertical resolution was 720 pixels. The vertical resolution could be displayed without loss of the scanning display resolution; however, the horizontal resolution could not be displayed at 1280 pixels, which was the scanning display resolution.
Fig. 18. Resolution evaluation results. When evaluating the area where the image was visible, 360 horizontal line pairs (vertical resolution of 720 pixels) and 240 vertical line pairs (horizontal resolution of 480 pixels) were resolved. The vertical and horizontal resolution evaluation areas are FOV 1° and FOV 3°, respectively.
7.2.5 Image position shift
If beams displaying the same image reconstructed by the reflection holograms were not aligned in parallel, the position of the observed image would appear to be shifted in each exit pupil despite the same pixels being displayed (Fig. 19(a)). The parallelism was examined by measuring the displacement of the coordinates of one specific pixel of the virtual image. The image presented in Fig. 19(b) was displayed, and the coordinate positions of the center point of the image were measured (Fig. 19(c)). Some virtual images of 24 exit pupils could not be measured due to poor quality holograms, and only images that could be evaluated were measured for the center point position. A 1 pixel shift of a point corresponds to a 1.9 arcmin shift. The maximum horizontal shift was 15 pixels (28.5 arcmin) and the maximum vertical shift was 5.5 pixels (10.5 arcmin).
Fig. 19. Measurement of image position shift. (a) Image position shift. If beams displaying the same image were not aligned in parallel, the position of the observed image would appear to be shifted. (b) Displayed image. The center point position was measured. (c) Coordinates of the center point of the virtual image
7.2.6 ANSI contrast
To evaluate the contrast ratio of the virtual image, the pattern presented in Fig. 20(a) was displayed, and the ANSI contrast (ratio of the average luminance of white to the average luminance of black) was measured in a dark room. Since the display was corrupted in some locations and measurement was not possible, the four rectangles in the center (two white and two black) were evaluated (Fig. 20(b)). The displayed virtual image was captured by a camera, the value of G in the RGB of the image was quantified in 256 gradations, and the average value of the white and black area was calculated. Some virtual images of 24 exit pupils could not be measured due to poor exposure conditions. Only images that could be evaluated were measured for the ANSI contrast. The highest contrast ratio was 8.4:1, while the contrast ratio of the Nreal Light image was 7.6:1. Nreal contrast is low due to stray light from reflections of optics and plastic case (holder)(Fig. 20(c)).
Fig. 20. ANSI contrast evaluation (a) Pattern for ANSI contrast evaluation. (b) Evaluation area of ANSI contrast. the four rectangles in the center (two white and two black) were evaluated. (c) Displayed image on the Nreal Light. Slight reflections from plastic frames and optical components cause contrast deterioration.
8 Discussion
8.1 Angular selectivity
In some holograms, multiple object lights are reconstructed for a single reference light. In determining the angles of the reference lights for the reflection holograms, the angles of the reference lights were designed so that the diffraction efficiency due to the incidence of light at angles other than that of the reference lights would not be greater than 0.5 with the diffraction efficiency at the Bragg match set to 1. Therefore, it is possible that light with a diffraction efficiency of less than 0.5 was observed. In addition, the calculation in Subsection 3.4 did not take into account the effect of material shrinkage. Shrinkage likely caused diffraction by light at angles other than that of the reference light.
In this study, a material with a thickness of 16 µm and a shrinkage rate of 1.4% was used. However, a thicker material with less shrinkage would provide better angular selectivity. If the same material is used, the problem of angular selectivity can be solved by simulating the effects of shrinkage and setting the diffraction efficiency due to light at angles other than that of the reference lights much lower than 0.5.
8.2 Resolution
In the evaluation system used in the experiment, the horizontal resolution was degraded because the scanning display image was projected from an angle. Since the scanning display image beam was incident at 23° on the horizontal plane of the hologram, the beam diameter from the scanning display was expanded horizontally by a factor of 1/sin (23°) (1/0.39 ≈ 2.56), resulting in a calculated horizontal resolution of 1,280 pixels/2.56 = 500 pixels. In the experiment, a horizontal resolution of 480 pixels was observed when the projection system was focused, indicating that the major cause of resolution degradation was oblique projection of the scanning display image light (Fig. 21).
