Judge invokes DMCA in upholding ban on RealDVD

The technology used to thwart DVD ripping software such as Slysoft's AnyDVD includes mechanisms that place intentional errors and even false logic, such as navigation menus that lead nowhere, on studios' DVDs. Ordinary DVD players would ignore this false information, but rippers may copy it and, in producing better images of it in the copy, produce DVDs with errors that ordinary players would not ignore. It's this technology which RealNetworks engineers actively worked to overcome, in their creation of a system enabling owners of DVD movies to create archival backups of their collections onto hard disk drives.

Yesterday, US District Court Judge Marilyn Hall Patel ruled that this act -- the creation of error-correcting code that does not discriminate between accidental errors such as scratches, and intentional errors used for copy protection -- is a violation of the Digital Millennium Copyright Act. This in her decision yesterday which upholds a lower court ban on the sale of RealDVD, and future Real products based on that product, imposed last October.

"The intentional, calculated sector and file-structure errors that are created by ARccOS and RipGuard as part of their copy protection systems [used by major studios' DVDs] present different characteristics from physical scratches. A typical scratch or smudge would not have the numbers of bad read errors in a row, i.e., sequentially from when reading a disk, that an ARccOS or RipGuard error does, because scratches are typically perpendicular to the path of data that spirals radially around the DVD (much like the grooves on a vinyl record)," reads Judge Patel's decision yesterday. "These differences would be distinguishable by a software program that copies DVD content."

By Real's own testimony in the case, its engineers sought legal and defensible means to thwart these two brand-name intentional error-generating mechanisms, ARccOS and RipGuard, in order to build a DVD archiving system. The initial project for Windows-based software for home media computers was code-named "Vegas" (named for the phrase, "What happens in Vegas stays in Vegas"); while a future edition using a branded hardware component containing a hard drive, was called "Facet." "The Real software engineering team... spent over a year creating software that would ensure that Facet never encountered ARccOS or RipGuard errors as it made a copy," according to the decision.

But Facet used a unique error-correcting method that went over and above Vegas, and could be construed as a next-generation copy-protection defeat system. It went so far as to create a virtual machine in the computer's memory, that would pretend to play a DVD just like an ordinary DVD player, but whose internal clock was intentionally accelerated. To the VM, it would be "playing" the DVD at normal speed; in the real world, it was copying the DVD at an accelerated rate. The process was called "DVD Walk."

Because DVD Walk could copy anything it played at the speed it created for itself, it had no reason to try to detect whether any errors it encountered were accidental or intentional...because it wouldn't encounter them. "Facet switches to DVD Walk regardless of the cause of the errors it encounters. There is no explicit mechanism or code in Facet to identify ARccOS or RipGuard errors with certainty; Facet only recognizes sector errors, whether intentional or inadvertent, such as from a scratch."

Yet when Facet did encounter those sector errors, it managed those errors differently than a normal DVD player would, since its intention was, after all, to make a copy. Effectively, it did a better job of error-correction than the error-correcting code that all DVD players have, and it did all this while still avoiding the low-level error detection that would have red-flagged an error as an intentional one. In other words, there are ways to spot an ARccOS or RipGuard error; by not doing so, Real was in violation.

What's more, the judge ruled, since the copy created by Facet or Vegas did not include all of the CSS copy protection mechanisms that the original DVD contained -- by virtue of the fact that 1) Real wasn't licensed to reproduce those mechanisms anyway; and 2) the copy itself was not on a DVD but a hard drive -- RealDVD constituted a circumvention of those mechanisms. Changing the form factor of the copy, thus, can be construed as circumvention.

"Real has violated section 1201(a) of the DMCA by trafficking in RealDVD products for the purpose of circumventing CSS technology. Specifically, RealDVD's 'play and save' feature, which allows the user to view the DVD content while RealDVD makes a copy of the content on the device's hard drive, or its 'save' feature, which commands RealDVD to simply copy the content from the DVD onto the device's hard drive so that it can be played later, both circumvent CSS technology by permitting RealDVD to access DVD content from the hard drive without going through most of the CSS protection steps, such as DVD drive-locking, CSS authentication and CSS bus encryption. Once RealDVD has copied a DVD, it does not authenticate the DVD drive or receive encrypted keys at that time. The record shows that CSS technology requires that a DVD drive 'lock' upon insertion of a CSS-protected DVD and prevent access to its contents until an CSS authorized player engages in an authentication procedure, none of which RealDVD does when it reads back DVD content from its hard drive. The process of authentication with the DVD drive, and subsequent content decryption, is thereby circumvented by the RealDVD products."

One final strike against Real in Judge Patel's decision concerns the fact that Real was not licensed to reproduce CSS technology. The fact that it did not reproduce certain key elements of that technology anyway (see above paragraph) did not serve in Real's defense. Real's argument was that a licensed owner of a DVD was already authenticated to use that DVD by the copy protection mechanisms that were on the DVD, so "once is enough" -- once copying the authentic DVD to a source owned by the authentic user, the same authentication was not required again. Patel saw that as a technical argument that was not borne out by the technology itself -- effectively, CSS was designed to keep authenticating licensed users, and thus that's how it should work.

So Real should have obtained a license, the logic goes on, to have effectively authenticated the user for the purposes which the original DVD's technology authenticated the user for, even though the complete authentication suite was missing from the copy (thus, the act of circumvention). But as the decision itself continues, the DVD CCA licensing authority probably would not have licensed Real to use CSS for the purposes of making a copy anyway. So while DVD CCA and Real did make a good faith agreement, as the judge found, "That the Agreement [between Real and DVD CCA] does not explicitly ban the order in which Real copies DVD content onto a hard drive has no force from a legal standpoint, given that the step of playing back DVD content from a RealDVD hard drive itself violates the Agreement, both expressly and impliedly."

The precedent set by Judge Patel's order may be profound: If a company that makes a good-faith agreement with the DVD CCA licensing authority uses the technology it's authorized to use, for purposes which probably would not have been sanctioned, it may be in violation of the DMCA.