Challenging streaming video in E-learning.

The success of the large scale inter-university IT course in Flanders in 1998 – 2000, using live ISDNvideoconferencing, persuaded the Institute for Continuing Education (IVPV) at the Gent University to organize a second edition, taking into account the remarks of both students and supporting industrial partners. Taking up current ‘best practices’, a renewed course framework based on streaming video on CDs has been set up. Considering the unsatisfactory performances of available commercial products, an inhouse system for streaming video has been developed, with emphasis on animation and interaction. Following a successful trajectory, the new student evaluation showed aims were largely achieved. Currently the IVPV started a new project to integrate the learning materials with a global open management system based on an e-learning framework. At the same time, contacts are being made with countries where student’s mobility is restricted in order to make all produced learning material available to those countries. Setting out the milestones. In [1] a large scale continuing education programme in IT organised in Flanders (Belgium) from 1998 to 2000, using massive ISDN-videoconferencing at 384 kbps (6 ISDN channels) has been described. The target of this programme was to train a new generation of IT-professionals by the year 2000. It was jointly organised by the 4 main Flemish universities (Ghent, Leuven, Brussels and Antwerp). In twelve sites, spread around Flanders, for 417 hours, two times a week in the evening hours, 3 hours of interactive videoconferencing were set up, complemented by 87 hours of hands-on exercises. . As much as 1145 participants subscribed to this massive continuing education programme of 206 213 [participants*hours], 80 of which passed the necessary assessments to obtain a full academic certificate and 317 obtained a modular certificate. The course was largely self-supporting. In the follow-up of this course, a scientifically based evaluation of the complete educational programme among the participants was carried out to gather information about the effects of the videoconferencing framework as such. More detailed data can be found in [1], but the mainstream results were as follows: • the well known obstacles of videoconferencing remain in effect: too static, little interaction, no ‘classroom feeling’,.. • the limited quality of the image together with a good sound quality was not seen as a real problem; • the real-live animation of the videoconferencing (animated powerpoint slides + mouse interaction) was considered as highly beneficial to enhance the learning behaviour; • the possibility of bridging time and space was seen as a main advantage; • the accompanying electronic discussion forum (Majordomo) was appreciated very much. All this resulted into an overall satisfaction degree of 79%. Following those assessment results, a fundamental discussion with main representatives from the industry in Flanders was engaged to set out the framework for the second issue of the programme (20002001). The output was that the main idea of the videoconferencing was all right, but that the necessity for people, working in industry under high pressure, to be lively present at fixed times was impracticable, and that a swing to more asynchronous learning was highly desirable. Therefore a modified framework for the course was designed as follows: • the videoconferences have to be digitally recorded on CD or DVD in a format which: − keeps in any case the animation of the slides (building up slides / mouse movements); − enhances possibly the image quality (audio quality was satisfactory); − maximizes the amount of lectures per disc; i.e. streaming video on CD; • the interactive electronic discussion lists have to be extended with more features; • a minimal live interaction with teachers / assistants should be safeguarded; • participants should be tightly kept into a study rhythm in order to avoid ‘free lunch’ behaviour. The Streaming Video framework. The lecturer typically prepares his/her lecture in Powerpoint (or similar: pdf,…), and breaks it down in pieces (‘paragraphs’) of some 5 to 15 minutes, depending on the topic at hand. Thereby he/she utilizes the whole scale of animations (slides transitions,…) available, albeit obeying some basic rules about fonts, colours, picture sizes,… He/she gives his/her lecture (cut in ‘paragraphs’ as said) live in the multimedia studio (eventually for a ‘live audience’), whereby he/she freely uses mouse / cursor movements to clarify his/her lecture. The presentation is captured from the computer output (vga…) itself together with an image by a professional video camera directed to his/her head. The studio technician mixes in real-time the computer output with the image of the teacher (being it in full scale, being it as a ‘stamp’ picture-in-picture) resulting into one digital signal recorded onto a digital Dvcam tape. One of the important