openwebworks

Sunday, October 10, 2010

I arrived in New York on Thursday, September 30th to attend the OVC which began the following morning. This is the 2nd annual OVC. I attended the 2009 conference as well. OVC is organized by the Mozilla Foundation to promote development of open source technologies and best practices in open video. The OVC mantra is “open source, open standards, open formats, and open access.” Mozilla (http://www.mozilla.org/) is a huge and credible player in this space, as the developer of the Firefox browser and other open source software. The OVC was attended by about 1000 people from all over the world. You can get more insight etc from the Twitter feed which is #OVC10.

The following notes present the main topics and takeaways from my perspective. I’ll throw in some personal impressions as well. Above all, it seems to me that many great ideas about tools and methods were only concepts at last year’s OVC. This year they are working products, albeit in some cases rough around the edges.

Some of the following notes will be somewhat technical, but I try to present what is significant about the projects covered from a user’s perspective. The basic themes of the OVC were:

• HTML5
• Openness and how to facilitate open access
• Open Source tools in development and practice
• Media archives, metadata, management, and storage
• Remixing media as a focus of culture, education, and digital citizenship

The full schedule of the OVC is here: http://www.openvideoconference.org/agenda/

Friday, October 1, 2010

HTML5 player showcase/How to Build an HTML5 player - (10:30 AM - 11:45 AM)

Presenters:
Chris Blizzard — Mozilla
Philippe Le Hegaret — W3C
Steve Heffernan — VideoJS and Zencoder
Jeroen Wijering — LongTail Video

Steve Heffernan from http://videojs.com presented details on video.js, an open source HTML5-based video player. HTML5 is the newest format for web pages, and includes many improvements over HTML4 including native support for playing audio and video in HTML5-compliant browsers. HTML4 relied upon external players to play media files, like RealPlayer, Windows Media Player, QuickTime Player, various Flash-based players, etc. This has been a real nightmare for web developers, since you have to account for different users and systems using different formats and players. In theory, HTML5 solves this by providing a standard way to play media files directly in the browser, by using a or tag which points to the media file you want to play. In theory. The problem is, the designers of the different browsers (Firefox, Safari, Internet Explorer, Opera, etc.) implement HTML5 in slightly different ways, just as they did with HTML4. The situation is vastly improved, but not yet as simple as we (website developers and users) want.

So video.js, like other HTML5 player code solutions, addresses the remaining cross-browser inconsistencies by providing “fall back” code that plays the media file in case the browser doesn’t support the recommended HTML5 method. This requires some JavaScript and some html code providing several options, including alternative media files in different formats.

JWPlayer (http://www.longtailvideo.com/) takes this same approach. We use JWPlayer on the Illinois Public Media website, although we have yet to implement the latest version which is 5.3. The developer of JWPlayer 5.3 presented his solution for HTML5 incompatibilities in various browsers. JWPlayer provides some very slick skinning and an API for injecting a wide range of variables into the player, creating playlists, and providing other advanced functionalities and features.

Philippe Le Hegaret from the W3C then did a live 20-minute demo of how to build an HTML5 player with skinnable controls using JavaScript and Scalable Vector Graphics (SVG). SVG is another W3C standard (technically, it’s a “recommendation” just like html itself) for creating flexible and scalable graphics based on pixel coordinates instead of actual images. With SVG you can create elaborate animations in the browser without using Flash. Philippe Le Hegaret demonstrated building a media player using inkscape (http://inkscape.org/), an open source SVG authoring tool. In 20 minutes, very impressive.

Public Spaces, Private Infrastructure - (11:30 AM - 12:15 PM)

Presenters:
Ethan Zuckerman - Co-founder, Global Voices, Senior Fellow at Harvard Berkman Center
Lawrence Liang - Lawyer, public intellectual, Alternative Law Forum Bangalore
Trebor Scholz - The New School

The HTML5 player session went late (like most OVC sessions ☹) so I caught only the end of this session. The discussion was about how we use “the cloud” without thinking about it, e.g. when we upload a video to YouTube, post our pictures on Flickr, or conduct our social interactions on Facebook. We don’t own these services, but we entrust them with our information and media. They seem to be public spaces, but their owners can change their Terms of Service any time they like with no recourse. This has serious implications for privacy, and for preserving access to public and private media objects. And most casual users of the Internet don’t even think about it.

