Some of you might have heard about DukeStream 2.0 –our longstanding search for a YouTube-like media publishing system, anchored in asset-level security and based on Duke NetIDs. When the DukeStream 2.0 project began a couple years ago, there weren’t any tools that had the requisite asset-level security, or who could implement a system such as this for Duke in a user-friendly and cost-effective way. The landscape is changing, however, and we are in discussions with a number of vendors who seem to be in a position to offer something that meets our core requirements.

Along they way we’ve also encountered products that may not be a great fit for DukeStream 2.0, but are nonetheless interesting. On the recommendation of Duke Libraries and CIT, we looked into Avalon Media System, currently in v. 1.0, and which we heard is being billed as a “Kaltura killer.” Avalon piqued our interest given that we are currently using Kaltura as a plugin for Sakai and are interested in new contenders in this space. Avalon bills itself as an “open source software system to enable academic libraries and archives to easily provide online access to their video and audio collections.” This would seem to place this tool in the Media Asset Management product niche, targeting libraries where “super users” manage media more than end users of enterprise systems who are simply looking to publish video (i.e., YouTube and Vimeo).

Based on our initial look, we identified the following issues that would be obstacles to our implementing Avalon for DukeStream 2.0 at present:

• Workflows seem typical of MAM/ DAM systems and as such may be more complicated than we could offer to general users. For example, the interface for uploading and editing files has 6 separate pages you click through sequentially to edit things like metadata, title and description, and access control. We are looking for something more consolidated.

• Opensource/ homegrown model makes it not a good fit for DukeStream at present because of constraints around development resources and our preference for buying rather than building for this scenario

• It looks like authentication isn’t currently enabled for the embedded player (i.e., embedding into 3rd party web sites). For this, they are looking at secure token-based authentication for the media player as per (https://wiki.dlib.indiana.edu/display/VarVideo/Authentication+-+Authorization)

• No specific mention of Shibboleth integration for the wider system (but also no mention that it is excluded–merits further looking into)

• Uses FFMPEG to transcode a wide range of input formats (https://wiki.dlib.indiana.edu/display/VarVideo/Avalon+Media+System+Supported+Source+Formats+and+Codecs) (WMV not listed, but actually worked in testing), but appears to only output to Flash. We are assuming the output is H.264/ .mp4 since fallback to HTML5 is listed as a player requirement as per https://wiki.dlib.indiana.edu/display/VarVideo/Video+players

• Only supports Adobe Flash Media Server and Red 5 streaming (support for Wowza not listed)

• Current player in test environment looked to be underdeveloped (no embed code option); however, integration with Matterhorn Engage player is specified, and this player does support embed code. Likely a matter of some custom configuration. One user on the forum commented that video playback performance on the test site was slow and suspected the player to be at fault, though this has not been confirmed.

• Current limit of 250MB for uploads in test site

• Confusing to get to a “My Media” view where a user can see all of their media. This may be another indication that the product targets MAM superusers who are more technically able than the average user who wants to upload video.

Overall, we were impressed with Avalon and intend to follow the progress of this project as it evolves. However, given the issues listed above and the fact that other candidates seem closer to our requirements, Avalon isn’t a tool that could be considered for DukeStream 2.0 at this time.

On the subject of cool new gadgets and accessories, my friend Joel Elliott from the National Humanities Center (who also clued me in to the Yeti, by the way), pointed out an interesting new iPhone accessory called the Glif. This one functions as a clamp for your iphone that lets you connect it to any standard tripod:

http://kk.org/cooltools/archives/9723?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+CoolTools+%28Cool+Tools%29

Cisco has recently announced the End of Life timeline for their QuickSet C20 codec and peripherals (details found here). Although this is a concern for many users, there is no need to freak out over the possibility of suddenly having an expensive paperweight. The popular VTC system will continue to receive support for the next 5 1/2 years under the SmartNet maintenance program as long as it is renewed annually.

Details on the End of Life timeline are summarized below:

• August 6, 2013 – Last day to order new C20 codec and peripherals
• August 6, 2014 – Last day of new software release for C20 codec
• August 6, 2014 – Last day of routine failure analysis for C20 codec hardware
• August 6, 2014 – Last day to order a NEW SmartNet support contract
• November 1, 2017 – Last day an existing SmartNet contract can be renewed
• August 31, 2018 – Last day the C20 codec can be covered under SmartNet support contract

As an alternative, the Cisco SX20 is the current model that Cisco is concentrating their development efforts on for this smaller type of conference room system. It has proven to be a very capable replacement while providing similar if not enhanced capabilities while remaining very affordable for a true HD video conferencing system. For a comparison, please take a look at Jack D’Ardenne’s blog post from last year.

As always, please contact oit-its-info@duke.edu if you have any questions concerning this announcement or other video conferencing products.

Here’s a scenario that I imagine is not entirely uncommon: Users have been putting photos onto various shared drives for years. Probably a few thousand images, many without clear labels, a few different folder systems, and no tags. So, how do you tackle this?

I’ve had discussions with a few people on how best to handle this, some of what was suggested:
1) Leave the archive as a wild no-man’s land. Focus on making an easy system for organizing/images images moving forward.
2) Cherry pick the most useful seeming folders, organize them, forget the rest (a friend’s advice)
3) Run all public images through Picassa, tag moving forward, and tag past images when convenient with a bit of assistance from the facial recognition (my current tentative plan)
4) Bite the bullet and manually go through all the images

Was just wondering if anyone at Duke had dealt with this successfully, and/or had suggestions for tools and best practices for working through this. (Note: This was previously posted to WebCom, though I feel some folks here might have suggestions as well). Thoughts?