The world on your desktop

27 Apr 2010

Quicker computation of a hologram for 3D scenes makes real-time operation viable on a personal computer

Fig. 1: Holographic reconstruction of a 3D model of a world globe from holograms generated using the new computation algorithm.

Fig. 1: Holographic reconstruction of a 3D model of a world globe from holograms generated using the new computation algorithm.

© 2010 Y. Pan

Three dimensional (3D) displays have become increasingly popular recently for cinemas and domestic televisions. Most existing commercial systems, however, are based on stereoscopy, meaning they have limited depth perception. For true 3D displays, however, holography is required, although its use has been hindered by slow computational algorithms and the need for high-powered computers. Now, a method that enables the fast computation of holograms with a fully realistic depth perception—even on a desktop computer—has been developed by researchers from the Data Storage Institute of A*STAR, Singapore.

“Computer holography has the ability to present all depth cues of 3D scenes and could find a wide range of applications from entertainment and simulators to medical and scientific visualization,” says Yuechao Pan from the research team led by Xuewu Xu.

Computational generation of holograms is a time-consuming task as it requires simulated propagation of a large number of light beams. To perform this task efficiently, computer algorithms use ‘look-up tables’ of pre-computed, frequently required computational values. During the actual generation of a hologram, the pre-computed results are retrieved for faster processing.

The algorithm developed by Pan and his co-workers, simplifies the calculations based on one, large 3D look-up table. They designed the algorithm to split this table into two separate 2D look-up tables with two planes in space, one in the xz direction and one in the yz direction. After separately completing the computations for each plane, the values of the xz plane were convoluted with those of the yz plane.

Compared with previous algorithms, this process reduces the computational complexity of the problem by an order, making it sufficiently efficient for a desktop computer to generate holograms (Fig. 1). Rather than the main computer processor, the computations are performed in parallel using the processors of three graphics cards, which are particularly well suited to the parallel computations needed for the main part of the team’s holographic reconstruction algorithm.

The computation time of a hologram using the team’s new algorithm now takes less than half a second; previously it took approximately 50 seconds. This puts the computation of holograms very close to real-time operation. Indeed, Pan is confident that, with a further optimization of the algorithm as well as the use of more powerful graphics cards, “real-time and interactive 3D computer-generated holography will be realized in the near future.”

This research was funded by the A* STAR HOME2015 Programme. The A*STAR-affiliated authors in this highlight are from the Data Storage Institute.

Want to stay up-to-date with A*STAR’s breakthroughs? Follow us on Twitter and LinkedIn!


Pan, Y., Xu, X., Solanki, S., Liang, X.,Tanjung, R.B.A., Tan, C. Chong T.-C.  Fast CGH computation using S-LUT on GPU. Optics Express 17, 18543 (2009). | article

This article was made for A*STAR Research by Nature Research Custom Media, part of Springer Nature