I spoke with a number of people about this dilemma during the OVC. The idea of a “public cloud” seems to resonate with lots of us, wherein non-profit institutions could provide cloud services including long-term storage, with a commitment to trust, openness, and sustainability. Universities and the supercomputing centers around the U.S. could play a key role in providing this public infrastructure. I know there is some interest on their part, especially since NSF funding for the computer centers has diminished in recent years. It seems to me this is a big idea worth exploring, but it will require vision and leadership to build a consensus and get something actually done.

One example illustrates the need. WGBH/Boston is far ahead of most public TV stations in building online media archives. Peter Pinch manages WGBH’s Open Vault project, which provides thousands of open media archives as curriculum materials. Peter told me they’re using Amazon Web Services to encode and host their media collections. The advantages: it’s inexpensive, easy to set up, and they don’t have to maintain the servers and infrastructure. The big disadvantage: if Amazon decides to change their policy, or 20 years from now goes out of business, WGBH’s media collections could simply disappear. They would have to recreate everything from the original source files, assuming WGBH continues to manage them.

Another possibility is simply using the Internet Archive. I spoke with Tracey Jaquith, Web Engineer & Data Archivist at the IA, about their capacities for hosting large media collections. She says the IA is committed and funded to provide unlimited, permanent storage of both access files (i.e. streaming media hosting) and original source files. The IA can provide various tools for uploading and managing content, including an API so you can build you own tools. My feeling is we should simply begin using the Internet Archive as one repository for our digital media collections. I would never put all our eggs in one basket, so to speak, and the best strategy would be to have several copies of everything in different baskets. How many and where is the question, but I think using the IA as one basket is probably a great idea.

HTML5 Delivery Across Platforms: Overcoming Challenges - (12:30 PM - 1:30 PM)

Presenters:
Brian Crescimanno - Brightcove
Michael Dale - Kaltura
Andrew Pile, Brian Joe & Ryan - Vimeo

I missed most of this session since everything was running late, and I got talking with Shay David from Kaltura instead of attending the session about Kaltura.  Kaltura (http://www.kaltura.org/) is an open source video platform for creating, publishing, and managing media in an institutional environment. The Kultura corporation sells a number of services including hosting and support, with clients like many of the Ivy League universities. Shay claims Kaltura easily integrates with Learning Management Systems, and CMSs including Drupal, Wordpress, and Joomla. (During a later session, representatives from Penn State, MIT, USC, and the University of Virginia discussed their use of Kaltura.) Kaltura can be downloaded and used by anyone for free, and it looks very powerful in comparison to the Ensemble Video system we’re currently using at Illinois.

Vimeo and Brightcove are also implementing HTML5, and I caught most of the Vimeo presentation. (We’re currently using Vimeo for some of our video hosting, since it’s very flexible and provides some great features like embedding, privacy settings, and HTML5 compatibility.) What struck me about the Vimeo guys is they are so young, yet they’re just building great tools on a continuing basis. All these companies’ systems are in perpetual Beta and it’s clear they’ll keep getting better…at least as long as they stay in business.

A side-note is appropriate here concerning the PBS COVE system. I talked with a number of public TV people at the OVC about COVE, and all hate it. Compared with other video publishing systems like Brightcove and Vimeo, COVE is a technical disaster. The administrative interface is confusing and difficult, the performance in a browser is terrible, and it’s not HTML5 compliant. I compared notes with the WGBH folks, who have also talked with people from other stations and we all see the same thing: GM’s love COVE because of the way it looks; the web staff who work with COVE all hate it because of the way it doesn’t perform. The PBS system has lots invested in COVE, and in my view that is a shame.

