By Gregg Favalora, Optics for Hire
This year’s Stereoscopic Displays and Applications (SD&A) conference – the premier conference for 3D innovation – took place in San Francisco, USA, 23-25 January 2012. This, its 23rd meeting, was home to an array of innovations spanning stereoscopic and multi-view capture, 3D display, computation, and perceptual topics, and hosted 53 oral presentations and 43 posters, with a series of intriguing keynotes, a 3D Theatre session, and author demonstrations.
The conference is the largest part of a symposium of great depth: Electronic Imaging – a venue for the dissemination of research across image analysis, display, processing, and perception. SD&A is anchored by world-class plenary and keynote speakers, on topics relevant to those in the stereoscopic industry.
Prof. William Freeman (MIT) surveyed the state of computational photography, the merger of computer science and optics that re-examines what cameras can do. For example, Prof. Ramesh Raskar’s flutter shutter freezes object motion with great clarity and brightness using a particular aperiodic sequence of aperture states (Raskar et al, 2006), whereas camera shake can be mitigated by recovering a blur kernel in work from MIT CSAIL and University of Toronto (Fergus et al, 2006). Other examples included plenoptic cameras that permit post-processed refocusing, such as the Lytro light field camera (Ng et al, 2005), and an algorithm that magnifies motion (Liu et al, 2005).
Panasonic’s Masayuki Kozuka’s keynote address summarised the firm’s activity in developing stereoscopic cameras and televisions, including ways in which they ‘pivoted’ to accommodate shifts in technology or the target market. Also, Pete Bradshaw and Debargha Mukherjee from the Google/YouTube group gave a very candid look at the development of YouTube’s 3D capabilities. One delightful titbit was that Pete turned what began as a side-project into a widely-used feature of YouTube by interjecting sly remarks from a commenter posing as Wheatstone.
This year’s SD&A was an embarrassment of riches. Let’s look at just a few of the advancements presented by the speakers:
There continue to be innovations in 3D image capture, from Cornell University’s ‘Angle-sensitive pixels’ (Wang et al, 2012) to the ‘Polarizing aperture stereoscopic cinema camera’ of industry legend Lenny Lipton (2012). Emblematic of industry support, there were numerous papers from companies such as H-P Labs, Technicolor, Qualcomm, and Texas Instruments. For example, TI spoke on ‘Auto convergence for stereoscopic 3D mobile cameras,’ (Kothandaraman and Batur, 2012) as well as real-time misalignment correction techniques (Pekkucuksen et al, 2012).
Relevant particularly to autostereoscopic displays, such as multi-view lenticular or barrier systems, was a paper reminding us about the distortions that occur when imagery rendered for one location or region is viewed from a different location (Said and Culbertson, 2012).
Stereo 3D’s reach into daily life was discussed and demonstrated by Sony Computer Entertainment Europe’s Ian Bickerstaff. Ian’s presentation was given entirely in stereoscopic 3D, and was about methods in which the developers of PlayStation video games can use zoom and depth scaling to improve the appearance of stereo 3D in video games. The talk was awarded ‘Best Use of Stereoscopic Projection’ in a presentation at the conference.
Significantly, advances in no-glasses 3D stretched across more than three sessions. Two groups, such as the MIT Media Lab’s Camera Culture Group, are applying the maths of optimisation and tomography to stacked barrier displays, enabling them to reconstruct imagery that floats outside the display’s physical boundaries (Lanman, 2012). Also notable is ongoing development of 360-degree ‘table top’ displays (Uchida and Takaki, 2012) which direct several projectors at a spinning horizontal disc.
Safety and comfort
As millions experience your stereoscopic content in theatres and handheld devices such as the Nintendo 3DS, safety and comfort are getting renewed scrutiny. SD&A joined forces with the Human Vision and Electronic Imaging XVII conference to present a double joint session examining 3D image quality – particularly Technicolor’s work in visual fatigue (Vienne, Blondé, and Doyen, 2012), the discomfort of stereo viewing while tilting one’s head (Kane, Held, and Banks, 2012), and others.
3D Theatre awards
An exceedingly popular aspect of the conference is the annual 3D Theatre Session, chaired by conference chair Andrew Woods (Curtin Univ.) and committee member Chris Ward (Lightspeed Design). Nearly 400 attendees watched clips from 41 stereo films on an 18-foot 3D projection screen. Judges chose the following winners: Best of Show (‘All Is Not Lost’ by OK Go and Pilobolus) and Honorary Mention (‘The Blackpool Tower 4D Experience’ by Sharp Cookies Ltd).
