ETH Zurich's weekly web journal - auf deutsch
ETH Life - wissen was laeuft ETH Life - wissen was laeuft


ETH Life - wissen was laeuft ETH Life - wissen was laeuft
Home

ETH - Eidgenoessische Technische Hochschule Zuerich - Swiss Federal Institute of Technology Zurich
Section: Science Life
deutsche Version english Version
Print-Version Drucken

Published: 15.07.2004, 06:00
Modified: 14.07.2004, 17:47
Films from the ETH Laboratory for Software Technology
A further dimension

They're always on the move with two cameras, directed in parallel at what they meet with: the ETH students from the 3D film project of the Computer Science Department. Everything is shot in double in order to conjure up a three-dimensional world to present to viewers on the screen. Last week they showed seven short video films.

By Richard Brogle

A hand emerges from the screen and moves towards a viewer, who draws back, and then moves to the floor to grasp the case lying there – and carries it off. No, not a vision of the future but a preview of the first short film in 3D from the Laboratory for Software Technology at the Institute for Computer Systems at ETH Zurich. A few of the 20-odd filmmakers are also present. They made the seven short films in the framework of a course of lectures on "Stereoscopic Imaging” from the Computer Science Department as a semester assignment. Lecturer Dr. Cary D. Kornfeld, who teaches the course, supervised the students. A reputation for quality seems to have spread quickly: from 19 students at the beginning the number has swelled during the semester to around 30. PhD student Patrick de la Hamette says: "Of the three courses I started out on, this is the only one I attended to the end. What was good is that we got to do practical work". But he also emphasises that the accrued credits were not free lunch, "I never applied myself more to a course than I did to this one".

Doctoral students Susanne Cech (left) and Patrick de la Hamette (right) in the cutting room. large

Prize for the best animal actor

The time and energy expended on the filming and cutting seems to have been immense. Kornfeld: "About 1,000 working minutes went into each minute of finished film." Nonetheless, PhD student Susanne Cech says she would do the course again. "It was a brilliant atmosphere. Cary spent huge amounts of time helping us." Her brother, "who's heavily into stereoscoping images in his free time", pointed out the course to her. "He said that this was a course that I absolutely had to take.“ The technical challenges were not negligent either. Over the entire project, data with a volume of two terabytes (the amount of data on over 3,100 CDs) was processed on super-conducting computers with four processors.

Without glasses the screen shows two pictures, with glasses only one picture is seen with each eye. large

At the end of the project, just as in a real film festival, prizes were awarded in diverse categories. But the prizes were arranged so that every film got at least one. The Sushi film, for example, was awarded the prize for the "Best Animal Actor". There was also a "Cutest Smile“ prize category.


continuemehr

Understanding the human eye

The film project was preceded by one of Kornfeld's lectures, in which students started out by learning the basics of image processing and recording technique. They learned, for example, how to use colour, how the human eye reacts to optical stimulation and how to control and direct a camera. Asked whether these things belonged in a computer science, or electrotechnical lecture, Kornfeld replies: "Yes, because only those who know how eye and brain react to a specific picture can develop methods to compress the data to a maximum without the viewer being aware of the process."

Around 100 people were present at the premiere of the 3D videos. large

The area of electronic image processing was also addressed in the course. The students found out, for example, that optical scanners cannot differentiate more than 10 bits of colour depth. Everything above that is only calculated. The situation is similar when it comes to resolution. According to Cary Kornfeld the typical commercially available flatbed scanner can only resolve a maximum of 300 dots per inch (dpi) mechanically, anything more is interpolated. Kornfeld: "Values of 2,400 dpi are mere sales patter."

The trick's in the filter

Students also applied themselves to the 3D effect. The first thing one notices when entering the laboratory is that two projectors are directed at the screen. To create a 3D effect the right eye and the left eye must see two different pictures. Just as in the real world, a pair of human eyes never see exactly the same thing because they are spaced a couple of centimetres apart. Processing this effect is what makes our brains produce 3D pictures. But how does one succeed in making the eyes see two different pictures on the screen? The trick of this three dimensional perception lies in the use of polarisation filters in front of the projectors and the lenses of special glasses worn by the viewer. The filter only lets light through belonging to a specific plan vibration. In front on one projector there is a polarisation filter that only lets horizontally directed light through, the other can only be permeated by vertically directed light. Two pictures are projected onto the screen that are perceived simultaneously by the naked eye, because our eyes can see both horizontally and vertically directed light. If one now puts on the glasses whose left lens only lets through vertical light while the right only lets through horizontal light, then each eye sees a single picture from each projector. In other words the pictures are initially "produced" separately and united on the screen, separated again by the glasses and finally combined to form a 3D picture in the brain.

Two differently directed polarisation filters in the glasses and in front of the projectors conjure up a 3D world on the screen. large


References:
Course on Stereoscopic Imaging: www.lst.inf.ethz.ch/teaching/lectures/ss04/230/



You can write a feedback to this article or read the existing comments.




!!! Dieses Dokument stammt aus dem ETH Web-Archiv und wird nicht mehr gepflegt !!!
!!! This document is stored in the ETH Web archive and is no longer maintained !!!