YouTube Announces Improvements to Content ID System
To fight the unauthorized use of copyrighted video, YouTube created the Content ID system nearly five years ago. It automates the process of checking for duplications by creating a fingerprint of uploaded reference material and comparing those stored fingerprints against all newly and previously uploaded videos. Surprisingly, the process takes only minutes.
Yesterday, YouTube announced improvements to the Content ID system that have recently been put into place. First is a new appeals process. In the previous system, when Content ID rejected an uploaded video the uploader could dispute the decision, but the original content owners could reject the dispute and it would end there. Now, uploaders can file an appeal. The content owner can then choose to release the claim or file a formal DMCA notification.
YouTube acknowledges that mistakes can happen in the Contend ID process, so it's improving the system's algorithms for detecting invalid claims for reference videos. Reference videos that may not actually be owned by the uploader are manually reviewed. YouTube notes that it will ban the use of Content ID to people who misuse it.
Finally, YouTube says it significantly improved Content ID's video matching ability earlier this year.
At the 2011 Streaming Media Europe conference, YouTube Europe engineering director Oliver Heckmann gave a keynote address that explained how the Content ID system works. Content ID is currently used by over 3,000 content owners who have uploaded more then 500,000 hours of reference videos.
In a sea of free ad-supported content, YouTube offers subscriptions. Does it have the clout to pull this off?
The video sharing leader touted its reach with young viewers, while Dreamworks announced its acquisition of Awesomeness TV.
It now takes three full days to watch all the videos uploaded to YouTube in the average minute.