Forensic watermarking: Next level content security from Video Toolkit
Cable companies and large streaming services are making huge investments in 4K/UHD content, hoping to corner the market for 4K/UHD TV streaming and differentiate from the competition. Yet studios are nervous at the potential piracy losses from superior quality 4K/UHD content, and are putting content providers under intense pressure to adopt the highest level of content protection.
DRM is a significant part of the content protection mix, yet combined with forensic watermarking offers the next level of protection from video piracy. Now available as part of our Video Toolkit service, session-based forensic video watermarking acts as a deterrent to pirates and enables rights holders to report the misuse of their video assets, and identify the person that leaked the video.
Irremovable and imperceptible marking
Embedded into the video content during the encoding process, the watermark payload is a unique serial number that defines the origin of the content without attaching any personally identifiable data. This mark cannot be manipulated either through encoding, resizing or scaling, or any other video processing technique. While acting as a deterrent to potential criminals, it doesn’t interfere with streaming users, appearing invisible to the naked eye and posing no detrimental effect to the viewing experience.
Verifying content ownership and the pirated source is uncomplicated, with the watermark payload extracted using a specialist cloud based service, possible even from low grade and recompressed video.
Session-based watermarking
Basic watermarking helps identify where a potential content rendition originated from, however, studios may require recognition down to a service’s individual user accounts, or streaming sessions. Session-specific watermarking can achieve this, but it presents a few challenges, including:
- Avoiding creating/storing as many different content versions as there are end-users
- Being able to cache watermarked content on CDNs (as session-specific content versions would become obsolete once a user’s session ended)
- Player integration requirements (traditional solutions usually require content to be modified by a watermarking library within the player)
Our approach to simplify this is to produce two ‘flavors’ of the same content, both using a different mark. We then generate session-specific manifests on the origin server with segment lists consisting of different chunk combinations. In this manner, each chunk is either used from content track A or track B. This enables a binary representation to be used to create unique manifests for each specific session, for example, based on the ID number your system uses to reference users.
Example: Session-ID to binary manifest chunks
| Manifest chunk | 1 | 2 | 3 | 4 | 5 | 6 | 7 | … |
| Content track | B | A | B | B | A | A | B | … |
| Binary session | 1 | 0 | 1 | 1 | 0 | 0 | 1 | … |
Using this method, only a maximum of two content versions are needed to be created and stored in order to support session-level identification which also then allows content to be regularly cached on a CDN. Additionally, there is no need for third party libraries to be integrated on the client-side because the relevant session information is stored within the manifest itself for server-side processing.
For an extra layer of efficiency, we can also even select specific chunk locations to watermark in order to save on content track storage space.
Get started with castLabs Video Toolkit today
In addition to forensic watermarking, our cloud-based Video Toolkit service enables high-quality encoding, encryption, subtitle conversion, and packaging of assets into VOD streams. Find out more by contacting our team to discuss your project needs.
View more blog posts