A throwaway comment was tucked away during Faultline’s last call with Viaccess-Orca, when CEO Philippe Léonetti told us that the company’s content security division was not where he wanted it to be in terms of business. Picking up the baton from where we left it nearly three quarters of a year later, we are reliably reassured that the new boss is more pleased with VO’s digital security efforts (pending his official response).
Of course, speaking with representatives from the French vendor’s dedicated anti-piracy team this week, they would say that.
Getting back to business, the real trigger for this catch-up with VO was Synamedia’s recent acquisition of ContentArmor, which prompted Faultline to send probes out into the wild to gauge the sector’s reaction. Mathieu Harel, MD of Anti-Piracy Services at VO, believes the deal is a good omen for the wider watermarking industry, predicting that it will help to spread the word while further shrinking the already tiny pool of watermarking options on the market.
In theory then, Synamedia swallowing ContentArmor should trigger an uptick in business for the likes of VO, although things are seldom this simple.
Back when VO launched its dynamic watermarking technology and had it officially audited by Cartesian some 18 months ago, it was easy to interpret this as a late entrance to the watermarking party. At the time, Nagra, Verimatrix, Irdeto and FriendMTS were already long in the game, while practically everyone else was licensing forensic watermarking patents from either Nagra’s NexGuard or ContentArmor, including VO.
However, years of research and product development on a proprietary dynamic watermarking system in collaboration with French R&D house b<>com, while making sure not to initiate a patent war, allowed VO to be free from its relationship with ContentArmor.
“Creating this technology from scratch helped us address piracy in a way that provides a set of services in a broader anti-piracy center. Watermarking is useless unless paired with other services like our watermarking tracking algorithm,” Harel explained to Faultline.
Importantly, we learned that a specific unconfirmed part of VO’s dynamic watermarking product is patented, but the core technology is not. This may be why Harel was so keen to dismiss the idea of VO forming a licensing business on the side as a complementary revenue stream, to potentially pick up defectors from the ContentArmor camp. The other reason is that VO would see third-party licensing of its watermarking technology as devaluing the overall anti-piracy package that it has worked hard to sculpt.
While we were spared the details, we suspect the patented element is something to do with the way VO’s watermarking algorithms adapt dynamically on the fly to different attacks and recover streams IDs.
“When we acquired the Squadeo OTT player business in 2017, this was using Verimatrix DRM and we carried on offering it,” was the cryptic answer to what VO would do if it was in Synamedia’s shoes, regarding the continuation of ContentArmor’s forensic watermarking licensing agreements. Synamedia told us separately that it plans to honor these, for the time being.
But surely there is an element of frustration among the VO security ranks that it grafted for years to add a purebred watermarking system to its portfolio, only for Synamedia to be carved out of Cisco while the iron was still hot and barge its way into forensic watermarking via M&A? We certainly didn’t sense any such frustration through the medium of Zoom, as Harel reiterated that the deal only confirms VO’s decision to have its own technology in place, in true French style.
As we have highlighted in recent coverage, forensic watermarking is a popular technique for protecting premium streamed content, and with VO being a subsidiary of Orange, it should know a thing or two about operator demands. We noted how, under Synamedia guidance, ContentArmor is targeting opportunities in edge watermarking and 5G, and Harel confirmed – after some encouragement – that VO is seeing similar demands.
Another strong driver is UHD content driven by studios which require watermarking as an extension to DRM for 4K UHD titles, as well as some early release content.
Getting a little more specific, Harel cited contribution watermarking and distribution watermarking as two specific trending techniques, both of which are used in the client-side (or client-composited) approach to embedding imperceptible watermarks into content. VO’s algorithms work on the client/contribution side, which caters for more adaptability as here the watermarks are easier to manage within the CDN. Dynamic watermarking effectively owes its name to the client-side approach, as the underlying machine learning algorithms can behave dynamically to confuse pirates and trigger takedown processes.
Another differentiator is VO’s emphasis on collusion attacks, a content pirate weapon which involves combining media files to produce a new copy without the original information, therefore rendering the watermark unreadable and so detection becomes redundant. This can bamboozle watermarks as well as digital fingerprinting technology to the benefit of pirates, and VO’s dynamic watermarking is tipped as one of the most effective methods out there at halting collusion attacks in their tracks.
Touching on pirate trends during Covid-19, Harel has observed increases in approaches such as adding black bars on top of the content, tilting content, and changing the resolution – which he concedes are not the most advanced techniques, but sometimes the simple things done right can be the most disruptive.
It is regrettable that VO still cannot talk about customers. As always, one can assume that Orange is actively using VO’s dynamic watermarking technology, but where and how the inter-continental tier 1 operator is using this remains top secret. Harel noted there are other customers in Europe and APAC, while the company is targeting Latin America as a key growth area, which is why VO is eager to get a team on the ground for NAB.