From broadcast flags to the analog hole
Note: This is Part II in a series; Part I is here, Part III is here.
Last week, I began the story of the Broadcast Flag, a law that would make it illegal to build a general-purpose computer unless it conformed to a set of privately negotiated restrictions. The law had been promised by Billy Tauzin, then a lavishly corrupt Congressman, and its contours were being hammered out in an inter-industry body called the Broadcast Protection Discussion Group (BPDG), convened by the MPAA and attended by movie studios, TV studios, broadcasters, consumer electronics companies, and PC companies.
All of this was in the name of hastening America’s “digital TV transition,” the difficult business of convincing Americans to replace their old analog TVs (and VCRs and other peripherals) with new digital versions. The transition was particularly politically fraught because broadcast TV watchers were overwhelmingly drawn from the cohort of older Americans, who are also the country’s most avid voters, and who are also notoriously uninterested in keeping up with the latest gadgets.
Further complicating matters: US electromagnetic spectrum rules prohibit scrambling over-the-air broadcasts. The movie studios swore that they would not permit their catalogs to be aired unless some way to “protect” them was found (they carefully did not promise that they would ever authorize broadcasts of movies, even if “protection” was delivered).
Hence, the Broadcast Flag: a rule that said that any device capable of tuning or demodulating a digital TV signal would have to incorporate restrictions codified by the BPDG. These rules would govern how files could be written and retrieved by the system, and what kind of data and audiovisual outputs could be accessed under which conditions. Digital TV programming would be restricted to digital outputs that had DRM on them.
Which DRM? Well, that was the point, really. The BPDG’s notional task was to define the technical characteristics of this mandatory DRM. Hence the presence of the CE and tech companies: each was there to make sure that whatever technical specifications the group came up with wouldn’t exclude their own proprietary DRM.
I attended these meetings on behalf of the Electronic Frontier Foundation (EFF), and our client, the GNU Radio Project, a free/open software-defined radio. Any computer capable of running our client’s code would come under the scope of the regulation, and every computer could run our client’s code.
In other words, we were there to point out the inconvenient truth that the Broadcast Flag wasn’t a rule restricting how digital TV equipment worked — it was a restriction on all general-purpose computers.
Our client certainly had an interest in the outcome of the Broadcast Flag negotiations, since one iron-clad starting principle was that all compliant devices would have to be “tamper-resistant.” Any device that could be modified by its owner would be banned under the rule — which meant that all free/open software would be snuffed out by the rule.
With that background, let’s carry on with the tale.
The BPDG was a creature of the MPAA, an organization that had been waging war on tech and CE companies for decades. The MPAA and its allies from broadcast and TV production were highly integrated, with interlocking owners and boards, and they had undergone waves of mergers and acquisitions that had reduced their number to a handful. The resulting cartel had a finely honed lobbying machine, pushing a common agenda in Congress, at federal agencies, in state legislatures, and in international bodies like the UN’s World Intellectual Property Organization (WIPO).
By contrast, the CE and tech sectors were still made up of hundreds of companies that were still reeling from the dot-com crash. They were constantly at each others’ throats, with a strategy that mixed convincing new users to get online, and poaching one another’s customers with direct attacks on their rivals’ products (for example, Apple’s iWork suite, which let Microsoft users switch to the Mac while preserving their documents and their ability to collaborate with their Windows-using colleagues).
As much as CE and tech companies mistrusted one another, they mistrusted the entertainment companies even more. It’s hard to imagine their being lured to the BPDG to lend it their legitimacy, save for the presence of Intel.
Intel was the MPAA’s partner in designing the MPAA’s Broadcast Flag. Intel was a trusted broker in the tech world, providing silicon to all parties. Intel provided the group’s technology chair and set its agenda. That was the bait that drew in the tech and CE companies.
What these companies missed is that Intel had ambitions beyond providing commodity silicon to different companies. Just a few years before, Intel had lured reps from the burgeoning peer-to-peer software company to a new industry body, the Peer-to-Peer Working Group. I was at that meeting, too, as co-founder of Opencola, a P2P search company. Halfway through the meeting, Intel revealed its master plan: to organize the P2P sector under the directorship of a handful of incumbent tech giants (“gold members”) who would give marching orders to the rest of us. Our own steep membership fees would entitle us to receive these marching orders and discuss them amongst ourselves, but not to have a say in them.
Which is to say, the Intel of the early 2000s had serious ambitions: not just to supply the burgeoning networked tech industry, but to structure it.
Intel was a natural choice for chair of the BPDG. They were already a dominant player in the two leading DRM consortia, the “5C” group (DTCP), and the “4C” group (CPPM/CPRM). They owned patents in each group’s patent-pools and received a royalty every time a device containing either got made. They were also the senior member of each consortium, so each was guaranteed to act in ways that pleased Intel’s shareholders.
