Low Cost Narrow Bandwidth Synchronous Streaming Video From the Classroom to the Remote Students Desk

1
Low Cost Narrow Bandwidth Synchronous Streaming
Video From the Classroom tothe Remote Students
Desktop

• Eric Flower and Stacey Sawa
• University of Hawaii-West Oahu

flower_at_hawaii.edu ssawa_at_hawaii.edu StreamingVideoO
nTheNet.com
NMC 2005, Honolulu
2
Our Local Problem
3
Our Global Problem
4
Table of Contents

• Introduction/Purpose Slides 5 11
• System Design Slides 12 18
• Methodology Slides 19 22
• Results Slides 23 31
• Methodological Problems Slides 32 33
• Classroom Management Slides 34 36
• Secondary Observations Slides 37 43
• Future Research Slides 44 46

5
Internet Streaming

• An Internet data transfer technique that allows
  the user to see and hear audio and video files
  without lengthy download times the host or
  source streams small packets of information
  over the Internet to the user, who can access the
  content as it is received
• See sample video and a discussion of our project
  at
• http//socrates.uhwo.hawaii.edu/BusAd/Flower/video
  /uhthisweekcredits.html

6
Video Was Delivered Directlyto the Students
Desktop

• We did not broadcast to studios or computer labs
  where students would have to gather to
  participate in a site-to-site modeldelivery was
  directly to the students desktop at home or in
  an office

7
(No Transcript)
8
Purpose of the Study 1

• To identify levels of preparation and support
  (equipment, software, staffing, training,
  facilities, infrastructure, etc.) necessary to
  produce and distribute good quality narrow
  bandwidth (384Kbps or less) streaming video to
  students in their home or at their offices

9
Purpose of the Study 2

• To see if we could design a low cost Internet
  streaming video system without a large capital
  equipment investment, heavy infrastructure
  requirements, or a large technical support staff

10
Purpose of the Study 3

• To see if we could design a system on a cart to
  make it portable
• To see if classes could be broadcast from any
  campus classroom or conference room and not be
  tied to a studio, lab, or other special purpose
  facility

11
Purpose of the Study 4

• To make some preliminary determination on the
  effectiveness of synchronous streaming of
  classroom-based instruction when compared to the
  traditional classroom setting

12
Streaming Video Process
Capture with RealProducer
Distribute with RealServer, view with
RealPlayer/ RealOne
Encode with RealProducer
13
Software We Used

• We used apps from RealNetworks.com to encode and
  serve our classroom-based streaming video
• RealProducer to capture and encode video
• RealServer to distribute the video
• RealPlayer and RealOne Player to view streaming
  video

14
Video Encoding Rates
RealProducer can encode streaming video files in
any combination of the following rates
Connection Encoding Rate Delivery Rate
Dial-up 56Kbps 34Kbps Single ISDN
64Kbps 50Kbps Dual ISDN
128Kbps 100Kbps LAN/DSL/Cable
150Kbps 150Kbps LAN/DSL/Cable
256Kbps 225Kbps LAN/DSL/Cable
384Kbps 350Kbps LAN/DSL/Cable
512Kbps 450Kbps LAN/DSL/Cable
768Kbps 700Kbps
15
Hardware/Production System
Videocapture
Videotape backup
Videomonitor
Mixer
Audio capture
Splitter
Audio monitor
Encoding PC
16
Distribution System
Chat session
UH ITS StreamingServer
RealPlayer/RealOne Clients
17
Equipment Cart
Click here to see our equipment list
18
(No Transcript)
19
Methodology 1

• Using the streaming video system we designed and
  built, we taught four semesters of Computer
  Skills for Administrators with some students in
  the classroom and other students at home or in
  their offices participating synchronously
• Course is an elective with no pre-requisites
  offered in the Professional Studies Division

20
Methodology 2

• We taught the class four times between 2001 and
  2004
• Over that period, there were 41 students in the
  classroom and 29 online
• Each semester, both groups had the same class
  presentations, readings, and assignments, wrote
  the same reports, took the same tests, and worked
  on similar projects

21
Methodology 3

• Both groups could watch the archived class video
  files
• In a retrospective study, we compared student
  course grade scores for each group (in-class vs.
  remote)
• Used a t-test for unequal variances
• Hou1 u2(Null hypothesis means of the two
  groups are equal)
• Hau1 ltgt u2(Alternative hypothesis means of the
  two groups are not equal)

22
Grade Score Components

• Testing throughout the semester 45
• Final exam 20
• Class participation/Quality circle
  participation 10
• Group presentation/Group presentation
  contribution 15
• Critical review of Visions How Science Will
  Revolutionize the 21st Century by Michio Kaku
  10

23
Results 1

• We found that off-the-shelf hardware and software
  were adequate to produce good quality narrow
  bandwidth (384Kbps or less) streaming video
• Significant investments in broadcast studios may
  not be necessary where high production values are
  not an essential part of the instructional process

24
Encoding Rates and File Size

• Triple stream video files encoded at 56Kbps,
  150Kbps, and 256Kbps use 3.9 megabytes of storage
  per minute
• Stream to the viewer at the highest reliable
  connection, typically 34, 150, or 225Kbps
• This multistream provides a good balance of sound
  and picture quality and targets dial-up, LANs,
  and broadband
• In 2004 we upped the top rate to 384 Kbps encoded
  and 350Kbps received

