Speckle noise is a serious problem for laser projectors because it deteriorates the image quality. The brightness of the image is one of the important parameters to determining the speckle perception. In this paper, the effect of brightness on speckle contrast is investigated using matte and silver screens for cinema. The brightness was changed by using ND filters or changing the on-and-off duty cycle of a DMD device in a digital cinema projector. It is shown that the brightness does not affect the speckle contrast measured by a CCD camera. Then, the effect of the brightness on human speckle perception is investigated. It is shown that the brightness of the image significantly affects human speckle perception. Speckle patterns created by the image with higher brightness are more noticeable than those with lower brightness when the speckle contrast measured by a camera is the same.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Lately, laser projectors have been rapidly developed along with the recent development of visible lasers. Laser projectors have various advantages over conventional lamp projectors, such as long lifetime, wide color gamut, and high brightness. However, as a result of high coherency of lasers, undesirable interference patterns appear on screen or on retina, which is called speckle patterns. Since speckle patterns deteriorate the image quality, it is important to suppress speckle to the acceptable level.
However, it is not easy to determine the acceptable level of speckle. The most widely used indicator to evaluate speckle is speckle contrast, but there is no agreed human perception threshold of speckle contrast yet. So far, there have been several different suggestions for human perception threshold [1–6]. Trisnadi et. al. mentioned that speckle image with the contrast of 8% is slightly perceivable for stationary image and perfectly acceptable for motion image . Manni et. al. suggested that speckle contrast of 5% is acceptable in some applications, but contrast of 1% is required for the most demanding case . Roelandt et. al. reported that stationary images with the speckle contrast less than 3-4% is acceptable , and Verschaffelt et. al. reported that motion image with the speckle contrast less than 5–7% is acceptable . One of the reasons why there are several different suggestions is that speckle contrast largely depends on its measurement system, such as a camera aperture size, screen-to-camera distance, pixel size, and measuring angle [7,8]. Therefore, the standardization of speckle measurement is important to measure speckle consistently [9–11].
Although speckle contrast is most widely used to evaluate speckle, it is not the only parameter that affects the human perception of speckle. For example, the random distribution of color on the xy chromaticity diagram due to speckle patterns of different primary colors, which is called color speckle, also affect the human perception of speckle [12,13].
The difference in primary color also has significant effect on the human speckle perception. It is shown that speckle of blue image is less noticeable than that of green or red images . One reason would be that the number of receptor for short wavelength in human eye (S cone) is less than those of medium and long wavelengths (M and L cones) . It was shown that the number of S cone is only about 5% of total cones for people with normal color vision. This would be one reason why blue speckle is less noticeable than green and red speckle.
Another factor that can affect the human perception of speckle is the brightness of the image. Even when the speckle contrast of two images are the same, human perception of speckle can be different when the image brightness is different. Lee et. al. reported that the minimum speckle contrast that people can perceive becomes lower when the image becomes brighter . They performed the sensory test with only two participants, so further study was required to develop a deeper investigation. In , the effect of the brightness on speckle contrast was also reported. It was shown that the speckle contrast decreased as the brightness increased. However, they did not mention how the brightness was changed. In , it was shown that speckle contrast decreases as the input current of lasers increases. Although the brightness increases when the input current increases, the change of the input current also affects conditions other than the brightness, such as the oscillation modes and spectra. Therefore, it has not been clear how the brightness affects speckle contrast.
In this paper, first, the effect of brightness on speckle contrast is investigated by changing the brightness. The brightness was changed by using ND filters or changing on-and-off duty cycle of a DMD device in a digital cinema projector. Unlike the method to change input current of lasers, the lasing conditions of diodes and spectra are kept the same in these methods. Thus, the effect of the brightness on speckle contrast can be directly investigated. Then, the effect of brightness on the human perception of speckle is investigated using the projector. The experiments were conducted using matte and silver screens, both of which are very popular screen for cinema. Matte screen shows Lambertian like scattering characteristics and has wide viewing angle. Silver screen has higher screen gain than matte screen, but it has narrower viewing angle. Since it preserves polarization, it is used for polarized 3D system. However, it does not have speckle reduction by polarization diversity, and typically shows higher speckle than matte screen.
