media spaces and awareness. readings robert s. fish, robert e. kraut, robert w. root, ronald e....

45
Media Spaces and Awareness

Upload: loreen-butler

Post on 28-Dec-2015

225 views

Category:

Documents


0 download

TRANSCRIPT

  • Media Spaces and Awareness

  • ReadingsRobert S. Fish, Robert E. Kraut, Robert W. Root, Ronald E. Rice, Video as a technology for informal communication, CACM, v36,n1, January 1993, pp. 48-61.http://portal.acm.org/citation.cfm?id=151237

    Paul Dourish, Sara Bly, Portholes: supporting awareness in a distributed work group, Proceedings of CHI 92, 1992, Monterey, pp. 541-547.http://portal.acm.org/citation.cfm?id=142982

    Scott E. Hudson , Ian Smith, Techniques for addressing fundamental privacy and disruption tradeoffs in awareness support systems, Proceedings of CSCW 96, November 1996, pp.526-533.http://portal.acm.org/citation.cfm?id=240295

  • The death of distanceSince television was introduced people have foreseen a day when technology would make it possible to be there without traveling (telepresence)

    When two-way television is added to telephones, [] direct interaction will finally be available to those who are widely separated- Erving Goffman, Behavior in Public Places, 1963

  • Supporting distributed work groupsWork on video phones starting in the 60s Failure in the marketplaceVery expensiveSmall fuzzy pictureAnd the 70s, and

    If only we could get the cost down and the quality up

  • The Technology is Now HereVideo to/from a cell phone & POTS line with HDTV+ to the desktop, but...

    Its still not the same as being there

    Why?

  • Being there is a very rich experienceIt turns out just a small picture (and sound) from the remote location isnt enoughVideo actually only gives a small increment over just audio (for direct communications)

    There are a lot of small / subtle things in collocated interaction that are lostThey turn out to be importantSocial interactions rely on very subtle (and inherently ambiguous) cues

  • Media SpacesUse new media (audio and video transmission) to create a space for human-human interactionAlways onYou dont connect to a space, you are in it

    Provide opportunities for engagementSupport social interactionsInformal and serendipitous interactions

    Support awarenessExistence and presence of colleaguesWork status (e.g., for communication coordination)

  • Early media space systemsStarted with PARC Media Space systemGroup split between Palo Alto and PortlandHow can we use technology to improve our interactions? Cameras and displays in offices & public spacesVideo crossbar switches (analog)Manual at first, later computer controlledInherently a point-to-point technologyA / V link to branch site in Portland

  • Additional early systemsRAVE (EuroPARC)William Gaver, Thomas Moran, Allan MacLean, Lennart Lvstrand, Paul Dourish, Kathleen Carter, William Buxton, Realizing a video environment: EuroPARC's RAVE system, Proceedings of CHI '92, 1992, Monterey, pp. 27-35. http://portal.acm.org/citation.cfm?id=142754VideoWindow (Bellcore)Robert S. Fish, Robert E. Kraut, Barbara L. Chalfonte, The VideoWindow system in informal communication, Proceedings of CSCW '90, September 1990, pp. 1-11. http://portal.acm.org/citation.cfm?id=99335Cruiser (Bellcore) Fish, et al. in readingsCaveCat (U of Toronto)Marilyn M. Mantei, Ronald M. Baecker, Abigail J. Sellen, William A. S. Buxton, Thomas Milligan, Barry Wellman, Experiences in the use of a media space, Proceedings of CHI '91, March 1991, pp. 203-208. |http://portal.acm.org/citation.cfm?id=108888

  • RAVE (EuroPARC)UI reflecting intention and levels of engagementFunctionsBackground (public areas)Glance1-2 second peek into anothers spaceOne-way with non-speech audio indicator (partial reciprocity)Sweep (glances iterated over group)All or user settable groupVideo phone & Office shareSame action, but with different expressed intentionExplicit circle of friends access control for privacy

  • RAVE (EuroPARC)Use of non-speech audio particularly interestingUsed for indicate of one-way (viewing) visitsDesigned to give partial reciprocity Compare with Cruiser (next)Came shortly before the visitWell designed soundsDesigned for minimal disruptionIntuitive meaningPassed do they turn it off test (most dont)

  • CaveCat (Toronto)Same low level infrastructure as RAVESame computer controlled analog switchUsed 4-up displaySmall group instead of simple point-to-pointHighlighted a set of issues (see later)Unexpected affordancesMirror was important Monitoring office and contacts while away