Intel’s implicit pitch to the other tech companies: if you don’t come to this meeting, we’re going to standardize the 4C and 5C technologies as mandatory for all devices, and you’ll have to pay us a royalty every time you make any kind of device. If you do show up, you can ensure that the criteria are broad enough to incorporate your own DRM.
Before EFF got to the BPDG, the group’s foundational consensus was that every computer should include some form of mandatory DRM, and the only controversy was: “What technical criteria does a DRM need to satisfy in order to be included in the Broadcast Flag rule.”
EFF — and the consumer groups that joined us later, like the Center for Democracy and Technology — raised a separate, even more foundational issue: Why do we need DRM at all? And, secondarily, Would this DRM even work, or would it trigger a cascade of even further-reaching mandates to achieve its goal of preventing redistribution of high-definition broadcasts?
Recall that the BPDG meetings were “public” in the sense that anyone who knew about them could show up, pay a $100 admission fee, and sit in. The schemers who masterminded the Broadcast Flag anticipated that they would face criticism of conspiring in secret to rig the market, so they instituted this open-door policy.
At the same time, the convenors didn’t actually want the public there. Their project was so obviously shady that any scrutiny could spook lawmakers and keep it from being formalized through an act of Congress. So these “public” meetings were also secret: there was no website for the BPDG, and there was no notice of its meetings. The only way to join the BPDG was to be invited.
This is “security through obscurity” at its worst. When an EFF supporter stumbled on the “open” (but unlisted) midnight meeting at NAB where the BPDG was announced, we were able to foil the group’s secrecy, simply by publishing the existence of upcoming meetings and minutes from the previous ones. We started a Typepad blog called “Consensus At Lawyerpoint” where we did just that. The doors blew open.
The Broadcast Flag process was not a smooth one. With the public in attendance and our detailed notes on each meeting going live within hours of its conclusion, it became exponentially harder to convince the tech and CE companies to go along with the gag.
Soon, movie studios and tech giants like Microsoft started to unilaterally announce private “subcommittees” whose meetings would take place behind closed doors. Every time a CE or tech company balked at some element of the discussion, one of these subcommittees would declare a meeting and the whole process would grind to a halt while they dragged the refusenik into another board-room in the LAX Four Points Sheraton. Minutes or hours later, the “subcommittee” would return to the main room to tell us that the difference had been “resolved.”
Each meeting was peppered with warnings that Congressman Tauzin, the MPAA’s patron on The Hill, was growing impatient with our foot-dragging and might soon lose interest in the project, or simply impose some half-baked proposal cooked up by his staffers. The MPAA reps wielded this threat as a goad to keep the process moving, despite the mounting drumbeat of bad press and public suspicion.
Strangely, in light of this urgency, most of these meetings were taken up with minutiae. Meanwhile, the actual technical criteria for acceptable DRM was relegated to an appendix, dubbed “Table A,” which was kicked further down the road with each meeting. Table A was the only thing any of the tech companies cared about: its criteria would determine whether their products could be freely made and sold, or whether they would have to license DRM from one of their competitors.
It should have been obvious that there was some funny business going on with Table A, but even I was shocked when Intel, the broadcasters, and the studios announced their proposal:
There are no objective criteria for which technologies are acceptable and which ones are not. Rather, inclusion on Table A will be determined by approval from a quorum of TV and film studios. The only technologies that currently have that approval are the 4C and 5C technologies — the ones that Intel gets a royalty from.
This certainly roiled the attendees! It was wild to see the expressions on the faces of the reps from big tech companies like Apple and Microsoft, as well as their smaller startup competitors when they realized they’d been played by Intel. Their presence had legitimated a document proposing criminalizing making future products without paying Intel for the privilege!
Subsequent meetings and calls got very heated. Intel and its co-chairs tried to whip the attendees into signing onto a “final report,” but the document ended up being dominated by sprawling appendices with the individual, sectoral, and corporate critiques of many of the participants.
The plan started to unravel. Billy Tauzin’s political career was in a nosedive (he declined to run again in 2004, and instead took up a lucrative career as chief lobbyist for PhRMA, at $2,000,000/year).
With Tauzin’s attentions elsewhere, the MPAA abandoned a Congressional route to a Broadcast Flag mandate. Instead, they turned to GW Bush’s FCC chairman, Michael Powell (son of Colin Powell). Chairman Powell — who would graduate from regulating the telecoms industry to representing it as president of National Cable & Telecommunications Association (NCTA) — was an enthusiastic champion of the Broadcast Flag. He understood that delivering the Broadcast Flag would allow the FCC to claw back the analog broadcast spectrum and then auction it off for billions to mobile carriers.