Hacking Public Domain Government Video - (2:30 PM - 3:00 PM)

Presenter: Abram Stern — UC Santa Cruz / Metavid.org

Abram Stern from UC Santa Cruz and Metavid.org presented on the joys and ridiculous difficulties of providing access to video archives of sessions and committee meetings from the U.S. Congress. Many parts of Congress produce their own streaming media archives, but each does it differently, e.g. in different formats, to a greater or lesser extent, etc., so the result is a mess. The metavid site (http://metavid.org/wiki/) attempts to bring order to the mess and make it searchable by topic. The site also makes it easier to locate specific sections within video archives by linking text transcripts to the video time code.

CSPAN has been providing lots of streaming archives since 1998, but these are mostly encoded in outdated formats like RealVideo. (Once a video is encoded to RealVideo, there is no good way to transcode it to a more useful format unless you have access to the original source files.)

The Sunlight Foundation is attempting to address this by capturing Congressional  video and encoding it themselves. They are very ambitious. They plan to provide very deep search tools, including the ability to search video by phrase or keyword. They are working to implement the popcorn.js video player code which will enable linking specific video segments to related information and resources. Example: you’re watching a video archive of a Senator discussing the Affordable Care Act, and makes a partisan claim about a specific provision in the Act. The video can provide a link to that specific provision, so the viewer can actually read it. The text of the provision can easily be displayed next to the video, while it continues playing. That’s just one example of what’s possible by creating time-based links to related resources. You could also display how much money a given Senator received from oil companies when that Senator is speaking on energy policy. (During the Hack Lab on Sunday, met Noah Kunin, Multimedia Content Producer at the Sunlight Foundation, who is very interested in collaborating on any government media projects we might engage in, including how to best use the tools and resources they’re building.)

You can see an example video using popcorn.js on the WebMadeMovies site: http://webmademovies.etherworks.ca/popcorndemo/. As the video plays, the popcorn script displays captions, and the user can select different languages for the captions. Translation is accomplished on the fly using Google Translate. Next to the video window, we see a Google Map showing locations relevant to portions of the video as it plays. Another window displays related Wikipedia entries, and another window shows related items from Google News. There’s also a live Twitter feed showing search results based on topics as they are mentioned in the video. Crazy yes! It could get a lot more useful, since you can include semantics in the video stream linking to literally anything, and display it any way you want. Popcorn.js is supported by the Mozilla Foundation, and is in active development.

Introducing the Pan.do/ra HTML5 media archive - (3:00 PM - 3:15 PM)

Presenters:
Jan Geber
Sebastian Lütgert

Pan.do/ra is intended as an open media archive and tool set for annotating, editing, remixing, and curating online video, but it doesn’t quite exist yet. It is based on a tool that does exist, http://pad.ma.  “PAD.MA - short for Public Access Digital Media Archive - is an online archive of densely text-annotated video material, primarily footage and not finished films. The entire collection is searchable and viewable online, and is free to download for non- commercial use.”  Pad.ma is in turn built from Oxdb (http://0xdb.org/about), a pretty amazing open source tool for harvesting metadata about video, creating full-text search, and other utilities need to make video content findable and usable on the web. These tools make it possible to not only find a video, but to find a specific point in a video which you can then view, link to, add further annotation, analyze, cite, etc. The whole point of these tools is to facilitate collaborative production, distribution, annotation, and sharing, and making open video more useful for all concerned.

State of Media Accessibility Session - (3:15 PM - 3:45 PM)

Presenters:
Silvia Pfeiffer — Mozilla, xiph.org, WHATWG

Silvia Pfeiffer from Mozilla and xiph.org presented a short session outlining the problems in implementing accessibility features with audio and video in various browsers. And there are many problems, because each browser handles accessibility features in different ways. HTML5 should provide good solutions, but won’t until it’s implemented fully and consistently. 

