by Tim Siglin
February 12, 2009

advertisement

One almost-universal truism in the world of streaming media is that licensing particular technologies can be a confusing and somewhat Byzantine process. MPEG-4 and H.264 licensing, in particular, have received the brunt of the disaffection born from confusion. Equal uncertainty abounds for the proprietary codecs, which use licensing models that are not publicly disclosed. Those who sell proprietary codecs use the perception of H.264 licensing as a labyrinthine ordeal as an ideal marketing tool to complement the proprietary technology’s “simple” license.

Most impacted, though, are the content owners and content distributors who express a common opinion about the licensing process: While necessary, at least as far as they can tell, it can be a convoluted process to even understand the licensing, enough so that it distracts from the real focus of creating, encoding, and distributing content.

Part of this fear, uncertainty, and doubt comes from misinformation about the types of licenses and the thresholds required for such licenses. If you do a Google search, you’ll find many of the top documents on H.264 licensing center on the 2003/2004 time frame, a point at which the license terms were in a state of disarray.

Much of the confusion stems from this era, with some “urban myths” persisting from documents that are readily available online appearing to be part of the legitimate final license issue. To better understand the issues, let’s first take a look at the H.264 licensing entity.

MPEG LA
MPEG LA bundles license portfolios together for ease of licensing. The company’s most famous license portfolio was for MPEG-2, which was used in everything from DVD players to satellite television boxes. When the MPEG-4 portfolio appeared, though, the MPEG LA consortium didn’t come out of the gate with the same clarity of purpose and transparency of licensing terms as it had with MPEG-2.

MPEG LA’s H.264 licensing approach learned a bit from the MPEG-4 debacle, but it still didn’t reach the ease of understanding it had with MPEG-2. The fault is not all MPEG LA’s, as some patent holders disliked the way that their parts of MPEG-2 were handled, and they subsequently made it more difficult to aggregate the block of licenses together into a single portfolio. Having said that, MPEG LA did expand some of its licensing reach (we’ll talk about one license specifically) when it came to final licensing, which also caused a bit of industry consternation.

Remembering that the H.264 codec can be scaled down to mobile, scaled up to high-definition, and used as a low-latency messaging, videoconferencing, or broadcast-delivery methodology, let’s shift now to the licensing vehicle. We’ll look briefly at an overview of the final H.264 licensing terms from MPEG LA, based on a summary document provided at www.mpegla.com/ avc/AVC_TermsSummary.pdf.

MPEG LA splits the H.264 license portfolio into two sublicenses: one for manufacturers of encoders or decoders and the other for distributors of content. The biggest uncertainty comes from the distribution side. Since there are very few companies that manufacture codecs, we’ll focus strictly on the distribution side, where content is encoded, transmitted to the end viewer, and then decoded for viewing.

Licensing Categories
The sublicense on the distribution side gets further split out to four key subcategories, two of which (subscription and title-by-title purchase or paid use) are tied to whether the end user pays directly for video services, and two of which (“free” television and internet broadcast) are tied to remuneration from sources other than the end viewer.

‘Free’ Television
Looking at the sublicenses from last to first, the “free” television broadcasting approach has caused particular uncertainty—great enough that it led, in part, to a suspension of some fees until 2010.

The back story is that the “free television” term in MPEG LA draft licensing was limited to traditional television broadcasts known as over-the-air (OTA) broadcasts. While covering a specific segment of broadcast, allowing broadcasters to use H.264 video transmissions within the MPEG-2 transport stream in a technical backward-compatibility move that might have warranted the licensing fee, the patent holders decided that only limited revenues would be derived from OTA broadcast transmissions.

MPEG LA subsequently broadened the scope of the “free television” term.

“Formerly limited only to over-the-air broadcasting, ‘Free Television AVC Video’ now refers to AVC Video that constitutes television broadcasting which is sent by an over-the-air, satellite and/or cable Transmission,” the company said in a May 2004 release announcing final licensing terms. “The party (e.g., a broadcaster) which is identified as providing Free Television AVC Video service is the Licensee and assumes responsibility for the applicable royalties.”

This point is important because it further solidifies the approach MPEG LA uses for “participation fees” that must be paid on indirect revenue based, in some part, on the use of H.264 decoding.

Like a state tax structure where a sales tax is counterbalanced by a use tax, MPEG LA's participation fees act in a similar manner and make the broadcaster liable for royalties or licensing fees. In terms of the state, the use tax means that some purchases from out of state that have not been charged sales tax will be liable for an equal use tax as a way to close a loophole in interstate commerce, meaning that any use of a taxable product purchased out of state is subject to the use tax upon first use. In terms of MPEG LA, these participation fees require content distribution companies to pay a fee if they make money on content distributed as H.264.

