Metadata working group
TX Metadata Working Group
Co-ordinator: Mick: Mickfuzz at rocketmail.com
A release candidate for the Transmission Metadata Standard 1.0 is available here for comment: TX Metadata 1.0 rc
Please join the mailing list below to add your comments for the final release.
admin of list Mickfuzz at rocketmail.com
Metadata is "information about information". The range and
A metadata standard for online video will ensure a common definitions for basic
If you are not sure what a 'feed' is and why it is useful then just look at these
- Miro - http:///www.getmiro.org - is a Video player that shows you a selection of channels. These channels are actually created from RSS media feeds.
- Gotomeeting - Allows conferencing between multiple participants. Using gotomeeting increases publicity for metadata working groups.
- Podcasting - ITunes reads podcasts from wherever they are published and allows you to play the audio and video files downloaded from these podcasts. It also automatically checks back for new
files as they are published. A podcast is another word for a RSS media feed.
What has been done to date
The initial draft standard was presented to the London Re:Transmission meeting in October 2006. Since then the standard has received substantial technical feedback, and it is now ready for testing. We recently released a 0.9 version (pre-pre-release) of the TX Metadata Standard. You can find it here: http://clearerchannel.org/docs/tx_metadata_standard_0_9.pdf
We plan to start implementing the standard into Plumi, Drupal and Wordpress video content management systems as soon as possible
Documentation of Discussions:
Metadata Standardisation discussion at Re:Transmission, London, October 2006
What needs to be done
We are not imposing a structure from above, but starting small, from the bottom up. Once the concept and
With the Transmission feeds, existing tools will offer more and better search facilities, so that you can click on a keyword or enter a search term on any one participating site, and immediately find the range of videos you seek, wherever in the network it is hosted.
The work has been being planned and co-ordinated here: Metadata Workspace
Metadata Project Method and implementation plan
Drupal implementation on content provider site
Suggested site clearerchannel.org
Drupal implementation on aggregator site
Current test case is ifiwatch.tv. For this to be fully useful, we need more video publishers of relevant material to adopt the standard and create search-based media feeds.
An initial list is here: List_of_sites_to_encourage_to_create_good_clean_suckable_search-based_feeds
some proportion of the funding for the metadata standard implementation needs to go towards the costs of getting these projects to create useful media feeds based on the metadata standard. this budget can be listed here: Ifiwatch.tv/start_of_2008/Dividing_up_Budget
Suggested site for next aggregator test case is transmission.cc.
Feedback to standard
This part of the document is not at all clear.
There is no mention of a unique identifier in the table detailing "Entry Metadata". However there is a tag called <id> in the sample feed which would serve this purpose.
Seemingly It is not defined by the standard what the <id> or unique identifier of the Items should be. This is an open field. In the example given the id is a URL.
It would be good to create some guidelines to help media creators find a unique identifier for their work. In particular, this will be necessary to refer to updatable hashes where subtitle files etc can be stored.
Media file hash info and tools for hashing
It would be good to have a hash or similar identifier for each known version of media files. This could then be used to look up information about the other metadata based on linking these hash files with the unique identifier of the Media item.
In other words it would be good to cross reference the Unique identifier for the Video with all known versions of the file.
Md5 is a popular way of identifying files.
using SHA1 instead is recommended http://en.wikipedia.org/wiki/SHA1
Bitpedia is a useful tool for this network as it helps the tracking of files.
There is more information here. http://bitzi.com/bitpedia/
But a summary of what it could be used for is...
When you have a video file that you are distributing it is useful for other people to be able to find out metadata about that file which currently isn't really easy to include inside the actual file. Ogg files can include some metadata but realistically when you create that file you will not be able to include some information that people will later want to know.
For example the location of subtitle files for the video file is not information that will be complete when you create the file.
There is a need to link the version of the
file you are distributing with an online source of metadata. Bitpedia allows the tracking of files by linking a hash of the file - a "fingerprint", with online information about it - a "ticket".
In the words of Bitpedia - Currently, Bitpedia information is in the form of Tickets, structured metadata summaries which exist one per file. Files are identified by their Bitprints, distinctive digital fingerprints calculated mathematically by examining the file.
A similar system doesn't necessary need to use the same ticketing system. In fact this one is full of really strange adverts. But the software is opensource, it is possible to create a different interface to see the information about the video, the metadata
This is significant for the Translation/Subtitling project as well, since it
Are we doing the right thing
A section for links to other ways of doing this and other approaches to the aims of the Metadata project.
A proposal for a structure of the information, a bit centralised but only in data integrity, not necessarily in the server themselves, could be this: Problems
- This Metadata database has to be filled out by people, and so our proposal needs to go to the people that will make it, sending a lot of information about procedures along with the data structure.
- It is irrational to assume that when we create this database, people will just submit their information or They will need to be actively encouraged to submit info.
- Video makers often lack information about how to order their material once it goes online, and how to allow people to make contact with them after the video went online. A process of educating video makers will need to be part of this project.
- We should provide an interface for video makers to make use of the
- We should provide a way for somebody to find the author, subtitles, a better encoded version of a video which they found on its own in the internet, without extra information.
- We should have some sort of unique identifier for this database - i.e. the video 419412 is just that one, not other)
- We should provide different ways of entering the information (when people upload them on a different server, for example, the metadata submission process could be embedded there too)
- We should allow the information to be entered by people other than the creator of the video, and at the same time provide some means of validation for this information from the video maker.
- We should present timestamps of "Last seen" for the different video formats and subtitles etc
- We should create an indexation page that can serve the video makers as a pointer to their material
- We could allow contact information changes related to a video
- We should be able to "merge" information about two different videos, if we find out they are the same
- We should be able to "relate" two videos in an RDF-like way: this video is --a resume--> of this video
- We see here already that there will be a need for moderation, maybe user driven, wikipedia-like
- Version control.
- Provide a REST service for uploading video sites (engagemedia/archive/clearerchannel/ifiwatch.tv) that allows the uploader to submit information to the Metadata database
- Being able to generate Unique identifiers based mainly in the metadata of the "video node" (author, date, title) that connect to different files associated.to that "video": the dvd cover, the DivX version, the VOB for making a DVD, the subtitles in Chinese, etc
- All this files will link back to the video node
- When submitting a video node we will have to provide, always, a version of the video. Once the metadata has been filled in, we could present the users with other metadata entries that look similar and give the option of adding the new file to an already existing entry in a really easy way (providing quick validations, like showing the video)
- We can also lookup videos on the database to see if there are better versions: You give the youtube link and the database tells you "but there is a HD ogg file here!"
- The "Video node" page is a pretty page you want to link to when presenting your video somewhere else.
- You can submit videos that are uploaded in services that don't use the metadata api services.
- You can get shitloads of rss feeds
- You can tag videos and moderator would be able to create collections of titles (archive.org like)