Week 3

Introduction and History of VR / AR

Warning about Jargon

Warning about jaded old guy talking about how we did all this back in the 90s in a "Hey you kids! Get off my lawn" voice

General Definitions

'Virtual Reality' and 'Augmented Reality' buzzwords that can mean a lot of different things depending on who you talk to, and the line between them can be blurred.

There have been various ways of of trying to create scales or continuums from fully real worlds to fully virtual worlds - https://en.wikipedia.org/wiki/Reality%E2%80%93virtuality_continuum - in general looking something like this:

- pure reality - you are completely immersed in the natural world with minimal access to any synthetic worlds (e.g. your smartphone)
- augmented reality - you out in the real world with some gadgets (phone, headset) that allow you to experience the real world and a synthetic world simultaneously where the tech could range from a watch to a smart phone to a see through head mounted display.
- virtual reality - you are completely immersed in a synthetic world with minimal access to the real world

There is a large grey area of 'mixed reality' that covers most of the space from purely real to purely virtual where we spend most of our lives. I may be walking down the street looking at a set of directions on my smartphone, or I may be sitting on my couch with some chips and a drink playing a videogame on a big TV.

We are all pretty familiar with the 'pure reality' side of the spectrum so lets start with Virtual Reality

The key element to virtual reality is immersion ... the sense of being surrounded.

A good novel is immersive without any fancy graphics or audio hardware. You 'see' and 'hear' and 'touch' and 'taste' and 'smell'

A good play or a film or an opera can be immersive using only sight and sound.

But they aren't interactive which is another key element.

The children's 'Choose Your Own Adventure' books in the late 70s added limited interaction to books giving the reader a handful of choices every few pages that would lead to 40 endings in the case of the first book 'The Cave of Time', but it was computers that would run with this concept.

(Image from https://en.wikipedia.org/wiki/Choose_Your_Own_Adventure#/media/File:Cave_of_time.jpg)

There is a sample page here: http://www.sundancepub.com/c/@.IXKWkq4ugJuY/Pages/cyoa_studentbook_zoom.html

(and if you wish more 'serious' literary context on the topic, please see https://en.wikipedia.org/wiki/Hypertext_fiction for several very serious pieces of literature with similar methods)

Older textual computer games from the late 70s and early 80s  such as Adventure, Zork, and the Scott Adams (not the Dilbert guy) adventures are immersive and interactive and place the user within a computer generated world, though that world was created only through text. You can play adventure online at http://www.astrodragon.com/zplet/advent.html. You can play the personal computer version of Zork online at  http://textadventures.co.uk/games/view/5zyoqrsugeopel3ffhz_vq/zork. The Scott Adams adventures are playable at http://www.freearcade.com/Zplet.jav/Scottadams.html

video: https://www.youtube.com/watch?v=TNN4VPlRBJ8

Games in the early 80s started to incorporate primitive computer graphics visuals to go along with the text, such as Mystery House (1980) below.
video: https://www.youtube.com/watch?v=asOhTnQv8PE

and even simple 1st person graphics in games such as Akalabeth (1980) and Wizardry (1981), though the screen refresh rate was something less than real-time. The screen took a long time (up to several seconds) to re-draw so these games tended to be more strategy-based on a turn-taking model.
video: https://www.youtube.com/watch?v=P0jSh_MKM1M

Myst in 1993 took the visuals to a whole new level using CDROM storage for all of its (for the time) very realistic imagery, though you could only move between a set of fixed locations and viewpoints and use the mouse to click on objects to interact with them - https://www.youtube.com/watch?v=4xEhJbeho7Q

Moving on towards more modern computer games, they are immersive and interactive. These also have the advantage of being real-time running at 30+ frames per second, another key element.

Another key element of VR is a viewer centered perspective where you 'see' through your own eyes as you move through a computer generated space. Akalabeth, Wizardry, and Myst were first person view games, though you could only look where the game allowed you to look. Modern first person shooters and other games use this view as you move through a virtual world and interact with objects there, and more often than not kill everyone you meet. The way you see the environment is limited to a screen with a narrow angle of view and you use a keyboard / joystick / gamepad to change your view of that scene, and interact.

One of the most successful early ones was Wolfenstein 3D from 1992 - https://www.youtube.com/watch?v=NdcnQISuF_Y

(image from Wikipedia)

Of course as time went on the visuals became better and some would stick with a first person perspective and others using a third person perspective to better show what was going on around the player.