2. Experimental method
Figure 1 shows a tabletop experimental setup to measure speckle. Laser light entered an optical fiber (0.39NA, core diameter 800 μm) and then the beam was homogenized in a hexagonal rod integrator (side length 2 mm, length 50 mm) and projected onto a silver screen (Finesilver240, Kikuchi Science Laboratory Inc.) or a matte screen (MattPlus, Harkness Screens Ltd.). A reflective ND filter with the optical density of 0.3 or 0.7 was inserted after the rod integrator to change the brightness.
Figure 2 shows an experimental setup using a DLP cinema projector. Light emitted from five lasers with the linewidth of about 0.1∼0.2 nm were bundled together using a bundle fiber (0.22NA, core diameter 400 μm) and entered the projector. The wavelengths of five lasers were chosen from Table 1. The wavelength interval of laser set 1 was about 1 nm, and that of laser set 2 was about 6 nm. In the DLP projector, a collimation lens, a first rod integrator, a ground glass diffuser, and a second rod integrator were placed in this order. The second rod integrator was slightly larger than the first one to avoid light loss between the two rods. The light emitting from the second rod illuminated a DMD device in the projector, and the image on the DMD device was projected onto a silver screen or a matte screen at the normal angle.
Speckle contrast was measured with the experimental setups shown in Figs. 1 and 2 with various brightness using a cooled CCD camera (BS-44DUV, Bitran Corporation). The speckle measurement condition is shown in Table 2. A circular aperture with the diameter of 1.0 mm was placed in front of the camera to simulate the human eye [16,17]. The speckle contrast was calculated by dividing the standard deviation of the intensity by the mean intensity. When the brightness of the image was changed, the exposure time was adjusted accordingly to obtain similar level of CCD signal. It was reported that the change of exposure time does not affect speckle contrast .
Then, the effect of brightness on human speckle perception was investigated using the experimental setup shown in Fig. 2.
Both speckle contrast measurement and human speckle perception was performed in a dark room. The luminance of the images was measured with a spectroradiometer (SR-3A, Topcon Technohouse Corporation).
3. Results and discussion
3.1 Effect of brightness on speckle contrast
First, speckle contrast was measured with the experimental system shown in Fig. 1. The brightness of the image was changed by up to a factor of 5 by using different ND filters. The exposure time of camera was changed as the brightness changed to keep the obtained CCD count roughly the same. Figure 3 shows the speckle contrast on two screens. It was shown that the speckle contrast did not change when the brightness was changed using ND filters.
Then, speckle contrast was measured with the experimental setup shown in Fig. 2. The laser set 1 in Table 1 was used and the image size was set to 840 mm × 1600 mm. The brightness was changed by changing the test patterns of the projector. The test patterns were chosen from two patterns shown in Fig. 4. Both test patterns have no spatial patterns and create full size rectangular images with uniform brightness. The difference between two patterns was brightness of the image. Test pattern 1 has the highest brightness and the brightness of test pattern 2 was about 20% of that of test pattern 1. Brightness of the two test patterns was changed by the on-and-off duty cycle of the DMD device. Figure 5 shows the measured speckle contrast with different on-and-off duty cycle of the DMD device. It was confirmed that there was almost no difference in speckle contrast when the brightness of the image was changed by adjusting the DMD duty cycle.
3.2 Effect of brightness on human speckle perception
Then, sensory tests to investigate human speckle perception were performed. First, the speckle contrasts of laser set 2 with the setup shown in Fig. 2 were measured for larger image (2400 mm × 4500 mm). This setup was the same as the setup used for the sensory tests. Table 3 shows the measured speckle contrast on the matte and silver screens. The speckle contrast of silver screen was about twice as large as that of matte screen because of the speckle reduction by wavelength diversity and polarization diversity. Wavelength interval needed to reduce speckle by wavelength diversity is smaller on matte screen than on silver screen, so the speckle reduction by using 5 wavelength lasers can be larger on matte screen [17,19]. Since the illumination light was depolarized in optical fibers in this setup, speckle reduction by polarization diversity can provide 50% speckle reduction in maximum on matte screen, whereas speckle is not reduced on silver screen because silver screen preserves polarization .
Two sensory tests of speckle were performed using laser set 2 to investigate the effect of brightness on human perception of speckle. The number of participants was 44. As reported by Goodman , those who have a farsighted or a nearsighted eye perceive speckle differently from those who have perfect vision or wear their vision correction. In our experiment, the participants who need glasses for daily life wore the glasses to correct their vision when they performed the speckle test. Other participant attributes of the sensory test are shown in Table 4.
