Open Access
Add to BibSonomy
Abstract
In this paper, we propose a time-division multiplexing parallax barrier system with reduced power consumption by using a lenticular lens to generate active barrier slits. In the conventional study, the ratio of luminance to power consumption is low because of the usage of two LCD panels with limited transmittance. In the proposed system, a fine slit pattern is generated by projecting LED striped light through a lenticular lens, where active change of slit pattern is realized by changing the positions of LED emission. We measure and confirm reduction of power consumption in the proposed method. We also compare the effects of time division on crosstalk in the conventional and the proposed methods.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Parallax barrier is one of the most well-known methods to realize autostereoscopy. However, the conventional parallax barrier systems have limited viewing zones and low spatial image resolutions because of its static barrier structure.
One of the solutions to maintain high image resolution is to use a directional backlight composed of a light guide film and a pair of light sources [1–3], where autostereoscopy is realized by time-division of images on the display panel and the directionality of backlight. The drawback of this method is fixed viewing zone due to static directionality of backlight.
To enlarge the viewing zone, the directionality of backlight has to be controlled to follow the motion of a viewer, which requires thick optical systems [4–15]. Some of them use a coarse lenticular lens to realize directionality of backlight [8–13]. To realize a thinner optical system by using a finer lenticular lens, a dot matrix light source with extremely fine pixel pitch is needed, which is not practical.
An easier and more compact way to realize full resolution autostereoscopy is time-division multiplexing parallax barrier [16,17]. The pixels of a stereo pair are divided into two frames by resolution, where one frame shows half of each view and the other frame shows the other half by shifting the phase of a barrier pattern and an image pattern. Since this system requires a pair of LCD (Liquid Crystal Display) panels layered with a short interval, the system can be thin and compact.
In addition, head-tracking technology widens the viewing zone [18–21]. By monitoring the position of the viewer, the image or the barrier pattern is adjusted accordingly to move the viewing zone so that it may always follow the position of the viewer to maintain correct stereoscopy.
To ensure a wider viewing zone, Zhang et al. have proposed time-division quadruplexing parallax barrier [22–25] as shown in Fig. 1, where the same image is delivered to two of the four viewpoints, which avoids emergence of crosstalk when each of the viewer’s eyes is positioned between the two viewpoints showing the same image.
Fig. 1. Time-division quadruplexing parallax barrier. A left-eye image is shown at pixels A and B and a right-eye image is shown at pixels C and D.
When the barrier is slanted, each slit can be shifted by subpixel or sub-subpixel unit [26], which enables fine tuning of barrier patterns and reduces crosstalk when the barrier pattern is properly controlled to follow the motion of the viewer. The theory on the viewing zone without crosstalk has already been established [27,28].
Another drawback of parallax barrier is reduction of luminance. When the aperture ratio is 1/4, the luminance becomes 1/4 of the original display. When time-division multiplexing is applied, the image becomes all the darker because of the usage of two LCD panels with limited transmittance. Therefore, higher power consumption is needed to maintain the luminance of image. One way to solve this problem is to use self-luminous fine slits with high intensity, which is hard to realize technically, however.
In this paper, we propose a method to realize time-division multiplexing parallax barrier with higher luminance and low power consumption by use of a lenticular lens to generate a barrier pattern.
This paper is organized as follows. Principle of the proposed method is explained in Section 2. Section 3 describes a prototype system applying the proposed method. Experiments performed with the prototype system and the results are described in Section 4. The paper is concluded in Section 5.
2. Principle
To realize a barrier pattern image composed of fine slits, we use slanted LED bars arranged side by side as shown in Fig. 2 for the backlight. An image of fine slit pattern is generated by projecting LED striped lights through a lenticular lens, as shown in Fig. 3. Note that the distance between the backlight and the lens is much larger than that between the lens and the real image of parallax barrier in reality. Due to the similarity of triangles, fine slits can be generated from a wide range of LED lights. Here one every four LED bars are luminous, while the luminous bar is alternated for time-division multiplexing. We can reproduce time-division quadruplexing parallax barrier image by switching the luminous part of the LED light bars at a high speed as shown in Fig. 4.
Fig. 2. Arrangement of inclined LED bar light. LED bars are placed close to each other, while one every four LED bars are luminous at a time.
Let f be the focal length of the lenticular lens, D be the distance between the backlight and the lenticular lens, and d be the distance between the lenticular lens and the real image of parallax barrier. Then
holds.Let W be the width of LED light bars and w be the slit width in the parallax barrier image. Then
holds because of the similarity of dotted triangles in Fig. 5.Let Wl be the lens pitch of the lenticular lens. From the similarity of thick-lined triangles in Fig. 5,
is derived. If these equations are satisfied, shift of slit positions generated by different light bars shown in Fig. 6 does not occur.Fig. 6. Misalignment of slit image. When Eqn. (3) does not hold, the barrier image from different light bars do not converge to the same position as the circled part in the figure shows.
