Four Ways To Protect Your Premium Video Content From Piracy (Part 2)

In the first part of this blog series, we covered two key approaches to protecting your premium video content and minimising piracy – Tokenisation and DRM systems. This blog will explore the two other strategies, Digital watermarking and Content fingerprinting, which are integral components within the broader landscape of DRM systems and tokenisation, and play pivotal roles in safeguarding intellectual property and ensuring content integrity. By examining their applications and significance, we will uncover the approaches employed by OTT platforms to uphold content security and rights management standards.

Digital Watermarking

Digital watermarks embed information or data into media, such as images, audio, video, or documents, in a way that is typically imperceptible to human senses. These embedded watermarks serve various purposes, including copyright protection, content authentication, and ownership tracking. Here is how digital watermarks work:

  1. Embedding the Watermark:
  • Payload Data: The information to be embedded in the digital media is referred to as the “payload data”. This can include copyright information, ownership details (i.e. email address, account ID, etc), or any other relevant metadata by which we can identify the user and take action against those who are involved in piracy by terminating or suspending their account, or by taking legal action against them.
  • Key: A cryptographic key is often used to control the embedding process and ensure security. This key is known to the party embedding the watermark and may be kept secret to prevent unauthorised removal.
  1. Modifying the Media:
  • Depending on the type of digital watermarking technique used, the media is altered in a way that incorporates the payload data without significantly degrading the quality or user experience.
  • In image or video watermarking, changes are made to specific pixels or frequency components of the media to encode the watermark. In audio watermarking, it may involve altering the audio sample.
  1. Imperceptibility:
  • One of the key principles of digital watermarking is that the embedded watermark should be imperceptible to the human senses. This means that viewers or listeners should not notice any visible or audible changes in the media.
  1. Detection and Extraction:
  • To retrieve the watermark and its embedded information, specialised software or algorithms are used. This software analyses the digital media to identify and extract the watermark.
  • Extraction typically requires knowledge of the watermarking algorithm and, in some cases, the cryptographic key used during embedding.
  1. Verification and Action:
    After extracting the watermark, it can be authenticated for various purposes:
  • Copyright ProtectionThe watermark may include copyright information, enabling copyright owners to establish ownership in instances of copyright infringement.
  • Content AuthenticationIt can verify that the digital media has not been tampered with since the watermark was embedded, helping to ensure its integrity.
  • Ownership Tracking: Watermarks can include information about the creator or owner of the content, aiding in tracking its origins.

On obtaining information about users who have engaged in content piracy, a media owner can initiate actions against those individuals.

Content Fingerprinting

Video fingerprinting, also known as video content identification or video hashing, is a technique used to uniquely identify and match specific videos based on their content. The method includes pulling out unique characteristics from a video and forming a shorter version or fingerprint. Here’s an overview of how video fingerprinting works

  1. Frame Selection:
    • Video fingerprinting often begins with the selection of keyframes from the video. Keyframes are representative frames that capture essential visual information. They are selected to reduce the computational load and focus on the most informative parts of the video.
  2. Feature Extraction:
    Various features are extracted from the selected keyframes. These features can include:
    • Colour Histograms: Representing the distribution of colours in the image.
    • Texture Features: Describing patterns and textures in the video.
    • Motion Vectors: Capturing the movement of objects between frames.
    • Keyframe Intervals: Determining the time intervals between keyframes.
  3. Hashing or Encoding:
    • The extracted features are then used to generate a unique hash or encoding for the video. This process condenses the information into a compact representation that serves as the video fingerprint.
  4. Database Storage:
    • The generated fingerprint is stored in a database along with metadata about the video, such as its title, duration, and any other relevant information.
  5. Matching Process:
    • When a new video is introduced to the system, its fingerprint is computed using the same process.
    • The system compares this new fingerprint with those stored in the database to identify potential matches.
  6. Matching Threshold:
    • A matching threshold is set to determine when a match is considered significant. This threshold helps control the sensitivity of the fingerprinting system and reduce false positives.
  7. Robustness to Variations:
    • Video fingerprinting algorithms need to be robust to variations in the content, such as different resolutions, compression formats, and slight modifications to the video.
  8. Temporal Consistency:
    • Video fingerprints often take into account the temporal consistency of the content, ensuring that changes over time are appropriately represented in the fingerprint.
  9. Scaling and Efficiency:
    • Systems must be scalable to handle a large volume of videos efficiently, especially in applications where content identification is performed in real-time or near-real-time.
  10. Security Measures:
    • Security considerations are important to protect the integrity of the video fingerprinting system, including measures to prevent tampering with the database or the fingerprinting process.