First person or viewer centered perspective on the left vs third person perspective on the right from the Jedi Knight series from the 1990s.

VR adds the concepts of head tracking, wide field of view and stereo vision

Head tracking allows the user to look around the computer generated world by naturally moving his/her head. A wide field of view allows the computer generated world to fill the user's vision. Stereo vision gives extra cues to depth when objects in the computer generated world are within a few feet.

As Dan Sandin, original co-director of evl likes to say, this gives us the first re-definition of perspective since the Renaissance in the 16th century.

Albrecht Dürer, Draughsman Drawing a Recumbent Woman (1525) Woodcut illusion from 'The Teaching of Measurements.'

Natural interaction is also important in VR. If you want to reach out and touch a virtual object then tracking the users hands lets the user do that, rather than using a keyboard or gamepad to 'tell' your virtual representation to interact.

Audio also plays a very important role in immersion (try listening to a modern Hollywood film without its musical score) and haptic (touch) feedback can provide important cues while in smaller immersive spaces.

And there is some work in trying to deal with smell (the HITLab in the late 90s, and Yasuyuki Yanagi, Advanced Telecommunications Research Institute, Kyoto more recently) and taste (Hiroo Iwata, University of Tsukuba.)

So here is a picture that puts a lot of this together ... Randy Smith of General Motors in their CAVE in the mid 1990s. Randy is real. The car seat Randy is sitting in is real. The rest is computer generated.

Augmented Reality has a very similar feature set but whereas Virtual Reality usually is set up in controlled settings - typically indoors within a fixed space where you can set up and calibrate tracking systems, and you have access to power for the computers to drive the graphics), augmented reality usually takes place out in the real world where tracking is less accurate, power needs to be portable, and computational power needs to be portable.

Augmented reality has the additional constraint that the synthetic world it is creating must match up with the real world.

Better batteries help with making power more available, and access to cloud computing resources can help offload the computation but accurate tracking is still difficult. In some minimal levels of AR where I want to know the weather, I probably only need accuracy down to the city level, if I want to know where is the closest coffee shop then I need accuracy down to the block level, if I want to see what power lines are running under the street or the names of the people who are walking past me then I need much more accuracy.

A Bit of History

1793 - Fixed 360 degree Panoramas - Robert Barker in Leicester Square, London - link

1840s - Moving Panoramas - John Banvard's Mississippi Panoramas - 3.6m (12 feet) high and 800m (2600 ft) long - link

1800ds - Stereoscope - https://en.wikipedia.org/wiki/Stereoscope

(image from https://en.wikipedia.org/wiki/Stereoscope#/media/File:Holmes_stereoscope.jpg)

1950 - The Veldt by Ray Bradbury

1960 - Morton Helig

Sensorama - https://www.youtube.com/watch?v=vSINEBZNCks

(image from http://www.mortonheilig.com/InventorVR.html)

patent for first HMD

(image from http://accad.osu.edu/~waynec/history/lesson17.html)

1965 - Ivan Sutherland - University of Utah

1966 - Ivan Sutherland

(image from http://accad.osu.edu/~waynec/history/tree/images/hmd.JPG)

1967 - Fred Brooks - University of North Carolina

1973 - The Recreation Room (later called the Holodeck) in Star Trek: the Animated Series

mid 70s - mid 80s Myron Krueger

(image from http://resumbrae.com/ub/dms424/05/01.html)

1977 - Richard Sayre, Dan Sandin, Tom DeFanti - UIC

1979 - Eric Howlett

1982 - Thomas Furness III

1984 - Michael McGreevy and friends

1985 - Jaron Lanier - VPL research

1986 - Kazuo Yoshinaka - NEC

1989 - Autodesk

1989 - Fake Space Labs

(image from: http://www.fakespacelabs.com/tools.html)

1991 - Virtuality - https://en.wikipedia.org/wiki/Virtuality_(gaming)

1992 - Electronic Visualization Laboratory, UIC

1992 Tom Caudell - Boeing

(image from http://thearea.org/augmented-reality-in-the-aerospace-industry/)

1992 Steve Feiner and friends - Columbia University

1993 - GMD - German National Research Center for Information Technology

1993 - SensAble Technology

1992 Steve Feiner and friends - Columbia University

(IMAGES FROM http://monet.cs.columbia.edu/projects/mars/touring.html)

