Monthly Archives: February 2011

Classes*v2: Media Gallery Pilot

Posted on by
Reply

This week Associate Director David Hirsch and Senior Programmer Peter Liu from CMI2 and Instructional Technologist Matt Regan and Instructional Support Manager Gloria Hardman from ITG spoke about the pilot program they are currently running on the Classes*v2 called, “Media Gallery.”  Through Media Gallery, professors, students and teaching fellows are able to upload, organize and present video and audio content on their Classes*v2 course sites.  Currently there are about fifteen courses involved in the pilot program, which has begun to reveal some of the benefits (and bugs) of the system. Tracking the pros and cons of the system will allow CMI2 and ITG to better adapt the program to fit Yale’s teaching and learning needs.

Kaltura, the media hosting platform that is being used as the platform for Media Gallery, is an exceptional fit for Yale’s needs for two reasons in particular: the program is open source (providing flexibility for Yale-tailored modifications) and all media is hosted on a dedicated media server (ie not a Yale server).

Basic functions of Media Gallery are easy enough for even the novice user. To upload material off of one’s personal computer, only a few simple steps need to be followed.  The first step is to access the course site by logging in classesv2.yale.edu.  Instructors or teaching fellows can then select a file, which will be uploaded to the Kaltura server.  Once on the Kaltura server, the media files will be transcoded into standard formatting before being stored as links inside the course site library.  Any metadata chosen to be associated with the content—title, description, tags—will appear in the course site library with the link to the media file.  Tags in particular are especially beneficial as course instructors begin to organize the media files inside the library by creating “collections.”

To explain the capacities and detractors identified during the pilot, Matt oscillated between his real account (the Yale Matt) and account (his alter ego with special powers).  Gloria, in turn, channeled Julia Child, whose included “files already sitting on the drive, ready for upload.”

Here are the pros and cons of the program.

PROS

Most video and audio formats are supported, with the notable exception of MP4 audio files. Media files can be easily incorporated into existing functions of Classes*v2, such as tests, quizzes, and even resources. Permission to use the files is done through Classes*v2, allowing for simple access control.  While instructors can download files and/or copy the files’ embed codes, students cannot.  Like with material currently being uploaded under “Resources,” responsibility lies with the individual professor to abide by copyright terms and restrictions of fair use.

There is a great deal of growth potential within the existing program. Not only does the open source nature of the program enable tailored modifications, but there are programs possible through Kaltura that are not currently implemented but might be incorporated in the future.  For insantce, remixing and clip creation through a web-based media editor is possible through Kaltura.  This function would allow film studies professors, wanting to teach students about clip order, to upload a number of clips, encouraging students to play with order to consider its effects.  Similarly, image uploading is current disabled, though possible.

CONS

The biggest benefit of the pilot program is that the testing process enables many of the program’s bugs to be rendered obsolete for future users.  For instance, there is a 2G upload limit, which makes Kaltura more suitable for shorter video clips than full-length films.  Communication with Kaltura, however, indicates that the upload limit is likely to change soon.  Also, the default option for media files is currently “hidden.”  That means, occasionally, uploaded media content will be hidden from student’s page views without the instructor realizing (as the content is not hidden from the instructor’s page).  Lastly, there is no “drag and drop” feature for the ordering of files within collections.  Instructors who are concerned with order should be conscious of the order in which files are moved into collections files.

THINGS FOR USERS TO KEEP IN MIND

  • Users, not courses, control media files.  Media content is stored in association with the user who uploads it and then shared, by that user, with a course.  Once the course has been completed, the media files stay with the individual who uploaded them to be reused for future classes.  This does mean, however, that Media files included in a course by a TF will need to be re-uploaded by professors the second time the course is taught.  Similarly, language courses using the same video content for multiple courses controlled by multiple professors cannot share the video content, but must instead upload videos separately.
  • The more files uploaded, the more important metadata becomes.  “My Content,” the automatic repository for media files, only shows ten clips at a time.  While there is a search function, it’s relatively useless without meaningful tags associated with the content.  Unfortunately, international characters are not currently supported in the metadata fields.
  • Transcoding off of the Kaltura server can take a little while.  Instructors can’t expect to upload a file to have it available for class five minutes later, unless the file is very small.  That being said, most files are likely to take less than an hour to transcode, so any files uploaded the night before will certainly be readily available for class the next day.  The transcoding, Gloria reminded us, “all takes place in Kaltura-land” so if instructors find themselves frustrated by these delays, they’re welcome to take it up with Kaltura.  Alternatively, this also means that once the instructor has sent the file he or she is welcome to “go take a nap.”