Implementing the Media Fragments URI Session - (3:45 PM - 4:15 PM)

Presenter: Raphael Troncy, EURECOMm W3C Media Fragments Group

The Media Fragments URI specification is currently being developed by the W3C Media Fragments Working Group. The group's mission is to create standard addressing schemes for media fragments on the Web using Uniform Resource Identifiers (URIs). “To make media a ‘first class citizen’ on the Web, it needs to be as easily linkable as HTML pages. Only when we are able to navigate through media resources based on semantics rather than random guesswork, will we really be able to master the full complexity of rich media.”  So the spec would allow use of a URI fragment (anything behind a "#" in a URI) and a URI query (anything behind a "?" and before a "#" in a URI) to address subparts of a media resource, e.g. to play, link to, or annotate.

Example: http://www.example.com/example.ogv#t=10,20 would address the segment of the video file example.ogv occurring between 10 and 20 seconds into the video. So if you were writing about something that was said in this video, you could not only link to the video but the exact spot in the video in which the quotation occurs. This allows annotations and metadata to be added to specific segments within media files, which opens up many new possibilities for semantic links.

The W3C Media Fragments URI draft spec was published in June 2010, and is available here: http://www.w3.org/TR/media-frags/.

An Introduction to PBCore 2.0: Metadata for Public Broadcasters - (4:30 PM - 5:30 PM)

Presenters:
Chair: Nan Rubin — PBCore Project
Linda Tadic — Audiovisual Archive Network
David Rice — Audiovisual Preservation Solutions
Chris Beer — WGBH Interactive

Nan Rubin, Linda Tadic, Dave Rice, and Chris Beer gave a good primer describing the reason for the development of PBCore and its structure. The room was full, so it’s clear there is great of interest in PBCore. We’re doing a PBCore 2.0 panel at the AMIA Conference next month, where we’ll talk about how we are using PBCore, the tools we’ve created, and what PBCore brings to the table in terms of archival practice. There is lots of momentum in the PBCore 2.0 project and with the American Archive making use of it, we’ll see some solid examples of its utility.

The Opencast Matterhorn Project Session - (5:30 PM - 6:30 PM)

“The Opencast Matterhorn Project is an international cooperation of academic institutions to build an open source solution to manage academic video. Embedded in the Opencast Community, it is to meet the needs of universities with respect to lecture capture especially while at the same time providing a flexible solution for the wide range of academic requirements.”

Opencast provides a system for publishing, managing, and sharing video among groups of producers and users. The Matterhorn Project (http://www.opencastproject.org/) is more ambitious than Ensemble, and less than Kaltura. It offers simple, inexpensive dedicated capture hardware for classroom lectures including both video and VGA inputs (e.g. for capturing both the lecturer and a Powerpoint presentation), automated encoding (using ffmpeg) and uploading to streaming servers, and a management interface for adding metadata (based on Dublin Core), publishing, and managing access to video archives and assets. It also manages source video files, instead of leaving them at the mercy of whatever practices producers use (or don’t use). Matterhorn is also open source and in very active development. The roadmap includes adding captions, speech-to-text, and many other features. The Opencast Community is trying to serve as a focal point for open source projects beyond Matterhorn.

I went to two session where the same Matterhorn presentation was given. It looks like a nice system. One takeaway for me is the question How many open source video management systems and focal points for open source development do we really need? Are we diluting out collective resources? I don’t have a good answer, and I do think Matterhorn deserves a closer look at the University of Illinois, along with Kaltura.

Saturday, October 2, 2010

Keynote: Mike Wesch - 10:00 AM - 10:45 AM

Dr. Wesch is an anthropologist and professor at Kansas State University. If you haven’t already seen his video The Machine is Us/ing Us, stop reading this now and watch it on YouTube: http://www.youtube.com/watch?v=6gmP4nk0EOE&feature=channel

Here’s another video he created with his class entitled A Vision of Students Today: http://www.youtube.com/watch?v=dGCJ46vyR9o. This one is especially relevant to the discussion of instructional technology and the renovation of classrooms at the University of Illinois.

