HUMAN COMPUTER INTERACTION
LECTURE 4
virtual reality and 3D
interaction
positioning in 3D space
moving and grasping
seeing 3D (helmets and caves)
positioning in 3D space
cockpit and virtual controls
steering wheels, knobs and dials
… just like real!
the 3D mouse
six-degrees of movement: x, y, z
+ roll, pitch, yaw
data glove
fibre optics used to detect finger
position
VR helmets
detect head motion and possibly
eye gaze
whole body tracking
accelerometers strapped to
limbs or reflective dots and video
processing
pitch, yaw and roll
yaw
roll
pitch
3D displays
desktop VR
ordinary screen, mouse or
keyboard control
perspective and motion give
3D effect
seeing in 3D
use stereoscopic vision
VR helmets
screen plus shuttered specs,
etc.
VR headsets
small TV screen for each eye
slightly different angles
3D effect
VR motion sickness
time delay
move head … lag … display moves
conflict: head movement vs. eyes
depth perception
headset gives different stereo distance
but all focused in same plane
conflict: eye angle vs. focus
conflicting cues => sickness
helps motivate improvements in technology
simulators and VR caves
scenes projected on
walls
realistic environment
hydraulic rams!
real controls
other people
physical controls, sensors
etc.
special displays and gauges
sound, touch, feel, smell
physical controls
environmental and bio-sensing
dedicated displays
analogue representations:
dials, gauges, lights, etc.
digital displays:
small LCD screens, LED lights, etc.
head-up displays
foundin aircraft cockpits
show most important controls
… depending on context
Sounds
beeps, bongs, clonks, whistles and whirrs
(sound of something in rapid motion)
used for error indications
confirmation of actions e.g. keyclick
also see chapter 10
Touch, feel, smell
touch and feeling important
in games … vibration, force feedback
in simulation … feel of surgical instruments
called haptic devices
texture, smell, taste
current technology very limited
physical controls
specialist controls needed …
industrial controls, consumer products,
etc.
easy-clean
smooth buttons
multi-function
control
large buttons
clear dials
tiny buttons
Environment and bio-sensing
sensors all around us
car courtesy light – small switch on door
ultrasound detectors – security, washbasins
RFID security tags in shops
temperature, weight, location
… and even our own bodies …
iris
scanners, body temperature, heart rate,
galvanic skin response, blink rate
paper: printing and
scanning
print technology
fonts, page description,
WYSIWYG
scanning, OCR
Printing
image made from small dots
allows any character set or graphic to be
printed,
critical features:
resolution
size and spacing of the dots
measured in dots per inch (dpi)
speed
usually measured in pages per minute
cost!!
Types of dot-based printers
dot-matrix printers
use inked ribbon (like a typewriter
line of pins that can strike the ribbon, dotting the paper.
typical resolution 80-120 dpi
ink-jet and bubble-jet printers
tiny blobs of ink sent from print head to paper
typically 300 dpi or better .
laser printer
like photocopier: dots of electrostatic charge deposited on drum, which
picks up toner (black powder form of ink) rolled onto paper which is then
fixed with heat
typically 600 dpi or better.
Printing in the workplace
shop tills
dot matrix
same print head used for several paper rolls
may also print cheques
thermal printers
special heat-sensitive paper
paper heated by pins makes a dot
poor quality, but simple & low maintenance
used in some fax machines
Fonts
Font – the particular style of text
Courier font
Helvetica font
Palatino font
Times Roman font
§´µº¿Â Ä¿~ (special symbol)
Size of a font measured in points (1 pt about 1/72”)
(vaguely) related to its height
This is ten point Helvetica
This is twelve point
This is fourteen point
This is eighteen point
and this is twenty-four point
Fonts (ctd)
Pitch
fixed-pitch – every character has the same width
e.g. Courier
variable-pitched – some characters wider
e.g. Times Roman – compare the ‘i’ and the “m”
Serif or Sans-serif
sans-serif – square-ended strokes
e.g. Helvetica
serif – with splayed ends (such as)
e.g. Times Roman or Palatino
Readability of text
lowercase
easy to read shape of words
UPPERCASE
better for individual letters and non-words
e.g. flight numbers: BA793 vs. ba793
serif fonts
helpsyour eye on long lines of printed text
but sans serif often better on screen
Page Description Languages
Pages very complex
different fonts, bitmaps, lines, digitised photos, etc.
Can convert it all into a bitmap and send to the printer
… but often huge !
Alternatively Use a page description language
sends a description of the page can be sent,
instructions for curves, lines, text in different styles, etc.
like a programming language for printing!
PostScript is the most common
Screen and page
WYSIWYG
what you see is what you get
aim of word processing, etc.
but …
screen: 72 dpi, landscape image
print: 600+ dpi, portrait
can try to make them similar
but never quite the same
so … need different designs, graphics etc, for screen
and print
Scanners
Take paper and convert it into a bitmap
Two sorts of scanner
flat-bed: paper placed on a glass plate, whole page converted into
bitmap
hand-held: scanner passed over paper, digitising strip typically 3-4” wide
Shines light at paper and note intensity of reflection
colour or greyscale
Typical resolutions from 600–2400 dpi
Scanners (ctd)
Used in
desktop publishing for incorporating
photographs and other images
document storage and retrieval systems,
doing away with paper storage
+ special scanners for slides and photographic
negatives
Optical character recognition
OCR converts bitmap back into text
different fonts
createproblems for simple “template
matching” algorithms
more complex systems segment text,
decompose it into lines and arcs, and
decipher characters that way
page format
columns, pictures, headers and footers
Paper-based interaction
paper usually regarded as output only
can be input too – OCR, scanning, etc.