When the periodical cycle of slits is close to that of interleaved image, the viewer can see a stereoscopic image without crosstalk when he or she is far from the screen. To maintain stereoscopy when the viewer comes closer to the screen, the interval of slits should be widened, or the widths of the interleaved pixel units should be narrowed to preserve the similarity of triangles whose vertex with the smallest angle is located at the eye position. When this relation does not hold, the images for different viewpoints are mixed up, which causes crosstalk as shown in Fig. 7 (top). Though the pixel width cannot be changed adaptively to meet this requirement, we can maintain the similarity of triangles in a discrete manner by removing a subpixel for a corresponding viewpoint periodically as shown in Fig. 7 (bottom). By adjusting the period of subpixel removal depending on the distance between the display and the observer, proper stereoscopy is maintained even when the viewer moves in the depth direction.
Fig. 7. Crosstalk caused by displacement of pixel pitch (top) and correction of image displacement (bottom).
3. Prototype
We have made a prototype system based on the proposed method. The optical design of the prototype system is shown in Fig. 8. We used a 24-inch full HD LCD panel. The pixel pitch of this panel was 0.276 mm. The interval between neighboring LED light bars was 17 mm and the width of the LED light was 8 mm. Ideally, the light bars should be placed without interval as shown in Fig. 2. Here a narrower LED bars were used to reduce crosstalk caused by low precision in alignment process.
The distance between the front panel and the lenticular lens was 6.5 mm, and the distance between the backlight bars and the lenticular lens was 100 mm. The lens pitch was 0.7 mm, while the radius of each lenslet was 0.5 mm. In this research, the lenticular lens tilted by ${\tan ^{ - 1}}\frac{1}{3}$ was used as shown Fig. 9. Therefore, ${W_l}$ is calculated to be $0.7 \times \frac{{\sqrt {{3^2} + {1^2}} }}{3}$ = 0.738 [mm]. The picture of the prototype system is shown in Fig. 10.
We used an AMD FirePro W8100 for the graphic board. Time-division lighting of LED bars was controlled by the signals from the mini-DIN jack of the graphic board, which is usually used to control 3D shutter glasses.
4. Experiments and results
4.1 Comparison of the ratio of luminance to power consumption
We measured power consumption of the backlight and luminance in the center of the display when displaying a white image for the right eye and a black image for the left eye using both the conventional system (using two LCD panels) and the proposed prototype system (using one LCD panel and a LED based slits). All the LCD panels used in both systems were 24 inch TN color panels (AUO M240HW01 V8). The luminance was measured with Topcon BM-7AC. We calculated the ratio of luminance to power consumption and compared them.
The results are shown in Table 1. The proposed method is about 23 times more energy-efficient than the conventional method. Though the system using two LCD panels can be three times more energy efficient when a monochrome panel is used for generating active barrier patterns, the proposed system still remains around eight times more efficient, which shows its superiority.
Table 1. Ratio of luminance to power consumption
4.2 Effects of time division on crosstalk
Next, we measured stereoscopic crosstalk levels in the conventional and the proposed methods. By shifting the position of slits, the luminance for different viewpoints was measured while the position of the luminance meter was fixed. Switching the luminous part of the LED backlight might increase crosstalk because it was not fully synchronized with image update on the LCD panel. To take this effect into account, we measured luminance under static (time-division disabled) and active (time-division enabled) barrier conditions with the conventional and the proposed systems.
The results using the conventional system are shown in Fig. 11 and those using the proposed system are shown in Fig. 12. View 1 in the figure means the luminance when black images were displayed for both eyes, which was used to measure the ambient luminance. View 2 means the luminance when a white image was displayed for the left eye and a black image was displayed for the right eye. View 3 means the luminance when a black image was displayed for the left eye and a white image was displayed for the right eye. Let Lv2 and Lv3 be the luminance at the left-eye position under View 2 and View 3 conditions, ${R_{v2}}$ and ${R_{v3}}$ be the luminance at the right-eye position under View 2 and View 3, and ${B_l}$ and ${B_r}$ be the ambient luminance at the left-eye and the right-eye positions. Then, the crosstalk level is defined by
Fig. 11. Change of luminance when time division is enabled (left) and is disabled (right) in the conventional system.
Fig. 12. Change of luminance when time division is enabled (left) and disabled (right) in the proposed system.