The procedure of the first test was as follows:
- 1. The research participants stand as close to screen as possible and judge if they can perceive speckle.
- 2. They step back by 1 m and judge if they can perceive speckle again.
- 3. They repeat procedure 2 until they reach the distance from which they cannot perceive speckle, or until they step back by 12 m in total.
Then, the second sensory test was performed as follows:
- 1. The research participants stand 4 m apart from the screen.
- 2. They evaluate speckle on a four level scale: 1. very annoying, 2. annoying, 3. visible, but acceptable, 4. imperceptible.
Figure 7 shows the result of the second sensory test. On the matte screen, most participants answered that the speckle was imperceptible when the luminance was low, but about half participants answered that the speckle was visible or annoying when the luminance was high. On the silver screen, nearly half participants answered that speckle was acceptable when the luminance was low. When the luminance was high, most participants answered that speckle was very annoying or annoying.
These two results suggest that even for the image with one primary color (in this case, green), speckle contrast is not the only parameter that determines the human perception of speckle. It was clearly shown that brightness of the image has significant effect on the human perception of speckle. Speckle is easier to perceive when the brightness of the image is higher.
One reason why the speckle with higher brightness is easier to perceive is that the brightness affects the pupil size of the human eye. Since speckle was reduced by angular diversity in our experimental setup using optical fibers and a diffuser , speckle contrast measured by a camera becomes larger when the aperture is larger [10,17]. When the luminance of the image is higher, the eye pupil size is expected to become smaller . However, it was reported that when the pupil size is smaller, the eye aberration becomes smaller and the pinhole size that simulates human eye becomes larger . Therefore, unlike the speckle measured by a CCD camera, it is expected that speckle contrast on human retina becomes larger when the pupil size is smaller. It would be possible to compensate the change of speckle contrast on human retina due to the brightness change by changing aperture diameter of a camera accordingly. Further studies are required to determine the optimum aperture diameter according to the image brightness.
Another possible reason why speckle is more noticeable for brighter image is that the size of speckle depends on the pupil size. The speckle size on a CCD sensor is determined by the autocorrelation function of the speckle intensity on the sensor, which is dependent on the diameter of the aperture D, the distance between the sensor and the camera lens z, and wavelength λ . Defining the speckle size on the CCD sensor as the radius at which the normalized covariance function of the speckle intensity becomes 0.5, the speckle size is written as r = 0.51λz/D.
Figure 8 shows a calculated result of the speckle size as a function of camera aperture diameters and measured speckle sizes on the CCD camera’s pixels using a matte screen and a silver screen with the experimental setup similar to Fig. 1. As predicted by the theory, the speckle size was inversely proportional to the aperture diameter. Therefore, it is expected that the brightness of the image affects the speckle size on human retina due to the change of the pupil size, and the difference of the speckle size can be one of the reasons why speckle is more noticeable for brighter image. However, as far as we know, the speckle size on aberrated human eye model and how the speckle size affects the human speckle perception has not been investigated, so further studies are needed in order to reveal the effect of speckle size on the human eye.
In conclusion, the effect of brightness on speckle was investigated. It was confirmed that speckle contrast did not depend on the brightness when the brightness of the image was changed by using ND filters or by changing the on-and-off duty of the DMD device in a projector.
The relation between the human perception of speckle and the brightness of the image was also investigated. It was revealed that the brightness of the image significantly affects the human perception of speckle by performing sensory tests. When the speckle contrast measured by a CCD camera was the same, speckle of brighter image was more noticeable than that of less bright image. A possible reason why speckle is more noticeable for brighter image is that the pupil size changes depending on brightness. The pupil size becomes smaller for the brighter image, and it is reported that the resolution of human eye is higher when the pupil size is smaller because of the eye lens aberration. Thus, it is expected that the speckle contrast on human eye becomes higher for higher brightness image.
