Understanding YouTube Content ID

When YouTube detects your music in a video, monetization (ads) will be turned on for that video. If there is a text or banner ad on the YouTube video with your music in it and the user clicks on the text or banner, or if there is an AD before or after the video and the user watches up to 30 seconds or the entire commercial, ad revenue is generated and collected by us for the original music creator.

The process would work something along these lines:

  1. User uploads a video.
  2. Youtube then queues up the video to be processed server-side. It is encoded and quite a few different versions are created (HTML5 version, HD, non-HD, etc).
  3. If the video contains audio a hash is then calculated based off a time-frequency graph called a spectrogram. The hash for audio works similar to that of sound discovery apps like Shazam or Soundhound. Target zones (peak points in the spectrogram) are marked, then the target area between them is also taken and hashed. The hash would be a string that might resemble something like this: kjdlas980394jmlkjkjlkdf=312938092183jhkjfdfkjsdf (completely unreadable in human form).
  4. Frames from a video (most likely minus blank screens and predominately text-based frames) a sample of a section of the video is taken. The frames are being analysed when the video is being converted server side, so this isn’t expensive or a separate operation. A hash is created from those sampled frames from the video. The hash would most likely resemble the one created above, a series of jumbled letters and characters that are unique to that content.
  5. Content ID is based on audio and video samples that rights holders have uploaded to Youtube. This is separate from the other process. Essentially Youtube has a database of legitimate source material to create hashes from.
  6. When a video is uploaded and those hashes are created, they are most likely compared to that of the rights holders database hashes. I am sure they probably have a likeliness factor there as well, as direct matches would be very rare.

Please watch the video below to understand Content ID:

What kind of Content is Eligible for Youtube’s Content ID?

You must have exclusive copyright rights to the material in the reference file for the territories where you claim ownership. Sound recordings submitted should be original, and not have any open or unlicensed samples, sound effects, sound bites, beats, royalty-free music, etc.

The following examples are ineligible for use in or as a reference:

  • Recordings that contain audio made available for free from Garage Band, Ableton, Logic, Fruity Loops and other DAWs are also prohibited.
  • Content licensed non-exclusively from a third party
  • Content released under Creative Commons or similar free/open licenses
  • Public domain footage, recordings, or compositions
  • Clips from other sources used under fair use principles
  • Video gameplay footage (by other than the game’s publisher)

More information about this from YouTube can be found via the link:

Original Video Game Soundtracks are only valid references when owned by the video game’s publisher.

Original Video Game Soundtracks are sound recordings created primarily for use in a video game, such as the original score or background music accompanying a video game.

These references are only valid when delivered under the video game publisher’s content owner.

All reference content must be sufficiently distinct.

The following examples are ineligible for use in or as a reference:

  • Karaoke recordings, remasters, and sound-alike recordings
  • Sound effects, soundbeds, or production loops and beats

Youtube Content ID by DIGITUNES

We’ll add this single to YouTube’s Content ID database, and continually scan for matches. When your music is detected in any YouTube video, you’ll be notified—and ad revenue will automatically go to you, instead of to the person who uploaded the video.

Leave a Reply