Xerox PaperWorks
glyphs – small patterns of /\\//\\\
used to identify forms etc.
used with scanner and fax to control applications
more recently
papers micro printed - like wattermarks
identify which sheet and where you are
special ‘pen’ can read locations
know where they are writing
memory
short term and long term
speed, capacity, compression
formats, access
Short-term Memory - RAM
Random access memory (RAM)
on silicon chips
100 nano-second access time
usually volatile (lose information if power turned off)
data transferred at around 100 Mbytes/sec
Some non-volatile RAM used to store basic set-up
information
Typical desktop computers:
64 to 256 Mbytes RAM
Long-term Memory - disks
magnetic disks
floppy disks store around 1.4 Mbytes
hard disks typically 40 Gbytes to 100s of Gbytes
access time ~10ms, transfer rate 100kbytes/s
optical disks
use lasers to read and sometimes write
more robust that magnetic media
CD-ROM
- same technology as home audio, ~ 600 Mbytes
DVD - for AV applications, or very large files
Blurring boundaries
PDAs
often use RAM for their main memory
Flash-Memory
used in PDAs, cameras etc.
silicon based but persistent
plug-in USB devices for data transfer
speed and capacity
what do the numbers mean?
some sizes (all uncompressed) …
this book, text only ~ 320,000 words, 2Mb
the Bible ~ 4.5 Mbytes
scanned page ~ 128 Mbytes
(11x8 inches, 1200 dpi, 8bit greyscale)
digital photo ~ 10 Mbytes
(2–4 mega pixels, 24 bit colour)
video ~ 10 Mbytes per second
(512x512, 12 bit colour, 25 frames per sec)
virtual memory
Problem:
running lots of programs + each program large
not enough RAM
Solution - Virtual memory :
store some programs temporarily on disk
makes RAM appear bigger
But … swopping
program on disk needs to run again
copied from disk to RAM
slows t h i n g s d o w n
Compression
reduce amount of storage required
lossless
recover exact text or image – e.g. GIF, ZIP
look for commonalities:
text: AAAAAAAAAABBBBBCCCCCCCC 10A5B8C
video: compare successive frames and store change
lossy
recover something like original – e.g. JPEG, MP3
exploit perception
JPEG: lose rapid changes and some colour
MP3: reduce accuracy of drowned out notes
Storage formats - text
ASCII - 7-bit binary code for to each letter and character
UTF-8 - 8-bit encoding of 16 bit character set
RTF (rich text format)
- text plus formatting and layout information
SGML (standardized generalised markup language)
- documents regarded as structured objects
XML (extended markup language)
- simpler version of SGML for web applications
Storage formats - media
Images:
many storage formats :
(PostScript, GIFF, JPEG, TIFF, PICT, etc.)
plus different compression techniques
(to reduce their storage requirements)
Audio/Video
again lots of formats :
(QuickTime, MPEG, WAV, etc.)
compression even more important
also ‘streaming’ formats for network delivery
methods of access
large information store
long time to search => use index
what you index -> what you can access
simple index needs exact match
forgiving systems:
Xerox “do what I mean” (DWIM)
SOUNDEX – Smith ~ Schmidt
access without structure …
free text indexing (all the words in a document)
needs lots of space!!
processing and networks
finite speed (but also Moore’s law)
limits of interaction
networked computing
Finite processing speed
Designers tend to assume fast processors, and make interfaces
more and more complicated
But problems occur, because processing cannot keep up with all the
tasks it needs to do
cursor overshooting because system has buffered keypresses
icon wars - user clicks on icon, nothing happens, clicks on another, then
system responds and windows fly everywhere
Also problems if system is too fast - e.g. help screens may scroll
through text much too rapidly to be read
Moore’s law
computers get faster and faster!
1965 …
Gordon Moore, co-founder of Intel, noticed a pattern
processor speed doubles every 18 months
PC … 1987: 1.5 Mhz, 2002: 1.5 GHz
similar pattern for memory
but doubles every 12 months!!
hard disk … 1991: 20Mbyte : 2002: 30 Gbyte
baby born today
record all sound and vision
by 70 all life’s memories stored in a grain of dust!
the myth of the infinitely
fast machine
implicit assumption … no delays
an infinitely fast machine
what is good design for real machines?
good example … the telephone :
type keys too fast
hear tones as numbers sent down the line
actually an accident of implementation
emulate in deisgn
Limitations on interactive
performance
Computation bound
Computation takes ages, causing frustration for the user
Storage channel bound
Bottleneck in transference of data from disk to memory
Graphics bound
Common bottleneck: updating displays requires a lot of effort -
sometimes helped by adding a graphics co-processor optimised to take
on the burden
Network capacity
Many computers networked - shared resources and files, access to
printers etc. - but interactive performance can be reduced by slow
network speed
Networked computing
Networks allow access to …
large memory and processing
other people (groupware, email)
shared resources – esp. the web
Issues
network delays – slow feedback
conflicts - many people update data
unpredictability