  • VideoWindow (Bellcore)Oft-repeated video wall conceptFor another attempt at joining break rooms see for example: Gavin Jancke, Gina Danielle Venolia, Jonathan Grudin, J. J. Cadiz, Anoop Gupta, Linking public spaces, Proceedings of CHI '01, March 2001, pp. 530-537 http://portal.acm.org/citation.cfm?id=365352Try to join to spaces (break rooms) by placing camera and (very wide) video display against a wallOther space looks like its past the wallHigh quality media setup (special camera & display, etc.)Some successRemote participants interacted (much like local)But not being thereFreq of interaction was ~2.5x less than collocated

  • Cruiser (Bellcore)Designed explicitly around social phenomena(Partially) modeled after walking down the hallCruisingAlso glance function and video phone style interactionPrivacy via reciprocityIn normal space: if I can see you, you can see meStrong social conventions can be leveragedProblem with reciprocity is disruptionCruising clearly failed More like random telephone calls than walking down hall

  • Early systems were a qualified successDefinitely not the same as face-to-face, but still usefulProvided opportunities for social engagement which did make a substantial difference

    Lots of lessons learned and issues uncovered

  • Some identified needs (a priori and from early systems)Informal communications supportSocial interactions are inherently informalNeed spontaneity and serendipityA lot of work gets done bumping into people in the hallRequires very lightweight interactions Low intentionalityNeed to support unstructured, complex, & equivocal communications(from Gaver et al, RAVE)

  • AwarenessGeneral (background / peripheral) information about collaborators makes interaction smootherLots of very informal informationExistence, presence/absenceActivities, busyness, mood, habits, Not (necessarily) task-oriented communicationsbut still makes tasks smootherE.g., coordination of communications (when to interrupt)Each piece of information has low value, but valuable in the aggregateCant incur much cost to get each piece Leads to need for ambient displays

  • Awareness

    Although seemingly the most invisible, the use of the media space for peripheral awareness was perhaps its most important use. - Bly et al., Media Spaces, CACM 1/93 http://portal.acm.org/citation.cfm?id=151235c

  • Some identified needs (a priori and from early systems)

    Smooth transition from informal to direct communicationsDirect communications still critical for getting work done (task oriented actions)Shared artifacts / documents very important for direct communications

  • Issues identified from early systemsMedia-related issuesViewField of view & resolution Perceptual and privacy implicationsAnisotropic viewsLoss of gaze awareness, social distance distortionsMissing view controlMissing movement parallax & stereopsisAudioGeneral quality and noiseLoss of localization

  • Issues identified from early systemsControl & PrivacyMost early systems (partial exception Bellcore) were used mostly by creators and their friendsHighly motivated and self-selectedStill privacy issues raised by all investigatorsCamera is like working in front of a 1-way mirrorCant tell whos looking (loss of reciprocity)Like real world situations E.g., the man lurking behind the bush loss of reciprocity breaks social convention unease

  • Issues identified from early systemsFour facets of privacy needs from Rave paperDesire for control over who can see or hear us at any given timeDesire for knowledge of when someone is in fact seeing or hearing usDesire to know the intention behind the connectionThe desire to avoid connections being intrusions on our workKnowledge (& generally all reciprocity solutions) Intrusions / Disruption

  • Issues identified from early systemsOther social aspectsMissing cues for public / private nature of spaceMissing cues for approach & negotiationSocial interaction prerequisites (from VideoWindow paper)Concentration of suitable partnersCo-Presence Opportunity to engageLow personal costVisual channel Cues for engagement are primarily visual

  • Additional issues identified in later systemsDistance related issues Time zone issuesCultural issues

  • PortholesDesigned to connect PARC with EuroPARCForced technology change due to distanceShared digital network instead of dedicated analogPartially piggybacked on older local infrastructure (cameras)Lower bandwidthRepresents a turning point (2nd generation)Shift to broadcast instead of point-to-pointMuch lower fidelity Cruder pictures (see also earlier PolyScope system at EuroPARC)Frame per minute instead of 10 per sec

  • PortholesA key lesson from portholesPrevious efforts indicated better fidelity needed Attempt to better replicate face-to-face realityBut Portholes worked as well, and in some cases better than higher fidelity systems

    Aspects other than fidelityE.g., dropped connection implied by point-to-pointGo beyond replicating face-to-faceConcentrate on unique properties of media spaces

  • Third generation systemsStart to abstract away from direct communications (images)Go after individual issues separately (e.g., privacy and disruption)

    My favorites :-)Shadow view privacy preserving videoMumble filter audioSynthetic group photo

  • Shadow View TechniqueJust enough informationRough outlines and movementHide identity and other details

  • Video Clip...