Mid 90s - Steve Mann - MIT

1998 - TAN / Royal Institute of Technology in Stockholm

1998 - Electronic Visualization Laboratory, UIC

1999 - Mark Billinghurst - HITLab at University of Washington

2003 - University of Arizona

2009 - UCSD Calit2 / KAUST

2012 - Electronic Visualization Laboratory, UIC

2013 - Google Glass

(image from https://en.wikipedia.org/wiki/Google_Glass)

2014 - Oculus / Vive / Gear

2014 - Google Cardboard

2016 - Microsoft

2017 - Dell, Asus and others release their Microsoft mixed reality based headsets ($300) with 'inside out' tracking where the hand-held controllers are tracked from the HMD

2018 - Magic Leap AR display Dev Kits released ($2,300 with no PC or tether)


VR has gone through several hype phases with the biggest being in the mid 80s and mid 90s. With the release of low cost headsets we are now in the midst of another hype phase. AR is in its first hype phase.

Where are we now?

(image from http://www.gartner.com/newsroom/id/3412017)

back in 1995, when the first of these charts came out, VR was just sliding down into the trough of Disillusionment
Gartner Emerging Tech Hype Chart 1995
(image from https://www.gartner.com/doc/484424/gartners-hype-cycle-special-report#1169528434)

VR Hardware

HDM, BOOM, and Fish Tank VR

For large format based systems, some companies that sell these things are:

For Head Mounted Displays, the previous generation of $10,000 - $20,000 displays by companies like NVIS have mostly been supplanted by a new generation of low cost gaming-related displays:

Instead of totally isolating the user from the real world, Augmented Reality displays overlay computer graphics onto the real world with devices like Google Glass and the Microsoft HoloLens and the Magic Leap

and there are other interesting solutions that have been in development for a couple decades such as the Virtual Retinal Display

and to some extent your smartphone or tablet with GPS, camera, and a Gyroscope already acts as an AR display.

Current VR Uses

There is quite a bit of work going on in various research labs in VR. New devices are being created, new application areas being worked on, new interaction techniques being explored, and user studies being performed to see if any of these are valuable. What is much harder is getting the technology and the applications out of the research lab and into real use at other sites - getting beyond the 'demo' stage to the 'practical use' stage is still very difficult.

Current AR Uses

Homework assignment due Friday at 9pm Chicago time

google translate in its smartphone app shows some of the more 'serious' potential of augmented reality as it allows you to automatically translate text into other languages. For this assignment you should find something in a foreign language in the real world (not by bringing up images in google) and take a photo of it, and then save 2 or 3 screens when google translate is translating it with varying degrees of success. Create another public webpage on your site and attach the photos and a page of text you have written on how you think this capability would be most effectively used. Right now on your smartphone it allows you to have a lens that you can move over the real world and see it modified on the phone's screen, but what if you were running this in a future AR pair of glasses or contact lenses, and it was automatically translating everything it sees to the language of your choice and hiding the original text from real world. What are the pros and cons of that? How much control do you think the user should have over the way the synthetic is mapped over the real.

VR and AR Components

I'm going to give a brief overview here and then we will go into each of these areas in more detail in the coming weeks


For virtual reality it is important to note that the goal is not always to recreate reality.

Computers are capable of creating very realistic images, but it can take a lot of time to do that. In VR we want at bare minimum 20 frames per second and preferably 60+ in stereo.

For comparison:

The trade off is image quality (especially in the areas of smoothness of polygons, anti-aliasing, lighting effects, transparency) vs speed of rendering. In some cases, like General Motors, they sacrifice frame rate (frames per second) for better visual quality.

Gamers also tend to want higher frame rates than people watching TV / Movies / YouTube videos.

In AR we are typically not covering the entire field of view of the user so the rendering requirements are lower, but there is a greater need to do a better compositing between the real and the synthetic (i.e. based on lighting conditions) and faster graphics updating.

If we want stereo visuals then we need a way to show a slightly different image to each eye simultaneously. The person's brain then fuses these two images into a stereo image.

One way is to isolate the users eyes (as in a HMD or BOOM) and feed a separate signal to each eye using 2 display devices where each eye watches its own independent display (as in older HMDs), or take a single wide display and render the left and right eye views onto the same screen and then make sure each eye can only see its appropriate half of the display (as in current less expensive HMDs).

Another way is to show the imagery on a larger surface and then filter which part of the image the user sees. There are several different ways to do this.

We can use polarization (linear or circular) - polarization was used in 3D theatrical films in the 1950s and 1980s and the current generation. One projector is polarized in one direction to show images for the left eye, and the other projector is polarized in the other direction to show images for the right eye. Both images are shown on the same screen and the user wears lightweight glasses to disambiguate them.