SD&A’s demonstration session has transitioned into a symposium-wide event; a collection of software and hardware techniques, as discussed by multiple authors, was on show. For example, Prof. Hideki Kakeya (Univ. of Tsukuba) showed the latest integral/volumetric display, Christopher Tyler used a Kinect depth camera to plot random-dot autostereograms, Raytrix showed its lightfield camera, and Ronald Karpf demonstrated active Pulfrich-effect glasses. These demonstrations were also judged, and the winner was the University of Tokushima’s work in periodic metal structures that create ‘floating’ images of LED signs placed behind it (Yamamoto et al, 2012).
The interface between technology and market adoption was the theme of a riveting panel discussion on ‘Will 3D moving images become ubiquitous?’ Lenny Lipton moderated panellists Bob Whitehill (Pixar Animation Studios), Ray Zone (The 3D Zone), Art Berman (Insight Media), and Jon Peddie (Jon Peddie Research). They expressed a shared belief that 3D is accepted by the public, certainly in movies, helped by the quality of particular 3D films such as Avatar and Hugo. Paraphrasing, it was concluded that 3D should be used as a subtle storytelling and mood-conveying device, rather than a gimmick. In terms of technology, Jon Peddie predicted that by 2015, 80% of mobile phones will be autostereoscopic – that year, we will wake up and ‘like a boiling frog’, 3D will be everywhere.
It is difficult to identify a core theme tying together progress in each branch of 3D research, but ‘Rapid innovation amidst widespread adoption’ seems fitting. This year’s SD&A had the hallmarks of a busy and productive twelve months. 3D movies are generating strong revenue, researchers are pushing ahead stereo and autostereo display, and vision scientists and computer software engineers are ensuring that it is safe and useful.
After generations of fits and starts, stereoscopic (and autostereoscopic) 3D has crossed the chasm into mainstream enjoyment and the generation of profit for the companies within its ecosystem.
Note: citations having ‘SDA12 [CID number]‘ are in Stereoscopic Displays and Applications XXIII, edited by Andrew J. Woods, Nicolas S. Holliman, Gregg E. Favalora, Proceedings of SPIE-IS&T Electronic Imaging, SPIE Vol. 8288 (2012).
R. Raskar, A. Agrawal, and J. Tumblin, ‘Coded exposure photography: motion deblurring using fluttered shutter,’ in ACM SIGGRAPH 2006, 25(3), (July 2006).
R. Fergus, B. Singh, A. Hertzmann, S. T. Roweis, and W. T. Freeman, ‘Removing camera shake from a single photograph,’ in ACM SIGGRAPH 2006, 25(3), (July 2006).
R. Ng, ‘Fourier slice photography,’ in ACM SIGGRAPH 2005, 24(3), (July 2005).
C. Liu, A. Torralba, W. T. Freeman, F. Durand, and E. H. Adelson, ‘Motion magnification,’ in ACM SIGGRAPH 2005, 24(3), (July 2005).
A. Wang, S. S. Hemami, and A. Molnar, ‘Angle-sensitive pixels: a new paradigm for low-power, low-cost 2D and 3D sensing,’ SDA12 8288-05.
L. Lipton, ‘Polarizing aperture stereoscopic cinema camera,’ SDA12 8288-06.
B. Zhang, S. Kothandaraman, and A. U. Batur, ‘Auto convergence for stereoscopic 3D cameras,’ SDA12 8288-09.
I. E. Pekkucuksen, A. U. Batur, and B. Zhang, ‘A real-time misalignment correction algorithm for stereoscopic 3D cameras,’ SDA12 8288-0J.
A. Said and B. Culbertson, ‘Analysis and management of geometric distortions on multiview displays with only horizontal parallax,’ SDA12 8288-1J.
D. Lanman, G. Wetzstein, M. Hirsch, W. Heidrich, and R. Raskar, ‘Beyond parallax barriers: applying formal optimization methods to multilayer automultiscopic displays,’ SDA12 8288-0A.
S. Uchida and Y. Takaki, ’360-degree three-dimensional table-screen display using small array of high-speed projectors,’ SDA12 8288-0D.
C. Vienne, L. Blondé, and D. Doyen, ‘Visual fatigue versus eye-movements,’ SDA12 8288-12.
D. Kane, R. T. Held, and M. S. Banks, ‘Visual discomfort with stereo 3D displays when the head is not upright,’ SDA12 8288-14.
S. Sawada and H. Kakeya, ‘Integral volumetric imaging with high resolution and smooth motion parallax,’ SDA12 8288-1R.
H. Yamamoto, H. Bando, R. Kujime, and S. Suyama, ‘Design of crossed-mirror array to form floating 3D LED signs,’ SDA12 8288-20.
Gregg E. Favalora is Principal of Optics for Hire, an optics-centric product engineering consultancy that develops LED illumination, medical devices, toys, 3D innovations, and other opto-electronic systems.
Note: Many presentations from the SD&A 3D technology conference are now available for free online viewing at www.stereoscopic.org/2012/program.html.