Wesch’s main point is that media changes the way we interact and thus changes our culture. Any profound change in media has a major effect on culture. Yet we are still teaching our students as if living in the 19th century. In an age of interactive media and anyone can create, we are still treating people like members of a passive audience. “We have to stop thinking about our schools (and I would add media channels) as information dumps,” Wesch says. “The best teachers are those who are learning alongside their students.” His method is what he terms “project-based learning”…with no syllabus, just a research project involving the entire class working collaboratively using Google Docs and whatever shareable tools and resources they find they need. He says we need to move from thinking about “information literacy” to “digital citizenship.”

To me this is very exciting. My takeaways: Applying this concept to journalism and storytelling, we as Illinois Public Media could think of ways to engage members of the “former audience” (Shirky) in the unfolding project of journalism and storytelling. Our job is no longer filling empty vessels with information, but engaging people in the shared project of understanding the world and making the best of our possibilities. Digital media now connects all of us in a global system, and it’s now both ubiquitous and personal. What can we do with that?

Open Video Innovation in Journalism session - (10:45 AM - 11:45 AM)

Emily Bell (chair) — Director, Tow Center for Digital Journalism, Columbia
Janine Gibson — Editor, guardian.co.uk
Nonny de la Pena — co-founder, Stroome (Knight News Challenge winner)
(There was a fourth panelist form the Huffington Post but I didn’t get her name)

Summary: Traditional media organizations no longer have a monopoly on the means of gathering and publishing media. How many cameras in the world are there anyway? More than there were a few years ago, to say the least. So we see the rise of “witness journalism” where non-professionals are essential sources of content. Of course this raises questions of credibility and responsibility, but the trend will not be stopped so we have to make the best of it.

Janine Gibson: When asked What is your digital strategy, her answer is “we ask our interns.” The Guardian is trying to open the door to more content contributed by readers etc. This has sometimes led to major breakthrough stories, like a piece of video that revealed corruption by an MP. Amateur video led to professional reporting by the Guardian. How do you get people to contribute? “The key bit is the asking.” But if we’re too afraid or risk-averse, we don’t ask…

Nonny de la Pena demonstrated Stroome (http://www.stroome.com/), a Knight-funded open source software tool allowing browser-based video editing, collaboration, and project management. These tools will only get better in the near future, and if traditional media organizations don’t use them to engage people, they’ll use them without us.

Tools for New Media Teaching and Learning - (12:00 PM - 1:15 PM)

Panelists:
Olaf Schulte - Opencast Matterhorn
Mark Phillipson/Schuyler Duveen - CCNMTL MediaThread
Peter Pinch - OpenVault (WGBH)

The highlight of this session was Schuyler Duveen demoing MediaThread, one of the coolest pieces of open source software I’ve seen for a while. MediaThread (http://ccnmtl.columbia.edu/portfolio/custom_software_applications_and_tools/mediathread.html) provides a browser-based collaboration space, and a bookmarklet. A bookmarklet is a bit of JavaScript that you literally drag into your browser toolbar. Then you visit a video site, like YouTube, Wikimedia Commons, the WGBH Open Vault site, and others, and when you see a video you’d like to annotate for whatever reason, you drag the bookmarklet onto the video. Suddenly you’re looking at the video in the MediaThread project site, where you can add notes and other metadata. You can play the video and select specific segments, then annotate those. You can then grab other videos from other sites and bring them into the project page. You can then share the project page with others. MediaThread becomes a simple way to add metadata to media, including specific sections of the frame.

I spoke with Schuyler during the Hack Lab on Sunday. It seemed obvious to me that one could use MediaThread as a user interface for building time-based links for use with popcorn.js, and he agreed. Both are open source projects, and it seems possible they could collaborate directly after discussions at the OVC.

Fostering Open Culture in Higher Education session - (2:45 PM - 3:30 PM)

Stanford University School of Medicine media folks struggled for years capturing classroom lectures on VHS videotape, which made it difficult to make the content actual usable. They recently built a media management system using Apple products including Final Cut Server and Podcast Producer, along with lots of custom code. They also worked with the Stanford Legal department to develop rights policies, including the use of Creative Commons licenses for classroom lectures etc. They built an interface for uploading and cataloging media content, including options for selecting different rights expressions. They got buy-in from the administration and the faculty, and they say it’s going very well now. Classroom lectures are actually accessible for the first time. A very nice presentation and a good success story.

