Contents technology protects copyright

By Yoon Ja-young

YouTube, the biggest video-sharing site in the world, has 48 hours of new video uploaded every single minute and over 3 billion video views every day. Most users, however, often forget one crucial issue regarding the videos ― the copyright. Most of the videos they upload, except the ones created purely by themselves, have copyright.

If someone uploads a video of their family enjoying vacation on a beach, and adds a song to it as background music, he or she is breaching the copyright of that song. If someone captured only a few seconds of a music video and uploaded it on YouTube, it is also a copyright violation.

If all these videos were to be deleted from YouTube, however, there wouldn’t be much content to enjoy. YouTube thus solves the problem with a “contents identification” platform.

Complication of copyright on YouTube

The contents identification platform deals with the complex nature of digital copyright management so that copyright holders can better control and manage their contents on YouTube.

For music, for instance, there are two copyright holders, the producer and the composer. When studios sell distribution rights in each country, it gets more complicated. The studios generally sell distribution rights to different entities in each country. For example, it will sell the rights to CBS in the United States, and to TF1 in France. “There will be separate global rights,” said David King, a YouTube Group Product manager who is in charge of the contents identification platform.

“When YouTube seeks licenses for using contents uploaded by users, we have to tie all these together to get global approval for the use of these contents. As we examine this, we end up with on average of five owners, sometimes 20, 30 or 40 different owners split over different continents. We tie this up together. Moreover, our users create mash-up of contents and this further complicates the problem.”

YouTube copyright technology

Since October 2007, YouTube spent over $30 million developing copyright-control technologies. Its mission was to provide scalable and automated solutions that would allow rights holders to choose whether and how their contents appear on YouTube.

“We analyze every one of the uploaded videos and, on top of that, 100 years’ length of videos are scanned by contents identification every day,” King said.

Over 2,000 media partners are using the contents identification platform, and they have claimed copyrights to over 120 million videos, which accounts for more than one third of monetized videos on YouTube.

The monetization process begins with the copyright holder’s uploading a file. It becomes the identification file in the database. When YouTube users upload videos, YouTube matches them with the identification files. When there is match, all copyright information is attached to the user content.

Notable of the contents identification technology is that it is sensitive to changes between reference and user videos. For example, YouTube finds the match even when the user video was made from TV screen clips captured with cameras. The entire matching process occurs within one minute, before the user videos are published on YouTube.

When it matches their content, partners have three choices. The first choice is to lock the content, preventing the matched material from going live on the site. However, they rarely choose this option since it can anger fans.

The second choice is to track the content, allowing the video and collecting detailed information on who is viewing it. The third option is monetization, putting advertisements on the video with YouTube and the copyright holder sharing the revenue.

How copyright holders make money, King said, is that partners are increasingly choosing to monetize rather than block the contents.

In a popular user video titled “JK Wedding Dance,” the contents identification platform found that the music played in the video was “Forever” by Chris Brown. As Sony Live had chosen monetization, they put an ad on the video and also added a link to iTunes. Though the song was released 18 months ago, it rose to the fourth place in iTunes. Sony, of course, made money from both advertising and iTunes.

Over 2,000 partners use the system, including the country’s major entertainment firms like YG Entertainment, SM Entertainment and JYP Entertainment. Currently the number of reference files is around 6 million.

“YouTube also provides detailed reporting and analysis to our partners, such as which of their videos are the most popular, about global viewers, or which part of the video was the most popular,” King added.

He said contents identification is beneficial for users as well. They can enjoy diverse contents without worrying about copyright infringement. This is a critical issue. In Korea, for instance, a father who uploaded a video of his five-year-old daughter singing and dancing to a song of Son Dambi, a popular singer, to his blog, had his video deleted from the Internet by the blog site operator upon request from the copyright holder. The father filed a lawsuit and the incident brought controversy over how complex copyright is.

YG Entertainment, which manages globally popular artists like Big Bang and 2NE1, partnered up with Google last August. It has recorded over 370 million video views so far.

“We didn’t expect that much when we signed up for the partnership. I think the K-pop boom owes the most to online services such as YouTube and Facebook,” said Simon Choi, managing director at YG Entertainment.

Most recently, the digital single “Lonely” by 2NE1 marked 5 million views in less than a week roughly averaging almost 1 million views each day.

Before the partnership, YG Entertainment hired some online copyright management companies which would search for copyright infringement on Websites and send the uploaders warnings. However, YG was neither expanding fan base nor earning profit this way. With contents identification, it becomes a win-win situation for both fans and YG.