If Nokia’s standards infringement lawsuit prevails over Apple, the costs could be steep to any company relying on H.264 or even HEVC. Here are the issues at stake.
On December 21, 2016, Nokia sued Apple for infringing eight patents related to H.264 encoding and decoding. By its terms, the complaint makes clear that Apple’s usage of H.264 is generic, and that similar infringement claims could be made against any products with an H.264 encoder or decoder without a license with Nokia. Though a quick glance at prior cases make the stakes appear minor, a more reasoned analysis leads to the conclusion that the costs to Apple, and others using H.264, could be very significant.
By way of background, this suit is only one in a flurry between the parties. According to a Nokia press release, “Across actions in 11 countries, there are now 40 patents in suit, which cover technologies such as display, user interface, software, antenna, chipsets, and video coding.” For its part, Apple is suing Nokia and related parties for antitrust, essentially alleging that Nokia is attempting to extract excessive fees from Apple. Obviously, the H.264 related suit is most pressing for the streaming industry.
Reasonable and Non-Discriminatory Obligations
When standards are formulated, standards bodies like the ITU require all contributors to agree to license the technology royalty free, or on reasonable and non-discriminatory (RAND) terms and conditions, which prevents any single contributor from blocking commercialization by demanding too high a price, or unfair terms against a competitor or other party. If a contributor refuses to agree with either the royalty-free or RAND alternative, its technology won’t be included in the standard.
In the complaint, Nokia declared that it had agreed to license on RAND terms. If the Court finds Nokia’s patents are valid, and that Apple did infringe, one key issue becomes the RAND value of those patents. One case that the court will undoubtedly consult is Microsoft v. Motorola, decided in 2013, where Microsoft alleged that Motorola breached its RAND obligations.
In that case, Motorola claimed that Microsoft infringed three H.264-related patents and demanded RAND compensation, which they asserted equaled 2.25 percent of sales of Windows, PCs running Windows, Xbox, and other products, amounting to over $4 billion dollars. In its finding of fact, the court explained that when setting RAND royalty rates, courts consider the value of the patents to the standard, and the value of the standard to the product. This means every RAND determination is separate and distinct.
In the Motorola case, the court considered other licensing arrangements entered by Motorola, but also comparables like the $0.20/unit royalty charged by MPEG LA for a pool that represents thousands of worldwide patents from 38 different companies. The court set the H.264 FRAND rate at 0.555 cents per unit ($0.00555), which amounted to around $740,000 in H.264 royalties per year. Note that the total award was reported at $1.8 million per year, but about 60 percent of that was for 802.11 patents also considered in the case (see page 207). To add insult to injury, the court ruled that Motorola did breach its RAND obligations and awarded Microsoft $14.5 million in damages. Correctly or incorrectly, this finding created the impression that the costs associated with H.264 patents not in the MPEG LA pool, such as the Nokia patents, would be very low.
Motorola Not Binding
I spoke with David Long, a practicing patent attorney and editor of the Essential Patent blog about this impression, and how much the facts of the Motorola case control the potential awards in this case. He responded that since all RAND calculations are unique the Motorola case “would not be binding as a data point. While the court would certainly consider the MPEG LA royalty rate, it would likely place a higher priority on actual licensing deals between Nokia and other independent third parties.”
That’s because in previous cases, plaintiffs have argued that royalties offered by patent pools are often poor benchmarks for the actual essential value of the patent. In an article entitled Apportionment, FRAND Royalties, and comparable licenses after Ericsson v. D-LINK, author J. Gregory Sidak explained why. (Note that FRAND stands for fair, reasonable, and non-discriminatory, which is often used interchangeably with RAND).
“First, the royalties from a patent pool may provide an inadequate benchmark to calculate a FRAND royalty if the pool’s participants have a business model that significantly differs from the SEP holder’s business model. For example, companies that are active in the downstream market might prefer to recover their investment in research and development through the services offered on a standard-compliant product, such as an app for on-demand video streaming offered on a smartphone, rather than through licensing fees…Patent pools are also not useful benchmarks for determining a FRAND royalty because they often reward contributors on the basis of the number of contributed patents, rather than the patents’ relative value.”
For these reasons, Long expects Nokia to argue that actual commercial licensing arrangements with third parties are more relevant than MPEG LA rates. I sent an email to Nokia asking for financial details about these other agreements, which are not provided in the complaint, but did not hear back. If the case goes to trial expect these details to become absolutely central to Nokia’s award claims.
The Nokia Wrinkle
Beyond Nokia not being bound by the rates set in the Motorola case, Nokia is also claiming that because the H.264 standard defines a decoder and not an encoder RAND licensing limitations don’t apply to the encoding-related patents. Here’s a snippet from the Nokia complaint.
“The H.264 Recommendation specifies the implementation of decoders and specifically defines the ‘decoding process’ as ‘[t]he process specified in this Recommendation | International Standard that reads a bitstream and derives decoded pictures from it.’ It does not, however, specify the implementation of encoders. In fact, it specifically defines ‘encoder’ as ‘an embodiment of an encoding process,’ and then defines ‘encoding process’ as ‘a process, not specified in this Recommendation | International Standard, that produces a bitstream conforming to this Recommendation | International Standard.’ Id. at 6 (emphasis added)). As a result, since encoder implementations are not specified under the H.264 Standard, claims covering such encoders are not essential under the Common Patent Policy, and thus any such claims are not subject to a RAND commitment under that Policy.”
This distinction doesn’t appear to have been raised in the Motorola case which did involve encoding-related patents. I asked Long about Nokia’s distinction, and he explained that RAND applies when a court finds patents “essential to the standard.” In Nokia’s favor is the fact that the H.264 spec does define the decoding process, as noted above. Against the claim would be the simple fact that H.264 decoders would serve no essential purpose without encoded streams to decode. Either way, it’s a finding of fact for the court to make.
Long did point out that under general-purpose (eg. Non-RAND) patent law, all royalties must be reasonable, but that if the encoding-related patents were not limited by RAND, Nokia could discriminate against Apple and raise the price, perhaps because Nokia plans to re-enter the smartphone market where Apple is a prominent competitor. This all could prove irrelevant if Nokia simply tries to recover the same rates paid by other parties, but could be used to claim a higher royalty if the court agrees that encoder-related patents are not limited by RAND restrictions.
So where does that leave us? Basically, at the starting point of a long patent suit that could redefine how standard-based codecs are licensed, and dramatically increase the cost of H.264 decoders and especially encoders included in hardware and software products. Eleven years after MPEG LA released its first H.264-related price list, the expected cost of H.264 licensing may be completely revamped. Since some or all of Nokia’s patents might also apply to HEVC, licensing costs here may also be affected. Let’s hope it doesn’t take eleven years to find out.
We’ll learn a lot from Apple’s response, which should take about a month or so.
Note: The author thanks David Long and Florian Mueller from the FOSS Patents website for sharing their perspectives on these issues.