Fig. 21. Comparison of horizontal resolution of (a) 240-line pairs and (b) 250-line pairs. (a) Experimentally observed horizontal resolution (240 line pairs = 480 pixels). (b) Actual visibility of the theoretically obtained maximum horizontal resolution (250 line pairs = 500 pixels)
The results of this study demonstrate that there is little resolution degradation due to the holograms. However, since a high-resolution virtual image cannot be displayed without controlling the image light projected onto the hologram, it is necessary to consider projection optics that not only parallelize the direction of the beam of the scanning display, but also compensate for astigmatism and lateral expansion of the beam due to oblique projection.
8.3 Image position shift
Possible reasons for the shift of the virtual image position include inconsistencies in the glass thickness, material shrinkage, and alignment accuracy of the exposure system. In the exposure experiment, the thickness of the glass between the transmission and reflection holograms was set to 1.1 mm so that the image pitch was 1 mm, and the lens for the object light of the reflection hologram was moved in parallel at a pitch of 1 mm. However, if the thickness of the glass is not exactly 1.1 mm due to the manufacturing tolerance, the position at which the beams duplicated by the transmission hologram hit the reflection hologram will change. As a result, the direction of the beams diffracted by the reflection hologram will change, and the beams will not be parallel (Fig. 22). If the glass is 0.01 mm thicker (1.11 mm), the angle of the beam diffracted by the reflection hologram will change by 6 arcmin depending on the inclination of the beam coming from the transmission hologram. This problem can be solved by accurately measuring the thickness of the glass and changing the movement pitch of the lens for the object light according to that thickness.
Fig. 22. Image position shift due to the manufacturing tolerance. If the glass thickness differs from the design value, the position at which the beams duplicated by the transmission hologram hit the reflection hologram will change. Consequently, the direction of the beams diffracted by the reflection hologram will change, and the beams will not be parallel.
Regarding the image shift due to material shrinkage, shrinkage changes the diffraction angle of beams reconstructed by a transmission hologram. Therefore, the angles of incidence of the beams on the reflection holograms change, and the diffraction angles of the beams reconstructed by the reflection hologram also change. This problem can be solved by exposing the hologram at an angle that takes into account the shrinkage or by using a material with low shrinkage.
In this experiment, optical elements, such as mirrors, were adjusted manually. However, precise adjustment is required to expose at the angles presented in Table 1, and it is difficult to achieve precise adjustment by hand. In this experiment, we prioritized adjustment flexibility and used a jig with an adjustment mechanism for all optical elements. It is possible to construct a high-precision exposure system without manual adjustment by creating a jig that does not have an adjustment mechanism and can fix the optical elements at a specified angle.
8.4 Hologram material
It is desirable to use thicker materials with low shrinkage for NED optics. However, since thicker materials with low shrinkage and no variation in properties were not available, Covestro materials with small individual differences were used. In our experiment, it was difficult to obtain reproducible results even with Covestro mass-produced materials. It is necessary to clarify the parameters that must be controlled and the control accuracy to enable stable exposures.
8.5 Hologram exposure quality
As illustrated in Fig. 23(a), there were several images that were difficult to see among the images observed from the 24 exit pupils. In addition, some images, although visible, did not display on the entire screen, as illustrated in Fig. 23(b). The reasons are as follows. First, the exposure time was long. The seventh exposure took 18 s, while the eighth exposure took 25.2 s, which is a long time for a single hologram, resulting in interference fringes and low-quality holograms. One method is to increase the laser power and shorten the exposure time; however, in this case, the power available for a single hologram could not be increased due to the numerous light branches required for the 24 multiple exposures. We believe that using a laser with higher power will shorten the exposure time and improve the quality of interference fringes.
Fig. 23. Low-quality holograms. (a) Seventh exposed hologram with low exposure quality. (b) The entire image does not appear.
