SeeFront Shows Eye Tracked Autostereoscopic Laptop at SD&A
At the Stereoscopic Displays and Applications conference recently held in San Jose, SeeFront (Hamburg, Germany) showed a laptop that demonstrated means to reversibly convert a conventional 2D LCD display into an autostereoscopic 3D display. The prototype system was composed of an Apple laptop with a conventional 2D LCD, a lenticular sheet mountedable in front of the LCD, eye tracking accomplished by use of the single camera built into the laptop and some simple-to-use software.
Let me start by acknowledging that the individual technological elements used in this system have been seen before. That is, a digital camera has certainly been used for eye tracking in other applications. The idea of a retrofitable lenticular has also been seen before, for example in the iArt3D product. But…the combination of the retrofittable lenticular, eye tracking and easy-to-use set-up and control software is novel and resulted in a laptop with impressive 3D capabilities.
On booth duty at the exhibition, Christoph Grossmann, SeeFront’s Founder and CEO explained that a characteristic of the lenticular approach is that the underlying LCD suffers almost no reduction in brightness in the 3D mode. This is an advantage over a 3D laptop system based on shutter glasses. In addition, there is little change in color coordinates in switching between the 2D and 3D modes in a lenticular system.
Since the autostereoscopic image is intended for a single viewer, the system is configured such that the resolution in the 3D mode is half that in the 2D mode. This reduction in resolution is similar to that found in a MicroPol + passive glasses based laptop 3D display.
When I sat down to use the 3D laptop, the first step was to set-up the eye tracking system. This took just a few seconds but required that I remove my glasses. The process was further assisted by not having others directly behind me or otherwise in view of the camera.
The eye tracking function considerably expanded the autostereoscopic 3D head box thus addressing one of the principle objections to conventional lenticular autostereoscopic 3D. An important point to note regarding SeeFront’s tracking implementation is that the user’s position is tracked not only along a fixed or optimal distance plane but also within a three-dimensional volume which is limited by a minimal and a maximal distance plane.
Next step was the process of mounting the lenticular in front of the LCD.
In a conventional lenticular-based autostereoscopic display, the geometry of the lenticular is determined by the pixel pitch of the LCD. As a result, a given lenticular will typically work only with an LCD having that specific pixel pitch. In addition, combining the lenticular and the LCD requires precise alignment. The reason for this is that each lenticule must be assigned to a discrete number of pixel-rows or slanted sub-pixel "chains." (Call them pixel entities.)
This is not the case in the SeeFront approach. There is no need for precise mechanical registration of the lenticular with respect to the pixel array of the LCD. A software set-up and control panel provides means for the user to adjust the image until the 3D effect is most pleasing.
In the SeeFront 3D display, each of the two perspectives assigned to a lenticule covers a larger number of pixel entities than in a system with multiple views. In addition, the number and location of these pixel entities change with the movements of the user. For each viewing position, the SeeFront algorithm determines the parts of the image that can be seen by each eye and then splits and assigns the information from the incoming video stream or image accordingly. This dynamic process takes place several times per second (normally at 30 Hz or more). In this way, the SeeFront 3D lenticular technology should work with LCDs having any pixel pitch or screen diagonal size.
The lenticular mounting procedure and the set-up adjustment process took less than a minute.
Perhaps the principle artifact I noted in the SeeFront 3D image was a slight Moiré pattern between the lenticular and the pixels. The artifact was visible when the screen was primarily white, as in a Word document page. Grossmann expects this effect to be minimized in the next generation prototype accomplished by changing the angle between the lenses in the lenticular and the pixel array.
SeeFront’s business model is to license its 3D technology. Since the technology fills a vacant product niche and since the resulting 3D is quite credible, I can see some level of success for this technology. We shall see.
By Art Berman, DisplayDaily
