This page is a work in progress and is intended to provide resources for working with media assets in a repository environment. |
These are areas of difficulty that we are all facing together. For each of these topics, there are at least a handful of people who are actively working to reduce these challenges to non-issues. Are you one of those people? If you are, make sure we all know about your work, and make sure it's reflected here on the wiki.
There is a lot of discussion about digitization and the tradeoffs of an uncompressed format vs a lossless format.
Other communities interested in digitization and preservation include:
Helpful utility: AJA Data Rate Calculator Mac | Win
WGBH has created a very atomistic content model to support their rich media archives repository.
Content types:
There are three basic approaches to delivering video files: progressive download, rtsp/rtmp streaming, and progressive download with HTTP range requests. Many media repositories are currently using Fedora redirect datastreams to reference content hosted on dedicated media servers (either locally, on a CDN, etc). The obvious advantages here are optimized delivery of media content and the ability to offload processing and storage to external processes. The biggest disadvantages are loss of integrated authn/authz from Fedora and additional effort for allowing Fedora to manage those assets.
One current limitation in Fedora is it uses the Java URL parsing library, which fails to recognize rtsp:// format URLs. This is further confounded by Quicktime's inability to follow HTTP redirects to rtsp:// streams. One possible work-around is to use the quicktime media link xml schema to redirect the Quicktime plugin appropriately. WGBH has developed this proof-of-concept.