The second reason is that the exposure schedule was not optimal. The inability to expose reproducible holograms made it impossible to determine the optimal schedule time. The holograms for which sensitivity curves were calculated did not have the same characteristics as the holograms for the exposure experiment, and in some cases, even if eight multiple-exposure holograms were recorded according to the schedule calculated from the sensitivity curves, all eight holograms were not reconstructed. Because the same exposure schedule did not yield the same results, the experiment was continued with longer exposure times for eight holograms until all eight holograms were reconstructed. The temperature, air vibration, and exposure light intensity were controlled; however, we believe that the control was insufficient, resulting in the failure of highly reproducible exposures. The hologram exposure conditions for high reproducibility and the accuracy required to control the exposure conditions must be clarified in future work.
The third reason for the low image quality is shrinkage. The diffraction angle by the transmission hologram changed due to shrinkage, and not all of the image light hit the reflection holograms. In addition, the Bragg angle changed for both the transmission and reflection holograms, resulting in areas where sufficient diffraction efficiency could not be obtained. This problem can be solved by exposing the hologram at an angle that takes into account the shrinkage or by using a material with low shrinkage.
Finally, crosstalk and uneven brightness due to multiple exposures are also causes of low-quality images. As mentioned in Subsection 8.1, beams other than the virtual image beam to be reconstructed are also reconstructed for a single beam coming from the transmission hologram, resulting in noise, which can be alleviated using thicker materials.
In addition, the laser beam used for exposure was adjusted so that the in-plane intensity distribution was uniform; however, it was assumed that there was a slight distribution. Even if the slight difference in intensity was not problematic in a single exposure, the multiple exposures overlapped slight differences in intensity and increased the effect, leading to uneven brightness. The precise control of the in-plane intensity of the laser beam used for exposure is needed.
A single reflection hologram (reference light: parallel light object light: convergent light) can display images with higher contrast than eight multiplexed holograms (Fig. 24(a), (b)). It also has less-brightness irregularity(Fig. 24(c)).
Fig. 24. Quality of the single exposure hologram. (a) Image of a hologram with a single exposure (displaying 240 line pairs of images). (b) Image of a hologram with eight exposures (displaying 240 line pairs of images). The same image is displayed in (a) and (b), but (a) has higher contrast. (c) Image of a hologram with a single exposure (4 × 4 checkerboard).
Exposing high-quality multiple holograms is one major challenge. In addition, combining transmission and reflection holograms in this method makes it more difficult to achieve high-quality images. However, we believe that this can be resolved by developing thick materials with low shrinkage, controlling exposure parameters for reproducible holograms, optimizing the exposure schedule, and adjusting the in-plane intensity of the exposure laser.
8.6 Full colorization
Although only green color was verified in this study, full color is possible if the material is RGB compatible. However, because thin holograms have wide wavelength selectivity, the effect of crosstalk will increase. Therefore, a hologram material with high wavelength selectivity should be used so that the RGB wavelengths diffract independently. Additionally, the lenses used for exposure should not have any chromatic aberration. The exposure optical system used in this project is designed to minimize chromatic aberration; therefore, colorization is possible if hologram materials for RGB with high wavelength selectivity are used. The balance of RGB laser power must be adjusted during exposure.
7. Conclusion
In this study, we developed a compact NED that can display images with a horizontal FOV of 60° and that has a 6.2 mm × 4.8 mm eye box and 720 pixels vertical resolution using two holograms and a scanning display. However, the image quality observed in the experiments was low and further study is required. Future work includes the following tasks: (i) improving the lateral resolution by examining the optics to project the image onto the hologram, (ii) simulating the shrinkage of the hologram material, and (iii) examining the conditions for stable exposure of the multiple holograms.
Acknowledgments
The authors acknowledge ColorLink Japan, Ltd., HolyMine Corporation., Nippon Telegraph and Telephone Corporation and NTT Advanced Technology Corporation for supporting our work.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. T. Hatada, “3D display and human visual psychophysical effect,” Journal of the Imaging Society of Japan. 45(3), 262–269 (2006). (in Japanese).