The licensing fee for “free” television is based on one of two royalty options. The first is a one-time payment of $2,500 per AVC transmission encoder, which covers one AVC encoder “used by or on behalf of a Licensee in transmitting AVC video to the End User,” who will decode and view it. If you’re wondering whether this is a double charge, the answer is yes: A license fee has already been charged to the encoder manufacturer, and the broadcaster will in turn pay one of the two royalty options.

“The AVC License separately grants (a) a right to manufacturers to make and sell AVC products,” said Tom O’Reilly, MPEG LA’s manager of research and public relations, “and (b) a right to content providers to use such AVC products to encode and decode AVC Video. A royalty applies to each.”

O’Reilly noted that MPEG LA’s intent with these two royalties is to spread the charges across multiple parties.

“Rather than place the total royalty on one party in the AVC product chain (e.g., an encoder maker),” said O’Reilly, “royalties are aligned with different points in the product/service chain where value is received.”

The second licensing fee is an annual broadcast fee, per “broadcast market.” This one is interesting, especially given the expansion of the terminology of “free television” to include satellite and cable. MPEG LA’s licensing terms define a broadcast market as “a geographic area within which an End User could use an AVC Decoder to view Free Television AVC Video sent by a single transmitter or transmitter simultaneously with repeaters by a single Legal Entity.” OTA appears to be at somewhat of a disadvantage here, as the signal for a geographic OTA area would be minuscule compared to the geographic area for cable and/or satellite transmission.

Regardless of the technological questions that arise from the broadcast-market scenario, the annual broadcast fee is broken down by viewership sizes:
—$2,500 per calendar year per broadcast markets of 100,000–499,999 television households
—$5,000 per calendar year per broadcast market of 500,000–999,999 television households
—$10,000 per calendar year per broadcast market of 1,000,000 or more television households

Some might ask, “Why not just pay the per-encoder fee?” This might be the most feasible model for a cable or satellite delivery system, as the content could be encoded at a centralized location, multiplexed, and then transmitted across the network to multiple geographies within a single broadcast market. But for local OTA, there may be multiple encoders involved. In fact, for those who already have the digital OTA converter boxes, take note of the fact that many broadcast stations are sending out three or four signals per single traditional analog signal.

The other two issues here, somewhat addressed by MPEG LA, are the question of developing countries’ ability to maintain the fee structure and the corollary of countries that choose to assess a viewership tax on their television sets. Countries across Europe assess a viewership fee for OTA, and it brings up an as-yet-unanswered question.

MPEG LA’s approach on the issues of the developing countries is to footnote the issue, stating, “It is recognized that broadcasters in developing countries may have different circumstances to be considered.”

Grazing the edge of the television viewership tax, MPEG LA also notes: “There may be AVC Video services that qualify as Free Television AVC Video rather than Subscription AVC Video or Title-by-Title AVC Video where a nominal (but no other) payment is required or where other relevant circumstances taking into account the nature of the video service may apply.”

While both of these may mean a case-by-case basis, no details have been published.

Internet Broadcast
With all the issues around “free” television, why should someone involved in nonbroadcast delivery care? As I mentioned before, the participation fees apply to any delivery of content. After defining that “free” television meant more than just OTA, MPEG LA went on to define participation fees for internet broadcasting as “AVC video that is delivered via the Worldwide Internet to an end user for which the end user does not pay remuneration for the right to receive or view.” In other words, any public broadcast, whether it is OTA, cable, satellite, or the internet, is subject to participation fees.

This is a big deal for those whose primary focus is distributing content on the web. It would be quite fair to say that those companies choosing the “give away content for free in hopes of getting advertising” business model together might want to think through their strategies.

Careful consideration should yield models that would work, but the point is this: Content encoded, distributed, and viewed by end users is subject to the participation fee, even if the viewership is advertising-driven or sponsorship-driven. There’s even a grey line around the use of content to drive donations, as it could be argued that these participation fees are due since the content was a compelling part of the donor’s decision to fund the not-for-profit.

The fees are potentially somewhat steeper for internet broadcasts, perhaps assuming that internet delivery will grow much faster than OTA or “free” television via cable or satellite. Adding the “free television” broadcast-market fee together with an additional fee, MPEG LA grants a reprieve of sorts during the first license term, which ends on Dec. 31, 2010, and notes that “after the first term the royalty shall be no more than the economic equivalent of royalties payable during the same time for free television.”

