The Streaming Video Alliance on Tuesday released a new document running down various uses for forensic watermarking as a means of fighting off video piracy.
The paper, "The Forensic Watermarking Implementation Considerations for Streaming Media," was put together by the Alliance's Privacy and Protection Working Group.
"The piracy of online video content is a significant issue for content owners and distributors. Watermarking is a useful tool to combat theft as it permits tracing stolen copies back to the source of the leak," said Jason Thibeault, executive director of the Streaming Video Alliance, in a statement. "Our technical document aims to educate industry members on different watermarking methods, rather than promoting any specific approach, to help identify what best fits their organization's specific media workflows and business needs."
RELATED: Google joins Viacom's Streaming Video Alliance
Although the document was specifically produced by the SVA, a spokesperson said it’s geared toward the video industry as a whole. Although the document doesn’t specifically promote one form of watermarking over another, there are parties that could benefit from more information getting out about forensic watermarking methods. Forensic watermarking technology is vendor-supplied and it would require payment to use. But more premium content owners are starting to require it, an SVA spokesperson said.
The document explores various watermark embedding approaches, such as server-side and client-side, as well as watermarking for on-demand and livestreamed distribution. The purpose of forensic watermarking is to ID leak sources at the distributor level, device level or the subscriber level.
The document also goes into detail on choosing between one-step and two-step watermarking. “While all combinations are in use, selecting one-step vs. two-step watermarking may vary from one deployment to the other and are typically driven by considerations to optimize performance, control over device base, and implementation efforts,” the document reads.
One-step requires access to the uncompressed video where the embedding algorithm can identify locations that allow for robust and invisible modifications. Two-step breaks the process down into two steps: creating variants or pre-watermarked variations of segments, and creating a serialized video bitstream.