2. Nreal: https://www.nreal.ai/
3. K. Kiyokawa, “A wide field-of-view head mounted projective display using hyperbolic half-silvered mirrors,” in 6th IEEE and ACM International Symposium on Mixed and Augmented Reality, 1–4 (2007).
4. J. P. Rolland, “Wide-angle, off-axis, see-through head-mounted display,” Opt. Eng. 39(7), 1760–1767 (2000). [CrossRef]
5. A. Sisodia, M. Bayer, P.T. Smith, B. Nash, J. Little, W. Cassarly, and A. Gupta, “Design of an advanced helmet mounted display (AHMD),” Proceedings of the SPIE 5801. Cockpit and Future Displays for Defense and Security, 304–315 (2005).
6. B.C. Kress and W.J. Cummings, “Optical architecture of HoloLens mixed reality headset,” Proceedings of the SPIE 10335, International Society for Optical Engineering, 103350K–103350K-10 (2017).
7. H. Mukawa, K. Akutsu, I. Matsumura, S. Nakano, T. Yoshida, M. Kuwahara, and K. Aiki, “A full-color eyewear display using planar waveguides with reflection volume holograms,” J. Soc. Inf. Disp. 17(3), 185–193 (2009). [CrossRef]
8. Y. Weng, Y. Zhang, J. Cui, A. Liu, Z. Shen, X. Li, and B. Wang, “Liquid-crystal-based polarization volume grating applied for full-color waveguide displays,” Opt. Lett. 43(23), 5773–5776 (2018). [CrossRef]
9. P. Saarikko, “Diffractive exit-pupil expander with a large field of view,” Proceedings of the SPIE. 7001. Photonics in Multimedia II, 700105 (2008).
10. C. Zhang, D. Cheng, J. Lin, and Y. Wang, “Retinal projection display system based on MEMS scanning projector and conicoid curved semi-reflective mirror,” Proceedings of the SPIE, AOPC 2020, 11565: Display Technology; Photonic MEMS, THz MEMS, and Metamaterials; and AI in Optics and Photonics. 1156506 (2020).
11. A. Maimone, A. Georgiou, and J. S. Kollin, “Holographic near-eye displays for virtual and augmented reality,” ACM Trans. Graph. 36(4), 1–16 (2017). [CrossRef]
12. J. Xiong, G. Tan, T. Zhan, and S. T. Wu, “Breaking the field-of-view limit in augmented reality with a scanning waveguide display,” OSA Continuum 3(10), 2730–2740 (2020). [CrossRef]
13. J. Yeom, J. Hong, and J. Jeong, “Projection-type see-through near-to-eye display with a passively enlarged eye box by combining a holographic lens and diffuser,” Opt. Express 29(22), 36005–36020 (2021). [CrossRef]
14. S. B. Kim and J. H. Park, “Optical see-through Maxwellian near-to-eye display with an enlarged eye box,” Opt. Lett. 43(4), 767–770 (2018). [CrossRef]
15. C. Chang, W. Cui, J. Park, and L. Gao, “Computational holographic Maxwellian near-eye display with an expanded eye box,” Sci. Rep. 9(1), 18749 (2019). [CrossRef]
16. J. Jeong, J. Lee, C. Yoo, S. Moon, B. Lee, and B. Lee, “Holographically customized optical combiner for eye box extended near-eye display,” Opt. Express 27(26), 38006–38018 (2019). [CrossRef]
17. B.C. Kress, “Digital optical elements and technologies (EDO19): applications to AR/VR/MR,” Proceedings of the SPIE 11062, Digital Optical Technologies. 2019, 1106222 (2019).
18. H. Kogelnik, “Coupled-wave theory for thick hologram gratings,” Bell Syst. Tech. J. 48(9), 2909–2947 (1969). [CrossRef]
19. https://solutions.covestro.com/en/products/bayfol/bayfol-hx200_86194384-20033146?SelectedCountry=JP
