Video watermarking involves overlay compositing, a fundamental operation in video production where a foreground element (the watermark) is blended on top of the background video frames. The technical process centers on alpha blending, a mathematical operation that combines pixel colors based on transparency values to produce a natural-looking composite.

Alpha blending uses the formula: output = foreground * alpha + background * (1 - alpha), applied independently to each color channel (red, green, blue) of every pixel. The alpha value ranges from 0 (fully transparent) to 1 (fully opaque). When a watermark logo has an alpha channel (as PNG images with transparency do), each pixel of the watermark carries its own alpha value, allowing smooth antialiased edges that blend naturally into the video. Without an alpha channel, the watermark appears as a rectangular block with hard edges, which looks unprofessional. This is why PNG logos with transparent backgrounds are strongly preferred for video watermarking.

There are two fundamentally different approaches to video watermarking: burn-in watermarks and metadata watermarks. Burn-in watermarks (also called hard watermarks) permanently modify the pixel data of each video frame, compositing the watermark directly into the image. Once encoded, the watermark is inseparable from the video content. This is the approach used by visible watermarking tools and provides the most practical protection against unauthorized use, since removing a burned-in watermark requires sophisticated inpainting algorithms that leave visible artifacts.

Metadata watermarks (also called soft watermarks) add the watermark as a separate overlay layer or track within the container format. Some professional video players can display these overlays during playback, but they can be trivially removed by stripping the metadata track or using a player that ignores overlay instructions. Metadata watermarks are useful for draft review workflows where the watermark needs to be removed later, but they provide no real protection against content theft.

In the field of video forensics, invisible watermarking is an advanced technique used by studios and broadcasters to track content distribution. These watermarks embed imperceptible patterns into the video frames, often by making tiny modifications to the least significant bits of pixel values or by subtly altering DCT coefficients during encoding. Forensic watermarks survive common transformations like re-encoding, cropping, and scaling, allowing content owners to trace the specific copy of a video that was leaked. This is how studios identify the source of pirated pre-release screeners. While invisible watermarking is beyond the scope of consumer tools, understanding its existence highlights the spectrum of watermarking techniques from simple visible overlays to sophisticated forensic tracking systems, each serving different purposes in content protection and rights management.