Let’s explore the practical application of fingerprinting for identifying copyrighted content.

Initially, we embark on creating a video fingerprint, which entails extracting data such as visual patterns, scene alterations, color distributions, or any distinctive markers present within the video. For instance, this could involve capturing a keyframe from the 15th minute of the content, analysing colour schemes within these frames, identifying the singer’s voice after 10 or 60 seconds, and so forth. The process of fingerprint extraction involves transforming these attributes into a digital fingerprint.

This fingerprint serves as a concise representation of the video, encapsulating its unique characteristics while remaining compact enough for easy storage and comparison with other fingerprints. Upon uploading new content to the platform, it undergoes a comparison with stored fingerprints in the database to ascertain its copyright status. If the content is found to be copyrighted, the platform may take necessary actions such as removing the video or halting any associated monetary benefits.

Video fingerprinting is widely used in applications such as content identification, copyright protection, video recommendation systems, and content filtering on online platforms. It is used in online platforms such as Youtube, Instagram and TikTok, where other content creators can upload the same video without the proper copyrights.


Watermarking and fingerprinting differ significantly in their approaches to content protection. Watermarking alters the content by embedding an identifier, requiring specialised tools for detection, and is used for tracing the source and owner. It is applied during or after content creation and, although difficult, the watermark can potentially be removed. Conversely, fingerprinting does not alter the content but analyses and extracts features, matching the content against a pre-existing database of fingerprints. It is used for identifying and monitoring content across platforms, applied post-distribution for monitoring purposes, and is non-invasive, meaning it does not involve altering the content.

Watermarking provides a method for tracing unauthorised distribution of the content. Once a content creator gets the distributor details, they can take legal action if necessary. Content fingerprinting excels in scenarios requiring content identification and monitoring across various platforms. It enables content owners to efficiently scan for matches or infringements, facilitating a proactive content protection approach.

Tokenisation and DRM serve as the initial measures for premium OTT platforms to combat content piracy, safeguarding content from unauthorised access. Following their implementation, media owners can further fortify content protection through techniques like watermarking and content fingerprinting.

At DIAGNAL, our mission is to deliver awesome experiences for our customers and to continuously develop ourselves. We have a strong commitment to efficiency, trust and flexibility, and work with major media companies around the world, including Celestial Tiger Entertainment, Intigral, LoungesTV, Optus Sport, WRC Promoter and more. 

To find out more about how we can help, please reach out to us.

Recent Posts

Sky Stream vs EE TV: The Battle for UK Super Aggregation market

This blog will explore the two other strategies, Digital watermarking and Content fingerprinting, which are integral components within the broader landscape of DRM systems and tokenisation, and play pivotal roles in safeguarding intellectual property and ensuring content integrity.

More on OTT Apps

5 Essential Features for OTT Streaming Apps

Maintaining a good user experience A premium streaming app experience is more than just branding, design, usability and function. The key fundamental of creating an high-end user experience is understanding the user and their engagement points within the app. We will talk about the 5 essential product features for OTT apps on the market today and why they are important

Reducing Churn Rate With Offline Playback

The Benefits of having Offline playback for End-users and Content Providers     Introduction Offline Playback is a high-end feature for OTT Applications. Allowing end-users to download DRM/Non DRM videos over WI-FI and/or Mobile Data Networks has been proven to increase user engagement and reduce churn rate for content providers. Whether the user is commuting to work, flying or simply

Subscribe to our Newsletter

Want to keep up to date with all our latest news and information? Enter your email below to be added to our mailing list.