Table 2. Crosstalk levels under 4 conditions
In the conventional method, time-division multiplexing does not affect the crosstalk level. However, in the proposed method, displaying in time-division increases the crosstalk level. The crosstalk level is higher even when the barrier is static and the curve of luminance change is unstable in the proposed method. Since the LED bars are aligned by hand, the barrier pattern is not as regularly aligned as the conventional method that uses two LCD panels with the precise pixel pitches.
4.3 Observed stereoscopic images
To confirm the stereoscopic effect with the prototype system, we showed Tsukuba stereo pair, one of the well-known standard stereo pair images. Figure 13 shows a left-eye image and a right eye image when time-division is enabled in the proposed system, while Fig. 14 shows a pair of observed stereoscopic images when time-division is disabled.
Fig. 13. Tsukuba stereo pair observed at the left-eye viewpoint (left) and observed at the right-eye viewpoint when time-division is enabled in the proposed method.
Fig. 14. Tsukuba stereo pair observed at the left-eye viewpoint (left) and observed at the right-eye viewpoint when time-division is disabled in the proposed method.
Though crosstalk is more perceivable when time-division is enabled, it meets the level of practical use in both cases. The slanted noise comes from the rough alignment of the LED bars by hand.
4.4 For Further Improvement
The luminance can be improved when we use more powerful LED light bars. Figure 15 shows the observed image when a 24 W light source composed of wider light bars was used. Here the width of elemental light bars was as large as 18 mm. The luminance increased to 61.8 cd/${\textrm{m}^2}$ and the striped noise was reduced because the gap of elemental light bars decreased, while the crosstalk became more apparent because of the narrower gap. When the light bars become wider and the gap becomes narrower, more precise placement is required to suppress crosstalk.
Fig. 15. Tsukuba stereo pair observed at the left-eye viewpoint (left) and observed at the right-eye viewpoint (right) when high luminance LED light bars are used.
Thickness of the optical system depends on the width of light bars and the focal distance of the lenticular lens. The system can be thinner if the light bars are narrower, though the alignment of light bars becomes even severer to suppress crosstalk. When we can apply precise production lines in place of hand-making to compose the optical system, a thinner and brighter display system with little crosstalk is expected to be realized based on the proposed method.
5. Conclusion
In this paper, we have proposed a time-division multiplexing parallax barrier system composed of an LCD panel and an active slit pattern using LED bars and a lenticular lens. We have made a prototype system and have confirmed that autostereoscopy with acceptable crosstalk level is realized. We have measured the ratio of power consumption to luminance, which is significantly improved with the proposed method. Further study is needed to decrease the crosstalk level and to improve image quality.
Funding
Core Research for Evolutional Science and Technology (JPMJCR18A2); Japan Society for the Promotion of Science (17H00750).
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. J.C. Schultz, R. Brott, M. Sykora, W. Bryan, T. Fukamib, K. Nakao, and A. Takimoto, “Full resolution autostereoscopic 3D display for mobile applications,” SID 2009 Digest, pp. 127–130 (2009)
2. A. Travis, N. Emerton, T. Large, S. Bathiche, and B. Rihn, “Backlight for ViewSequential Autostereo 3D,” SID 2010 Digest, pp. 215–217 (2010)
3. M.J. Sykora, “Optical characterization of autostereoscopic 3D displays,” Proc. SPIE. 7863, 78630V (2011). [CrossRef]
4. T. Hattori, T. Ishigaki, K. Shimamoto, A. Sawaki, T. Ishiguchi, and H. Kobayashi, “Advanced autostereoscopic display for G-7 pilot project,” Proc. SPIE 3639, 66–75 (1999). [CrossRef]
5. A. Hayashi, T. Kometani, A. Sakai, and H. Ito, “A 23-in. full-panel-resolution autostereoscopic LCD with a novel directional backlight system,” J. Soc. Inf. Disp. 18(7), 507–512 (2010). [CrossRef]
6. P. Surman, I. Sexton, K. Hopf, W.K. Lee, F. Neumann, E. Buckley, G. Jones, A. Corbett, R. Bates, and S. Talukdar, “Laser-based multiuser 3-D display,” J. Soc. Inf. Disp. 16(7), 743–753 (2008). [CrossRef]
7. D. Miyazaki, Y. Hashimoto, T. Toyota, K. Okoda, T. Okuyama, T. Ohtsuki, A. Nishimura, and H. Yoshida, “Multi-user autostereoscopic display based on direction-controlled illumination using a slanted cylindrical lens array,” in Stereoscopic Displays and Applications XXV, Proc. SPIE 9011, 90111G (2014).