  • Simple ImplementationCapture a static reference frameKeep previous frame and do frame-to-frame differencingCompare (8x8) grid cells, if above threshold change, darken cell in reference imageIf no change, lighten cell

  • Does it Work? Small study (n=20) of alternate designs [Zhao CSCW 98]

  • Detection of Presence

    Is there someone there?

    (100%) 320x240 100% 100% 100% 100%(100%) 80x60 100% 100% 100% 95%

    All provide pretty good info

  • Activity IdentificationPick from 4 activitiesUsing machine, reading, talking on phone, two person meeting

    (100%) 320x240 94% 94% 91% 62%(98%) 80x60 94% 91% 97% 62%

    Shadow view reduces but not removes

  • Person Identification

    Pick from 5 alternatives (guess = 20%)Fairly similar images (white males, ...)

    (92%) 320x240 78% 77% 62% 25%(80%) 80x60 80% 61% 54% 31%

    Shadow view hides most of this

  • Another Study Published in CSCW 00Boyle,Edwards, & Greenberg, The Effects of Filtered Video on Awareness and Privacy

    Slightly different displays, but similar conclusions: There is a compromise position that provides useful information, but preserves privacy

  • Other Settings for Different Effects

  • A Privacy Preserving, Low-Disturbance Shared Audio SystemWanted awareness properties of an open mic, but with privacy and reduced disruptionCreated a non-speech rendition of speech sounds

  • Implementation GoalsWant to leave eyes free for main taskUse audioWant to transmitWhos speakingAt least some prosody (e.g., is it a question?)Do not want to transmitContent (privacy issues)Any words (demand attention)Anything else that demands attention

  • ImplementationCreate an audio icon for each speakerRecord ~30 sec speech sampleRemove silenceRepeatedly mix sample over itself at random offsetsAdd noiseLow-pass filterNormalize volumeA crowd of one (too loud)

  • Implementation (cont.)Apply live volume envelope to audio icon Result: Preserves typical frequency distribCadence informationSpeaker id but no words, and can fade into background

  • Evaluation of Speaker RecognitionPilot study (n=12): mixed resultsVery bimodal response in subjectsMost (10/12) had solid (> 90%) recognitionBut 2/12 close to guessing

    Needs another round of design

  • Awareness of Groups: Synthetic Group PhotosGroup gestalt informationCompact but meaningful (123 people)Differential use of space

  • Presence: S : 100 100 100 100 95L : 100 100 100 100 100Actor Identification/filter: (% correct answers)S : 80 78 60 54 31L : 91 88 81 62 29Activity Recognition/filter: (% correct answers)S : 97 94 91 97 62L : 100 94 94 91 62Actor Identification/activity: (% correct answers for activities"computer", "phone", "magazine", "other", "empty room")S : 62 65 46 59 91L : 80 82 52 75 100Presence: S : 100 100 100 100 95L : 100 100 100 100 100Actor Identification/filter: (% correct answers)S : 80 78 60 54 31L : 91 88 81 62 29Activity Recognition/filter: (% correct answers)S : 97 94 91 97 62L : 100 94 94 91 62Actor Identification/activity: (% correct answers for activities"computer", "phone", "magazine", "other", "empty room")S : 62 65 46 59 91L : 80 82 52 75 100Presence: S : 100 100 100 100 95L : 100 100 100 100 100Actor Identification/filter: (% correct answers)S : 80 78 60 54 31L : 91 88 81 62 29Activity Recognition/filter: (% correct answers)S : 97 94 91 97 62L : 100 94 94 91 62Actor Identification/activity: (% correct answers for activities"computer", "phone", "magazine", "other", "empty room")S : 62 65 46 59 91L : 80 82 52 75 100Presence: S : 100 100 100 100 95L : 100 100 100 100 100Actor Identification/filter: (% correct answers)S : 80 78 60 54 31L : 91 88 81 62 29Activity Recognition/filter: (% correct answers)S : 97 94 91 97 62L : 100 94 94 91 62Actor Identification/activity: (% correct answers for activities"computer", "phone", "magazine", "other", "empty room")S : 62 65 46 59 91L : 80 82 52 75 100