aspects of this recording system is that – starting from a lecturer with slides at hand (the ones he normally uses in old-style lectures) the extra workload on the lecturer is kept to a very strict and acceptable minimum. We postulated beforehand that the preparation time may only be as long as the lecture time itself, i.e. the total time may be maximum the double of the lecture time. Post-processing is done as follows: • the video-fragments are read from the Dvcam, via I-Link (Firewire), into a MPEG-2 video editing station for real-time compression to MPEG-2, 4 Mbps (IPPP). • these MPEG-2 video fragments can be edited if necessary; normally this is never done; • the finished video fragments are compressed to the RealNetworks video format (chosen because of its neutral computer independent platform); thereby the choice of optimal conversion parameters is very critical: i.e. they largely determine the quality of the resulting sound and image, and consequently indirectly the amount of space occupied on the CD; • the RealNetworks video files are incorporated as multimedia objects in a standard HTML document, together with the other learning material as one CD-R master copy: − the powerpoint files as such (or pdf’s thereof), so the student can visualise it eventually in full screen resolution or print it out; − per video ‘paragraph’ a short description of the contents; − the background material: course texts, URL-pointers, book references, exercises, solutions, application software,… • the net result (for the IT course) was an image with the following characteristics: 512 kbps / 12 fps / 512 * 348 pixels / 3 (wall)hours of lecture per CD; • the master CD-R is multiplicated and labelling is done in a CD multiplication robot. Compared with off-the-shelf commercial products for streaming video, the main difference is that the full animation both with slide transitions and cursor movement are maintained on the CD, giving huge educational benefits, since the student’s attention is constantly kept during the lecture: the animation is really the glue between the lecturer’s voice and the image. Although originally not intended, it turned out that some companies put a copy of the CDs on an internal streaming server (provided the bandwidth of their intranet was sufficient). The overall course structure. The IT-course was composed of 5 distinct modules: A: computer architecture / P1: programming / P2: information and programming structures / O1: analysis and design / O2: object-oriented software development. The layout of each of the 5 modules was identical: 1. one videoconference session: the purpose was to get the students acquainted with the physiognomy of the lecturer, and to create a practical opportunity to distribute the CD-sets to the students; 2. ‘home study’: a number of weeks for asynchronous learning by the student; 3. lab exercises: on Saturdays (to avoid traffic jams) students came to the university to perform hands-on training on PCs under the guidance of qualified university assistants; 4. one live feedback session (also on Saturday’s): were the students could forward final questions / remarks to the professors; 5. time for project work and final preparation time for the examination; 6. the examination: the form depended somewhat on the module: some were written exams, some were PC-exercises, some were projects. The course language was English. The table below summarizes the course load as calculated. Module Theory Videoconf. Home Lab exercises Feedback Project / study Total hours Number of weeks A 3 18 6 3 18 48 4 P1 3 18 12 3 42 78 7 P2 3 18 9 3 42 75 7 O1 3 12 9 3 30 57 5 O2 6 29 15 3 70 123 11 Total 18 95 51 15 202 381 34 i.e. 11 hrs/ week The course structure was furthermore complemented with enhanced discussion lists (Agora): one per module, and one global for organisational aspects. This platform enabled interactive discussions among students and with the professors / assistants. It was largely used during the course: e.g. for module A, some 197 messages / answers were posted. For each module an assistant professor was assigned as an online coach, who was continuously available to the participants in a module, but only for a fixed amount of weeks: after the predetermined period the list was closed. This mechanism intended to put some strain on the students to stimulate them studying the corresponding module during the period as scheduled, at the same time giving them enough freedom to do their own timemanagement. The discussion platforms and the way they were implemented, were thoroughly discussed beforehand with the industrial partners, and was also seen as a substitution of the social control among students during in-house videoconferences (as in the previous course issue). The course was subscribed by 206 students (approximately 150 students per module), 70 of which finally received the academic certificate after successful examinations and project work. Organisationally the course was self-supporting.