8. S. Ishizuka and H. Kakeya, “Flat panel autostereoscopic display with wide viewing zone using time-division multiplexing backlight,” SID Digest of Technical Papers 44(1), pp. 1173–1176 (2013).
9. S. Ishizuka, T. Mukai, and H. Kakeya, “Viewing zone of an autostereoscopic display with a directional backlight using a convex lens array,” Electron. Imaging 23(1), 011002 (2014). [CrossRef]
10. T. Mukai and H. Kakeya, “Enhancement of viewing angle with homogenized brightness for autostereoscopic display with lens-based directional backlight,” Proc. SPIE 9391, 93911A (2015). [CrossRef]
11. S. Ishizuka, T. Mukai, and H. Kakeya, “Multi-phase convex lens array for directional backlights to improve luminance distribution of autostereoscopic display,” IEICE Trans. Electron. E98.C(11), 1023–1027 (2015). [CrossRef]
12. G. Borjigin and H. Kakeya, “An autostereoscopic display with time-multiplexed directional backlight using a decentered lens array,” Proc. Digital Holography and 3D Imaging 2019, W2A.2 (2019).
13. G. Borjigin and H. Kakeya, “Autostereoscopic displays with time-multiplexed directional backlight using curved lens arrays,” ITE Trans. on MTA 9(1), 80–85 (2021). [CrossRef]
14. G. Borjigin and H. Kakeya, “Autostereoscopic display for multi-viewers positioned at different distances using time-multiplexed layered directional backlight,” Appl. Opt. 60(12), 3353–3357 (2021). [CrossRef]
15. G. Borjigin and H. Kakeya, “A backlight system using a novel interleaved Fresnel lens array that attains a uniform luminance and two-dimensional directional light control,” Opt. Lett. 47(2), 301–304 (2022). [CrossRef]
16. K. Perlin, S. Paxia, and J. S. Kollin, “An autostereoscopic display,” Proceedings of the 27th Annual Conference on Computer Graphics and Interactive Techniques, pp. 319–326 (2000).
17. H. J. Lee, H. Nam, J. D. Lee, H. W. Jang, M. S. Song, B. S. Kim, J. S. Gu, C. Y. Park, and K. H. Choi, “A high resolution autostereoscopic display employing a time division parallax barrier,” SID 2006 Digest, pp. 81–84 (2006).
18. G. J. Woodgate, D. Ezra, J. Harrold, N. S. Holliman, G. R. Jones, and R. R. Moseley, “Observer-tracking autostereoscopic 3D display systems,” Proc. SPIE 3012, 187 (1997). [CrossRef]
19. N. A. Dodgson, “On the number of viewing zones required for head-tracked autostereoscopic display,” Proc. SPIE 6055, 60550Q (2006). [CrossRef]
20. S.-Y. Yi, H.-B. Chae, and S.-H. Lee, “Moving parallax barrier design for eye-tracking auto-stereoscopic displays,” Proceedings of 3DTV Conference 2008, pp. 165–168 (2008).
21. J.-E. Gaudreau, “Full-resolution autostereoscopic display with all-electronic tracking system,” Proc. SPIE 8288, 82881Z (2012). [CrossRef]
22. Q. Zhang and H. Kakeya, “An autostereoscopic display system with four viewpoints in full resolution using active anaglyph parallax barrier,” Proc. SPIE 8648, 864811 (2013). [CrossRef]
23. Q. Zhang and H. Kakeya, “Time-division multiplexing parallax barrier based on primary colors,” Proc. SPIE 9011, 90111F (2014).
24. Q. Zhang and H. Kakeya, “A high quality autostereoscopy system based on time-division quadplexing parallax barrier,” IEICE Trans. Electron. E97.C(11), 1074–1080 (2014). [CrossRef]
26. A. Hayashishita and H. Kakeya, “Time-division multiplexing parallax barrier with sub-subpixel phase shift,” SID Digest of Technical Papers 49, pp. 1515–1518 (2018).
25. Q. Zhang and H. Kakeya, “Time-division quadruplexing parallax barrier employing RGB slits,” J. Disp. Technol. 12(6), 626–631 (2016). [CrossRef]
27. H. Kakeya, K. Okada, and H. Takahashi, “Time-division quadruplexing parallax barrier with subpixel-based slit control,” ITE Trans. on MTA 6(3), 237–246 (2018). [CrossRef]
28. H. Kakeya, A. Hayashishita, and M. Ominami, “Autostereoscopic display based on time-multiplexed parallax barrier with adaptive time-division,” Jnl Soc Info Display 26(10), 595–601 (2018). [CrossRef]
