Supplementary material to “Seminar 2.0: Learning With Skype and Video Podcasts”
Published 28 April 2009
Paul Wessel, Tiffany Anderson, Regan Austin, Asdis Benediktsdottir, Michael T. Chandler, Michaela M. Conley, Seung-Sep Kim, Robert L. Michaud, M. Elise Rumpf, Jonathan D. Sleeper, and Jonathan R. Weiss, Department of Geology and Geophysics, School of Ocean and Earth Science and Technology, University of Hawai’i at Manoa, Honolulu
Citation:
Wessel, P., et al. (2009), Seminar 2.0: Learning With Skype and Video Podcasts, Eos Trans. AGU, 90(17), 145–147. [Full Article (pdf)]
While there are many ways to accomplish the same end result, we will here report on the setup we evolved during the seminar. The steps reproduced here will no doubt be modified as we gain more experience with remote seminars, and perhaps new software will simplify some of the processing steps. We will maintain an up-to-date recipe at this website.
Hardware considerations
Anyone attending an AGU meeting where PowerPoint presentations developed on a Mac were viewed on Windows (or vice versa) knows that pitfalls are plentiful. In particular, presentations with animations will refuse to play due to lack of installed video codecs, and prior assumptions of font availability will inevitably turn out to have been foolishly optimistic. We therefore decided to mimic the hardware and presentation software used by the speakers. Of the eight participants, six prepared their presentations using the Mac 2004 version of PowerPoint; one used KeyNote (Mac only), while one used the Windows 2003 version of PowerPoint. Upon receipt of their files we ran their presentation locally, using the same setup (Windows or Mac, PowerPoint or KeyNote) as they did. The Mac running Skype had a wired Ethernet connection to ensure high throughput.
The two video projectors could vary from session to session, depending on what was available in the department. The unit projecting the videoconference can be very basic as the video signal is only 640x480 pixels when connection is fast but can drop to half resolution if there is a bandwidth bottleneck. However, a quality projector should be used for the PowerPoint presentation to render detailed scientific illustrations and maps at the highest resolution possible.
For our experiment we used a cheap set of external computer speakers. While this adequately amplified the speaker’s narration we did sometimes experience a feedback-loop with our microphone picking up the speaker’s voice which was then played back on the speaker’s computer and picked up by his microphone, and so on. We mitigated this by carefully positioning the speakers relative to our microphone, asking the speakers to lower their own playback volume (or use a headset), and occasionally muting our microphone during the talk.
The external web-camera was simply an iSight camera that we duct-taped to a tripod. Any webcam will do, but note Skype has certain requirements that your camera must meet to enable high-quality video. Having the camera on a tripod made it easier to pan from the teacher (when introducing the speaker) to viewing the students (when engaged in discussion with the speaker). During the talk itself the camera pointed to our projection screen so that the speaker could see which slide we were watching; this helped in synchronizing the presentation as slides were advanced locally on queue.
Software considerations
Not counting PowerPoint or KeyNote, the only software required to run the seminar was Skype, a free download. However, to record the conferences we purchased a USD 15 plug-in for the Mac version of Skype (Call Recorder by Ecamm Network) that allowed the videoconference to be recorded (similar plug-ins exist for Windows and Linux). Call Recorder created a single QuickTime movie of the entire conference. Conveniently, the recordings saved the two voice tracks separately so that we could mute or remove our track that usually only contained noises such as people coughing, chairs screeching, or audience laughing following humorous remarks by the speaker.
To produce podcasts we purchased a license for Apple’s QuickTime Pro (a USD 30 expense). This license activates the authoring capabilities of the free QuickTime program and is available for both OS X and Windows. While there is much software available for producing video podcasts, our experience was that QuickTime Pro was the simplest tool that could handle the relatively simple production we set out to perform, as detailed below. While we have no experience with Linux video editing we believe similar capabilities are available. We used QuickTime Pro for all audio/video-editing tests below, such as handling all the video cut, paste, assembly, tagging, and exporting to a variety of formats, including iTunes-compatible MPEG-4 with H.264 encoding.
Pre-Seminar Preparations
Prior to the seminar we emailed various instructions to help the speaker prepare for some peculiarities of a Skype talk. This included items such as numbering the slides, leaving space in the upper-right corner for a picture-in-picture video of the speaker (podcast only), and adding arrows to highlight items in lieu of using a laser pointer. We would typically test the videoconference connection some days prior to the actual seminar in order to troubleshoot the setup and discuss the procedures. The speaker would send us their presentation at least 24 hours prior to the talk, and the student responsible for running the slideshow would examine the presentation to become familiar with the flow of the slide show and to determine that animations (if present) played properly. This familiarity was helpful when the speaker forgot to say “Next Slide” and the student would have to make that judgment based on context.