This same technology can be used on televisions by adding a polarized film in front of the display where even lines are polarized in one direction and odd lines are polarized in the other direction. The user only sees half of the resolution of the display with each eye. This is the technology we use in CAVE2.

We can use colour - this has been done for cheaper presentation of 3D theatrical films since the 50s with red and blue (cyan) glasses as you only need a single projector, or a standard TV. It doesn't work well with colour and is somewhat headache inducing after an hour.

We can use time - this was common in VR in the 90s and the 00s as in the original CAVE.  Here we show the left eye image for a given frame then the right eye image for the same frame, then move on to the next frame. The user wears LCD shutter glasses which ensure that only the correct eye sees the correct image by going opaque on the eye that should be seeing nothing. These glasses used to cost over $1000 each in the early 90s. They were the basis for the early 3D televisions and cost around $100 per pair. Now they are down to $30 per pair.

In all these cases both of the eyes are focusing at a specific distance - wherever the screen is located. There is no way for the user to change focus and bring parts of the scene into focus and let others go out of focus as in the real world . 

"people hate helmets, but people like sunglasses"

ergonomics and health issues of various displays

Typically museums and other places with many visitors it is necessary to either give the glasses away to the user (with the paper ones) or wash them (with the polarizing ones) to keep things sanitary. This is more difficult with HMDs and AR headware where people have tried using alcohol wipes.

People typically don't spend all day in VR but people may spend all day in AR. VR also tend to be done in private where AR is more done outside. AR headware has to be light and unobtrusive, but still be able to operate. The entire computation system may be in the headgear as well, or some may be offloaded to a smart phone or to the cloud. Google Glass was one light solution. Headware for bikers like the Solos http://www.solos-wearables.com/ are another, as is the Microsoft HoloLens.

Image Generator

Need a computer capable of driving the display device at a fast enough rate to maintain the illusion.

In the past (i.e. the 90s) that usually means either simple scenes, very specialized graphics hardware, or a lot of work in optimizing the software. But this is less true today where scenes are getting more complex, the hardware more commonplace, and the software more capable, mostly thanks to the video-game industry.

Benchmarks on CPUs and graphics cards aren't really very meaningful. They can give ballpark figures but there are a lot of factors that combine to give the overall speed/quality of the virtual environment.

Multiple processors are usually required, since there tend to be multiple simultaneous jobs to be performed - i.e. generating the graphics, handling the audio, synchronizing with network events.

Multiple graphics engines are pretty much required if you have multiple display surfaces

Ability to 'pipeline' the graphics is pretty much required

With a very fast network it is possible to render the graphics remotely on a more powerful computer and just use the local display as a receiver.

Tracking System

At minimum you want to track the position (x, y, z) and orientation (roll, pitch, yaw) of the user's head - 6 degrees of freedom.

You often want to track more than that - 1 or 2 hands, legs?, full body?

How accurate is the tracking?

How far the user can move - what size area must the tracker track?

Can line of sight be guaranteed between the tracker and the sensors, which is necessary in many tracking systems?

What kinds of latencies are acceptable?

Input Device

Input devices are perhaps the most interesting area in VR research. While the user can move their head 'naturally' to look around, how does the user navigate through the environment or interact with the things found there?

Audio System

Ambient sounds are useful to increase the believably of a space

Sounds are useful as a feedback mechanism

Important in collaborative applications to relay voice between the various participants

Spatialized sound can be useful


Often useful to network a VR world to other computers.

We need high bandwidth networking for moving large amounts of data around, but even more important that that we need Quality of Service guarantees, especially in regards to latency and jitter.

what we want

Photo of the classic evl CAVE from the early 90s with 4 1-megapixel screens with active (shutter glasses) stereo giving typically 2 megapixels to each eye depending where you stand) , a 10' by 10' area to move, magnetic tracking for the head and one controller. Total cost was around $1,000,000 in 1991 dollars (about $2,000,000 in 2018 dollars) with about $500,000 of that 1991 price for the refrigerator sized computers to drive it.

to put this hardware into context, in 1991 we had

In 2017/18 the VIVE could send roughly 1-megapixel to each eye, gives the user a similar space to walk around in, IR camera tracking for the head and two controllers for about $2,500 including the computer.

Coming Next Time

Vision / Visuals and Audio

last revision 9/9/18