Why the topic "Lecture Recording"?
lecture capture currently hot topic on campus
many different approached of several departments
with Apple's Xserve being discontinued, the need for new technology is emerging
Introduction of LKSC
recording since 70s, now fully automated HD video capture system, 16 spaces, centralized scheduling system starts and initiates the capture of sessions
running for 1.5 years now
seamless process for end-user, no recording costs involved
capturing about 12 sessions a day
massive amounts of storage, increasing each year
Motivation
give students the opportunity to follow lectures online, asap after lecture
few minutes after lecture students can access whole file (video and audio) online
Features
completely automated: to reduce manual work, no human resources for all rooms
streamline processing process, no human interaction needed while filming
camera presets: slide-optimized, whiteboard-optimized, etc.
feedback monitor on podium to check recording/capture
production control room to feed several cameras with manual editing/switching
analysis: track video analytics on file download, (content page with links, google analytics on it)
evaluation: student surveys, where: home vs. class, how: mobile vs. pc, when: how soon after lecture
Challenges
whiteboards capturing challenge: notes can be made in reservation system, giving presets for whiteboard capture: lighting, camera problems; lecturers are encouraged to use digital ink in 5 classrooms, file output as pdf
switching between whiteboard/powerpoint: not ideal, camera not optimized, lighting not great
sessions with faq/discussion: hand-held microphone to capture question, lecturers should ideally repeat questions for recording audio
manual editing for slides that should stay private: copyright, privacy etc.;
course coordinators/tas who check videos and point out what to cut
Future Developments of LKSC
manually control system, lecturers can start and stop the recordings
close captioning
mashing multiple videos online
social media/chatter components tagging into profile, link profiles and people appearing in videos to social profiles
Possible collaboration with LKSC
infrastructure of lksc is stable enough to take other files
LKSC needs 2 fiber feeds from any location could be connected
system not used at night: anybody interested is welcome to contact Trent Tanaka
capturing can work for all events in LKSC, just let them know at reservation
Teaching models
koller's work on chunking: using lecture time to do other stuff
charles prober (senior ass. dean for medical education): small videos (chunks), experiments and brainstorming on ideas how to use that
self-capture model for the faculty, many have self-capture space at home, providing lecturers with spaces and ressources to record own videos
different styles: experimental studies, focus groups with students
small recording booth in old office, to play with ideas: faculty member with khan present to ask questions
keynote/powerpoint recording with voice, supplemental information - "flipped" model of teaching, basic study material over video, interactive sessions with students in class
Daphne Koller: record lectures with capture systems, video-quiz based courses
General concerns with lecture recording
accessibility issues: ca has strict rules for disability act: any disabled person has the right to sue institution if it's not accessible to disabled people (production of media HAS TO BE close captioned), john foliot as contact for accessibility issues
enable students to search videos (through captioning)
some students don't go to class, just use videos to study
Specific Xserve issues
apple xserve storage: running out next year, continuation unclear: LKSC has 18-24 months before xserve gets replaced (maybe linux, open source), how much re-invention, integration???
possibly get rid of pipelines all together, capture cards: eliminate processing step, encode full resolution, capture card would allow to switch presets - more flexible, faster delivery of content
General workload issues
Camtasia used at radiology dpt., not enough staff for editing and recording
captioning services: captiontool.stanford.edu - 1.25 usd per minute 90% accurate, but manual post-production, labor-intensive
Developments outside of stanford?
national form group: how to move transition from face-to-face to e-learning
(please give more information)
harvard merging xserve to open source
berkeley will use opencast matterhorn, many feature sets