Seminar Operations
Once we could enter the classroom we frantically set up all the equipment, typically in ~10 minutes. We would then connect to the speaker about 10 minutes prior to the start of their talk. We would let our camera pan to give the speaker an idea of the audience. The recording of the videoconference was started before the introduction of the speaker and ended once the talk had completed. We did not record the question/answer session when students would walk up to the camera to ask their questions. Once the seminar was over the recording was placed on a central server and a group of three students would begin the post-production.
Podcast Production
After each seminar a rotating group of three students would produce a video from the slideshow. This was greatly simplified once we realized that by running the speaker’s live presentation in “Rehearse mode” we could capture the variable slide durations while viewing the talk. When the talk completed we saved the durations to a copy of the PowerPoint file. It then became trivial to produce a video of the slideshow since PowerPoint has a “Make movie” option that uses the recorded slide durations to export a QuickTime movie. Anticipating a need to make podcasts and even DVDs at different resolutions we chose to export video at 1440x1080 resolution (HD1080p at 3:4 aspect ratio). The resulting movie clip has the correct duration and shows all slides. Any slides that contained animations needed special care, since PowerPoint would only play an animation clip once when exporting the movie rather than loop repeatedly to fit the slide duration. We ended up temporarily removing animations from those slides prior to exporting the movie; we then later added in the various animation clips, as explained below.
The next step was to process the videoconference recording, such as trimming the recording to fit the start and duration of the slide slow. We could then overlay the two videos and change the size and position of the speaker’s video to fit in the upper right corner of the slide show (we had told the speakers to make room for this in their slide layout). If there were animations skipped earlier we would next paste those in as well, and trim the total length of the animation to fit the recorded slide time. For identification, we added a simple 20-second video clip with seminar and talk titles to the beginning and a similar 20-second credit video to the end, both set to the music “Tectonic Plate Shake” by Gary Shutt (Palisade Records) and created in KeyNote. We also added appropriate metadata tags (author, title, copyright, year, etc) so that podcasts uploaded to iTunes University would have the correct information. Finally, we would boost the speaker volume (if needed) and then export the podcasts as MPEG-4 Quicktime movies in two resolutions: iPod (640x480) and HD (720p). Both of these were uploaded to iTunes University under the U Hawaii section; they were also posted on our local website so that users unable or unwilling to use iTunes could access them.
Given our lack of prior podcast experience and the difficulty of scheduling common time for the students outside class hours, we typically were able to publish the video podcast about one week after the seminar (actual workload was in the 1–3 hour range).
Finally, to show our appreciation to the guest speakers we provided each with a regular DVD-R that, in addition to a DVD-playable version, contained both miniature and high-definition QuickTime versions for their use, as well as a copy of their original PowerPoint materials.
Technical Glitches
For the videoconferance with David Sandwell we used iChat and the Conference Recorder iChat plugin from Ecamm. However, during the hour-long session the video feed froze three times and the only way forward was to restart the videoconference and tell Dave to start from the current slide; the resulting three video recordings then needed splicing and minor edits before podcast assembly could begin. While the origin of the problem remains unsolved we speculate that the iChat plugin was less robust than the more widely used Skype plugin. We also had KeyNote freeze during Müller’s presentation, and this required a reboot to correct. This occurred on a Mac Mini permanently used in our large conference room (used to accommodate Sydney time) and the problem was later traced to the failure to install many required OS X and KeyNote updates. Since it was nobody’s personal computer, nobody had felt responsible to do the updates. Other glitches were minor: there was the occasional video stuttering and garbled audio, but only for a few seconds during an hour-long talk.
Future improvements
We will continue to experiment with the setup to make it as simple and robust as possible. Since speaker audio quality is more important than the video we need to improve the speaker/microphone arrangement. The podcast production is now fairly straightforward after learning how to do it by trial-and-error; we are not aiming for a professional-quality production so simplicity and speed of production are key considerations. Given the experience we feel a final podcast like the ones published should take less than one hour to prepare, unless there is major editing (splicing, many animations, etc).
Final thoughts
In addition to running these seminar talks we have routinely used Skype or iChat during student thesis meetings to bring in one or more external thesis committee members. While Skype allows a single video connection, Apple’s iChat allows up to four people to participate in a multi-person chat, which is more than enough for most small impromptu committee meetings. With no need to record the setup simplifies considerably. For Windows a similar capability can be achieved using ooVoo.