After Matt and Gloria concluded their demo of the pilot, Martin Kane, Special Assistant for Spanish for Health Care Professionals at the CLS, closed the talk with a demo. His course, “Español for Healthcare Professionals,” is using the webcam capacities to record videos of students doing portions of a patient history in Spanish.

Want more? View their powerpoint presentation.

For full coverage of this session, please click the video below (note a slight delay upon initial playback):

Donatello – A Browser-Based Image Annotation Tool

Posted on by
Reply

Marq Jefferson '11 demonstrates Donatello.Instructional Innovation Intern (i3) Marq Jefferson and Ken Panko of ITG lead our first TwTT of the new year, discussing an image annotation tool that Marq created called Donatello.  Marq began the talk by telling us a little bit about the history of Donatello.  The program was developed in response to archeologist professor Harvey Weiss’s desire to show building plans, excavation field notes and images associated with the Leilan Akkadian palace excavations in Syria.   To create a tool that would do this without reinventing the wheel or requiring software like Flash, Marq built the browser-based tool using existing technology: a javascript library called Raphaël.  With Donatello, administrators can create shapes in an array of colors with varying opacities on top of the image to which they want to add more information using the square, circle, or polygonal tool.

The plan of Prof. Harvey Weiss' Tell Leilan dig visualized in Donatello.Donatello keeps all of the data on Yale servers, and is easily customizable to specific needs.  The instance created for Professor Weiss’s class shows a map of the dig, with annotations for different parts of the site, both in hoover text and displayed on the side bar with a double-click.  Professor Weiss wanted the tool to be a useful reference tool, one that allowed schematics of the site to look more realized by associating them with photographs.

The plan of Prof. Harvey Weiss' Tell Leilan dig visualized in Donatello.After explaining how Donatello had come about, Marq spoke about other instances of the protype, speaking to the evolution of the tool.  A current beta version of the project annotates a painting by Raphael, featuring hot-spots on the image which provide a deeper comprehension of what’s being seen.  Sidebar data is shown through double-clicks, and can be edited through html content – anything from a youtube video to plain text.  In addition to the data that can be shown in the mouse-over and the sidebar, Donatello allows an administrator of the page to link images to outside links (think: Wikipedia) or other images (such as different images of the same person being depicted).   Marq explained that webpages could also be constructed through the site, and linked in to the network of data that an instance of Donatello would contain.

Raphael's "The School of Athens" visualized in Donatello.Marq closed by explaining that one of his motivations for working on the project – beyond assisting Harvey Weiss – was to provide a tool that would help students in classes.  He felt that the most interesting and memorable classes in which he had participated were those with lots of media; a graphic view helped make information easier to ingest.

Audience members were curious to know more about access to the program. Marq explained that professors could request new instances as they’ve reached a stable version – though the program is “always sort of in development.”  [Ken noted that those interested in pursuing use of the program would need to keep in mind that they would be early adopters of the program and should go in with those expectations.]  With regards to privileges, as it stands, the program can be used from any web browser, but only the administrator can add content – although use of the site could alternatively be contained through Net ID restrictions.

Ken asked the audience if they had any suggestions, or noticed any functionalities that appeared to be missing.  Suggestions included enabling freehand drawing of shapes (or an imported polygonal tool set), adding audio capabilities to the site (a project Marq has already begun work on), providing multi-lingual character support (which lead to a brief demo of Marq’s other project, Rubicon), the capacity to thread together multiple background images and similarly, the capacity to annotate one image on multiple occasions.

Wrapping up the session, Ken wanted to draw attention to the fact that Donatello – and Marq! – are excellent examples of why the i3 program is such a success and awesome opportunity.  Anyone with project suggestions (or student suggestions) should get in touch with Robin Ladouceur who is currently in charge of the internship program.

For full coverage of this session, please click the video below (note a slight delay upon initial playback).