The FCC eventually adopted the Broadcast Flag rule in substantially the same form as the final BPDG report, taking little notice of the extensive objections in the appendices. Thankfully, EFF understood the Broadcast Flag inside and out, and we were able to successfully challenge the FCC’s authority to regulate the design of all general-purpose PCs. In 2005, the DC Circuit Court of Appeals unanimously threw out the Broadcast Flag rule.
Remember when I said that one of the many objections EFF brought to the BPDG was that the group’s work couldn’t possibly solve the problem they said they were there to fix?
The Broadcast Flag was always going to be a terrible way to prevent the unauthorized capture and redistribution of digital TV broadcasts, for a whole host of reasons, but the principle among them was redigitization.
You see, the Broadcast Flag only regulated the digital processing of DTV broadcasts. There were already millions of DTV receivers in the market that had high-resolution analog outputs that could be plugged straight into a video-capture card and turned back into a digital file with almost no signal loss. What’s more, anyone with a high-resolution digital video camera could position a 4k DTV screen in a perfectly configured screening room, lock off the camera on a tripod, and record a nearly indistinguishable copy.
It turns out the MPAA was alive to this problem. They even had a catchy name for it: “The Analog Hole” (no, seriously). The Analog Hole problem was that fundamentally, your head is full of analog holes, your eyes, and ears. In order for you to experience digital media, it would have to be converted to analog — pictures, and sound — so that it could be beamed into those analog holes of yours.
At the height of the BPDG negotiations, the MPAA’s Copy Protection Technical Working Group set up another subcommittee, the Analog Reconversion Discussion Group (ARDG), whose meetings took place immediately after the BPDG meetings, in the same LAX Four Points Sheraton.
If the BPDG was far-fetched, the ARDG’s work was even weirder.
The group’s mission was to come up with a watermarking system that would imperceptibly embed images and sounds into video and audio streams. Then they would make it illegal to produce any kind of “analog-to-digital convertor” (including, but not limited to, digital audio recorders, digital video recorders, and digital still cameras) unless they checked for this watermark. If they detected the watermark, they would switch themselves off.
This idea was simultaneously technologically audacious and nonsensical. Let’s break that down.
- First, it assumed that there was such a thing as a watermark that was both robust (hard to remove) and imperceptible (invisible and inaudible —after all, no one wants to pay for movies where the picture is marred by visible “DO NOT COPY” wordmarks). These are fundamentally irreconcilable goals. By definition, if a watermark adds no perceptible information to a signal, then removing that watermark will make no perceptible difference to it.
- Second, it treated analog-to-digital converter (ADC) mandates as easy-to-administer regulations, implying that ADCs are esoteric, specialized equipment. The reality is that ADCs are commodity components, frequently homebrewed by high-schoolers for science fairs. Regulating an ADC isn’t like regulating a car — it’s more like regulating the wheel itself.
- Finally, it ignored the second-order consequences, both intentional and unintentional: a watermark detector can’t tell whether it’s stopping an act of piracy or an incidental inclusion of a copyrighted work, whether that’s your attempt to record your toddler’s adorable dance to Price’s “Let’s Go Crazy,” or a cop blasting Taylor Swift to deliberately trigger a copyright takedown.
As with the Broadcast Flag, “plugging the analog hole” wasn’t really a technical proposal, it was a legal one — a rule aimed at structuring whole industries. The idea wasn’t to prevent piracy — it was to use the pretense of fighting piracy to allow the studios to decide which companies could make computers, and how those computers would act.
And as with the Broadcast Flag, the Analog Hole had champions in Congress: in 2005, Reps. Sensenbrenner and Conyers introduced the Digital Transition Content Security Act, AKA the “Analog Hole Bill.”
This bill would have required all devices capable of performing analog-to-digital conversions to license a technology called VEIL, which produced watermarks that were billed as both imperceptible and robust.
VEIL didn’t work, of course. Its only real-world deployment was in a Batmobile toy that would respond to audio watermarks in cartoons by racing around your living room. No one was allowed to test whether these watermarks could be forged, altered or removed, because VEIL was secret, and implementers had to pay $1,000 just to get a look at its specifications, agreeing to nondisclosure agreements in the process.
But again, the point wasn’t to prevent piracy — it was to control the legal market. A VEIL mandate would let the studios decide who could make products, and how those products could work.
At the time, many people — myself included — didn’t grasp the scope of this vision. The studios’ plans may have appeared to be a jumble of half-baked, technologically illiterate fantasies, but that’s only true if you assume that the point was to control technology, rather than the industry. The real play was far more ambitious.
Next week, I’ll wrap this series up with a discussion of the Authorized Domain, an attempt to convert this kind of legal power into the power to define the very structures of society itself, down to the definition of the family.