25
Connections LAN/Broadband

• Local Area Network (LAN) and broadband
  connections (cable, DSL) are clearly better
• Video encoded at 256Kbps is received at 225Kbps
• Video encoded at 384Kbps is received at 350Kbps
• Sound and video are both good
• Acceptable for motion, detail, and close-up work

26
Connections 56K Dial-up 1

• Streaming video encoded for 56Kbps dial-up modems
  is received at 34Kbps
• Sound is good
• Picture is not nearly as good as higher speed
  streams
• Video is choppy and there may be frequent
  rebuffering if there is network congestion
• More like a slide show than a motion picture

27
Connections 56K Dial-up 2

• May not be acceptable for classes or demos
  involving motion, detail, or close-up work
• May be acceptable for discussion-based classes
  with little movement
• May be acceptable for large image, large font
  PowerPoint presentations with little or no
  animation

28
Results 2

• Marginal operating costs were relatively small
  and limited mainly to the salary of the
  technician operating the camera and performing
  post-production work
• It was possible to stream from virtually any
  classroom or conference room on campus with the
  mobile cart

29
Results 3

• To determine effectiveness of synchronous
  classroom-based streaming video, we performed a
  t-test for unequal variances on the course grade
  scores for the two groups
• Hou1 u2(Null hypothesis means of the two
  groups are equal)
• Hau1 ltgt u2(Alternative hypothesis means of the
  two groups are not equal)

30
Results 4

• Analysis of the student course grade scores for
  each group (in-class vs. remote) shows a strong
  similarity of the means
• Using the t-test for unequal variances, the
  p-value was 0.84 this provides a strong
  statistical conclusion that the means are not
  different

31
Results 5

• Low cost narrow bandwidth synchronous streaming
  of classroom-based instruction appears to be as
  effective as the traditional classroom
  environment in this instance

32
Methodological/Data Problems 1

• Groups are not random
• Groups are self-selected by enrollment in a
  classroom or online section
• Neighbor Island students have no choice they
  must enroll in an online section
• There is no information about the students aside
  from their group and their final course grade
  score

33
Methodological/Data Problems 2

• There is no information about computing ability
  or knowledge at the beginning of the class
• There were no pre- and post- tests
• Not blind instructor knows who is in each group

34
Classroom Management 1

• Instructors teaching style and course content
  must be amenable to this delivery method if it is
  to be successful
• Instructors dont need to change their classroom
  methods very much to succeed with Internet
  streaming video broadcasting long periods of
  training appear to be unnecessary

35
Classroom Management 2

• Preparation, rehearsal, and timely distribution
  of supporting material are critical to successful
  streaming video presentations

36
Classroom Management 3

• Faculty workload increases in this environment,
  but not nearly as much as in an asynchronous text
  or multimedia-based online teaching environment
• Time spent on preparation and production is less
  with streaming video, though post-production work
  may be more, depending on the faculty members
  decisions on how much to do after class

37
Secondary Observations 1

• Students watched archived classes they missed due
  to illness, travel, work, etc.
• Students reviewed or intensively studied classes
  or portions they did not understand when the
  material was originally presented

38
Secondary Observations 2

• Instructor could review class files to correct
  errors in presentations or to add supplementary
  material to a web-based errata file
• Instructor could review class files to examine
  their methods for teaching effectiveness and make
  improvements where necessary

39
Secondary Observations 3

• Class files could be reviewed for teaching
  effectiveness when making decisions relating to
  contract renewal, tenure, post-tenure review, and
  professional development

40
Secondary Observations 4

• Synchronized Multimedia Integration Language
  (.smil) files may be used to create, or
  contribute to, learning modules of any length to
  improve instruction
• .SMIL file-based learning modules may be used to
  make material available to other classes or groups

41
Secondary Observations 5

• Students want more classes offered via streaming
  video if they have access to fast connections
  like Oceanic Cables RoadRunner service

42
Secondary Observations 6

• System was reliable
• There were very few incidents of hardware
  failure, software failure, links to server, or
  server failure
• Slow connections, network congestion, and lack of
  Quality of Service (issues of network packet
  priority) were problems for some dial-up students

43
Secondary Observations 7

• Network bandwidth is the critical variable in
  streaming video viewer satisfaction
• Faster connections support notably better viewer
  experiences
• No amount of preparation can overcome problems
  associated with a slow connection or network
  congestion
• Viewers will blame you or the technologyif their
  picture is bad

44
Future Research 1

• More rigorous study and analysis of
  classroom-based synchronous streaming video needs
  to be performed to confirm these preliminary
  findings

45
Future Research 2

• Students thought viewing archived streaming video
  class files on demand (asynchronously) could be
  as effective from an instructional viewpoint as
  participating in the live class (synchronously),
  provided they could then interact with the
  instructor via e-mail, chat, or videoconferencing
• Asynchronous delivery with a scheduled online
  class meeting component could increase access

46
Future Research 3

• Archived files could be viewed when network
  traffic was low, or, archived files could be
  distributed on CDs which would eliminate delivery
  problems associated with slow network connections
  or network congestionthis could be a boon to
  less developed areas with weak infrastructures

47
Thanks for reading!

• Eric Flower and Stacey Sawa
• University of Hawaii-West Oahu