1. J. I. Trisnadi, “Speckle contrast reduction in laser projection displays,” Proc. SPIE 4657, 131–137 (2002). [CrossRef]
2. M. S. Brennesholtz and E. H. Stupp, Projection Displays (Wiley, 2008).
3. Y. M. Lee, D. U. Lee, J. M. Park, S. Y. Park, and S. G. Lee, “A Study on the Relationships between Human Perception and the Physical Phenomenon of Speckle,” SID Symp. Dig. Tech. Pap. 39(1), 1347 (2008). [CrossRef]
4. J. G. Manni and J. W. Goodman, “Versatile method for achieving 1% speckle contrast in large-venue laser projection displays using a stationary multimode optical fiber,” Opt. Express 20(10), 11288–11315 (2012). [CrossRef]
5. S. Roelandt, Y. Meuret, A. Jacobs, K. Willaert, P. Janssens, H. Thienpont, and G. Verschaffelt, “Human speckle perception threshold for still images from a laser projection system,” Opt. Express 22(20), 23965 (2014). [CrossRef]
6. G. Verschaffelt, S. Roelandt, Y. Meuret, W. Van Den Broeck, K. Kilpi, B. Lievens, A. Jacobs, P. Janssens, and H. Thienpont, “Speckle disturbance limit in laser-based cinema projection systems,” Sci. Rep. 5(1), 14105 (2015). [CrossRef]
7. J. W. Goodman, Speckle Phenomena in Optics (Roberts and Company Publishers, 2007).
8. H. Yamada, K. Moriyasu, H. Sato, and H. Hatanaka, “Effect of incidence/observation angles and angular diversity on speckle reduction by wavelength diversity in laser projection systems,” Opt. Express 25(25), 32132–32141 (2017). [CrossRef]
9. T. Fukui, K. Ito, K. Suzuki, H. Tokita, Y. Furukawa, and S. Kubota, “Effective Calibration Method for Absolute Speckle Contrast Measurement,” Proc. LDC LDC8-3 (2012).
10. K. Suzuki, T. Fukui, S. Kubota, and Y. Furukawa, “Verification of speckle contrast measurement interrelation with observation distance,” Opt. Rev. 21(1), 94–97 (2014). [CrossRef]
11. S. Roelandt, Y. Meuret, G. Craggs, G. Verschaffelt, P. Janssens, and H. Thienpont, “Standardized speckle measurement method matched to human speckle perception in laser projection systems,” Opt. Express 20(8), 8770–8783 (2012). [CrossRef]
12. K. Kuroda, T. Ishikawa, M. Ayama, and S. Kubota, “Color Speckle,” Opt. Rev. 21(1), 83–89 (2014). [CrossRef]
13. J. Kinoshita, K. Yamamoto, and K. Kuroda, “Color speckle measurement errors using system with XYZ filters,” Opt. Rev. 25(1), 123–130 (2018). [CrossRef]
14. A. Roorda and D. R. Williams, “The arrangement of the three cone classes in the living human eye,” Nature 397(6719), 520–522 (1999). [CrossRef]
15. T.-T.-K. Tran, Ø. Svensen, X. Chen, and M. Nadeem Akram, “Speckle reduction in laser projection displays through angle and wavelength diversity,” Appl. Opt. 55(6), 1267–1274 (2016). [CrossRef]
16. S. Kubota, “Simulating the human eye in measurements of speckle from laser-based projection displays,” Appl. Opt. 53(17), 3814–3820 (2014). [CrossRef]
17. H. Yamada, K. Moriyasu, H. Sato, and H. Hatanaka, “Dependency of speckle reduction by wavelength diversity on angular diversity in laser projection system,” J. Soc. Inf. Disp. 26(4), 237–245 (2018). [CrossRef]
18. K. Suzuki and S. Kubota, “Understanding the exposure-time effect on speckle contrast measurements for laser displays,” Opt. Rev. 25(1), 131–139 (2018). [CrossRef]
19. H. Yamada, Ushio inc., 1194 Sazuchi, Bessho-cho, Himeji, Hyogo 675–0224, Japan, K. Moriyasu, H. Sato, H. Hatanaka, and Y. Kazuhisa are preparing a manuscript to be called “Theoretical calculation and experimental investigation of speckle reduction by multiple wavelength lasers in laser projector with different angular diversities.”
20. H. Yamada, Ushio inc., 1194 Sazuchi, Bessho-cho, Himeji, Hyogo 675–0224, Japan, K. Moriyasu, H. Sato, H. Hatanaka, and Y. Kazuhisa are preparing a manuscript to be called “Speckle reduction in laser projectors by angular, wavelength, and polarization diversity.”
21. J. Pokorny and V. C. Smith, “How much light reaches the retina?” in Colour Vision Deficiencies XIII. Documenta Ophthalmologica Proceedings Series, Vol 59, C. R. Cavonius, ed. (Springer, Dordrecht, 1997).