Your browser is not supported. Please update it.

5 December 2019

VO bites back – promising early 2020 dynamic watermarking launch

Studios and broadcasters had a strong presence at Video Exchange in London this week, meaning piracy was on everyone’s lips. Yet while many speakers warned of ballooning revenue losses at the hands of content pirates, very few were willing to disclose their specific anti-piracy strategies. Thankfully, a few specialist security technology vendors were on the hunt for customer prey.


Viaccess-Orca was eager to clear the air after a webinar last week left Faultline scratching its head about VO’s role in an evolving security world. Dynamic watermarking was the French firm’s term of choice to sweeten us up, promising that a new product launch is penciled in for early next year.


“It has taken a long time to develop our own watermarking system, but we have learned from the market,” Kevin Le Jannic, VO’s Director of Security, explained to Faultline in London. Or rather, VO was caught napping.


Forming part of VO’s broader anti-piracy initiative launched officially around IBC, we were told this newly developed dynamic watermarking technology has specific algorithms for specific families of attack, which can dynamically change based on the flavor of attack, to confuse pirates and spearhead a takedown process.


Le Jannic gave the example of collusion attacks, a technique harnessed by content pirates to bamboozle watermarks as well as digital fingerprinting technology. Collusion attacks basically combine media files to produce a new copy without the original information, therefore rendering the watermark unreadable and so detection redundant.


The way we see it, anti-piracy strategies are heavily reliant on watermarking and we have yet to see another technology enable stream source identification. Le Jannic disagreed about watermarking being the only real takedown tool, taking the opportunity to upsell his company’s wider security services. He could not comment on whether VO’s rivals are already offering so-called dynamic watermarking and therefore if VO is late to the party. It certainly sounds like technology that majors like Nagra and Synamedia already have in the locker, and we have observed such a system at Friend MTS, although we’ll have to wait until the official launch next year to get our hands on the juicy details.


As for VO’s breach detection techniques, Le Jannic said these algorithms are still a work in progress with more PoCs, but unfortunately nothing the company can disclose just yet. VO’s detects breaches by monitoring of illegal streams from where it surfaces anomalies to customers, allowing them to revoke devices and it can even help customers take pirates to court in some cases. It also emphasizes the crucial nature of upselling, which Viaccess-Orca reckons could open the door to a new way of selling content and a great way of justifying security purchases.


VO is also making use of existing resources, essentially reusing things like an operator’s network logs to feed into machine learning algorithms and ultimately help develop anomaly detection techniques. Le Jannic highlighted that pirates themselves are now using techniques they know can make anti-piracy machine learning algorithms see or not see certain things, like watermarks, and so he underscored the importance continually adapting.


Last week, Faultline picked up on a comment from VO inadvertently suggesting that anti-piracy was easy. Of course, while it might appear easy to a security expert, VO wanted to assure us the reality is world’s away from easy – and put the slip of the tongue down to translation trouble.