Educational Video Service Strategy - (3:45 PM - 4:45 PM)

Panelists: Chris Millet, Manager of Educational Technology (Penn State)
Steven Gass, Associate Director for Public Services (MIT)
Wendy M. Chapman, Director of Web Technologies, Facilities and Technology (USC)
Jama Coartney, Head of the Digital Media Lab (University of Virginia).

All have now adopted Kaltura as a video management system, and pretty much raved about it. Didn’t get into technical details, but discussed how they are moving pretty quickly to create and publish lots and lots of educational media content. MIT has revamped their Open Courseware project using Kaltura.

New Models for Production: In Conversation with Sally Potter - (4:45 PM - 5:30 PM)

Presenters:
Dr. Eric Faden — Bucknell University
Clare Holden — Adventure Pictures/SP-ARK
Peter Kaufman — Intelligent Television
Sally Potter — auteur (via Skype)

Filmmaker Sally Potter and her associate Clare Holden demonstrated SP-ARK (http://www.sp-ark.org/), an open source video archiving system allowing detailed annotation of films and related materials (scripts, call sheets, etc). To me this is yet another example of an “information environment” providing an entry point into a universe of related data and content.

Peter Kaufman spoke of building new partnerships among filmmakers, new media news organizations, archivists, and open source developers, and with established corporate media entities who after all have a great deal of interest in media access and preservation.

Dr. Eric Faden spoke about the challenges in teaching today’s college students about media. He makes students produce videos instead of writing papers, and while some of them take to it, others hate it. He told of one student who struggled with it but graduated, then got a job with an insurance company. Her company needed someone to produce a video about a new product for discussion at a corporate meeting, and she was able to do it. He says video production skills are increasingly in demand in a growing number of professional environments. “We need to teach not just media literacy,“ he says, “but media competence. They have to be able to actually make things.”

The Medium of Record: The Daily Show's Adam Chodikoff - (5:30 PM - 6:15 PM)

Adam Chodikoff is the guy who finds all the archival video clips for Jon Stewart. Like the interview where Dick Cheney says he never claimed there was a direct connection between Saddam Hussein and Al Qaeda, and then we see a clip of another Cheney interview where he makes exactly that claim. Adam Chodikoff played a number of segments from The Daily Show so that was fun. I asked him later if they ever have to license content for use on the show, and if they ever got complaints from rights-holders. He told me he doesn’t know about that stuff, as it’s “above my pay-grade.” One interesting thing I learned is that Jon Stewart comes up with many of the ideas that make it onto the show. They have a team of writers, but Jon is deeply involved every day.

Sunday, October 3, 2010

“Hack Day” was a very informal day of hands-on work at NYU's Interactive Telecommunications Program. I was able to attend for about three hours before heading to the airport for the trip home. I spent most of this time working with about a dozen other people to implement popcorn.js for a sample video. We almost got it done before I had to leave, and of course I brought back all the code. The scene in this space at NYU was astonishing. On a Sunday afternoon, there were about 120 people crammed into the forth floor of this building on Broadway, all hammering out software projects of one kind or another. Several projects were being personally guided by Ben Moskowitz, who works for Mozilla and is the main organizer of the OVC. Many of the people working in the popcorn.js session were filmmakers and archive managers, not programmers. More to say about popcorn.js but I’ll save that for my team. Hack Day at OVC was a fantastic model of how to leverage the expertise of programmers to educate non-programmers on the use of open source software and methods, and facilitate use of these new tools among a global community.

Summary Thoughts

Systems and tools are rapidly being built to make best use of “open source, open standards, open formats, and open access.” Mozilla is trying to foster a sense of community among those engaged in various projects, and it’s not just about technology but also “open culture,” digital citizenship, and news ways of thinking about media, education, and community. Yet most of us are so focused on our own projects that we don’t see the potential intersections of our work with others. If we intend to move forward in media, education, and technology, we must collaborate.

(0) Comments/Permalink | Categories: media • metadata • technology