The biggest potential long-term pitfall here lies, again, with the definition of the “broadcast market,” as MPEG LA and patent holders could argue that the geographic boundaries of an OTA local market would constitute the same. The near-term pitfall, footnoted in the MPEG LA summary document is that, since no royalties are charged during the first term of the license for internet broadcast, “royalties for the license for Internet Broadcast AVC Video Use granted pursuant to Section 2.5 are not subject to this limitation.”

In other words, as section 2.5 deals with internet broadcasts, there’s more confusion in store. To add to the confusion, readers of the summary document are recommended to “see footnote 23” in the summary document, which only contains 17 footnotes.

Title by Title
Beyond participation fees for indirect revenue (revenue not directly from the user), MPEG LA also sets out amounts for title-by-title (rental or per-view). For videos less than 12 minutes long, there is no royalty; but for videos beyond 12 minutes in length, the amounts are decided at 2% of the retail price paid to the licensee or 2 cents per title. The retail price is specifically noted as a “first arm’s length” transaction, specifically between the end user and the seller of on-demand, pay-per-view, and electronic downloads to end users.

Provisions are also made for a category of licensee that replicates physical media, in which case it appears the 2% of the amount paid would be based on wholesale prices for the “first arms-length” transaction rather than retail pricing. Confused yet?

Subscription
Where an end user chooses a subscription service that does not charge by title or a per-order basis, legal entities with fewer than 100,000 subscribers per year pay no royalties, although there is no definitive answer on two key questions. The first is whether “first year” refers to calendar year or rolling annual year, while the second deals with the question of the “fewer than 100,000 subscribers” being simultaneous or aggregate users. The latter is important information for a company that may, perhaps, have 50,000 subscribers at any given time but may experience churn greater than two times per year.

For any subscriber based above 100,000, the following annual fees apply:
—$25,000 for 100,000–250,000 subscribers
—$50,000 for 250,000–500,000 subscribers
—$75,000 for 500,000–1,000,000 subscribers
—$100,000 for greater than 1,000,000 subscribers

As it is possible for a company, or a group of commonly controlled companies, to have multiple business models—from indirect ad-based revenues to subscription or pay-per-view—MPEG LA also has set a maximum participation fee. It bases this maximum fee cap per enterprise. On an annual basis, these fees cannot exceed $4.25 million per year in 2008–2009, increasing to $5 million per year in 2010. Interestingly, MPEG LA calls out that fees cannot increase by more than 10% per year, but the bump from 2008–2009 to 2010 is almost a 20% annual increase.

Summary
In conclusion, the MPEG LA aggregation of a patent bundle into a licensing portfolio is an attempt to make the licensing process easier for the user or manufacturer of an H.264 encoding or decoding system. While the process is not simple to understand, a complaint often lodged against the H.264 licensing process by competing codec solutions, it is much easier than trying to gather licenses from a large number of patent holders. While the process of figuring out which areas are applicable might be tricky, the fees themselves are easily calculable, with the exception of a few questions we’ve noted throughout this document.

Another key point to remember is that a licensing portfolio can be modified at any moment, and the licensing structure is very much a “living document” in the sense that patent holders may opt to use a different approach or market forces may shift the terms of the licensing to meet economic or growth realities. As none of these participation fees for “free” television and internet delivery have been adjudicated, over time there may be a shift in licensing structures as someone may challenge the current structure of the dual royalty charges or even on the more nebulous definition of the “broadcast market” term. It is the responsibility of MPEG LA (and its patent holder pool) to better define several of these areas, especially when other codec options exist.

In the meantime, two things will continue to occur. First, proprietary codec manufacturers will continue to market the fact that they charge no decoding license fee. Ignoring the fact that they do charge an encoding license fee that may or may not be equal to the total fee charged for encoding and decoding of H.264, the “no decoding license fee” marketing strategy has struck a chord that MPEG LA and its patent partners might need to further address.

Second, objections from nonlicensees will continue, but for reasons other than the actual licensing fee structure. Given the known cost of a standards-based codec, unlike the unknown cost of a proprietary codec, it appears that many of the criticisms leveled at H.264 licensing are based on critics being upset about paying a known amount for something they equate to open source. Mistaking standards-based for open source, they act on the first part of a famous quote (“data wants to be free”), forgetting the second part of the same quote: “and data wants to be very expensive.”

In the end, for a fair comparison of licensing fees, it is up to the proprietary codec creators to lay the licensing cards on the table rather than using the fact that MPEG LA does so as a means to spread fear, uncertainty, and doubt.

“Patent holders enable MPEG LA to offer a valuable service to both Licensors and Licensees,” said MPEG LA’s O’Reilly. “We daresay the market is better off for having this opportunity than not, and based on the success of the AVC License program (23 Licensors and more than 550 Licensees), the market seems to agree.”