[00:00:01] Speaker 03: We are honored to have sitting with us this morning as a visiting judge, Judge Paul Etkin of the Southern District of New York. [00:00:10] Speaker 03: Welcome, Judge. [00:00:12] Speaker 03: We have five cases on the calendar this morning, one from the District Court, one from the Claims Court. [00:00:21] Speaker 03: One from the Trademark Board, one from the Merit Systems Protection Board, one from the Court of Appeals for Veterans Claims. [00:00:32] Speaker 03: Only the District Court and Claims Court cases are being argued. [00:00:36] Speaker 03: The rest is submitted in the briefs. [00:00:39] Speaker 03: Our first case is Constellation Designs versus LG Electronics, 2024, 1822. [00:00:46] Speaker 03: Mr. Schwentker. [00:00:50] Speaker 01: May it please the court, Andy Schwentker on behalf of LG. [00:00:54] Speaker 01: I'd like to begin with patent ineligibility, which is dispositive of the appeal. [00:01:00] Speaker 01: The district court committed two fundamental errors in granting summary judgment of eligibility for Constellation Design's four patents. [00:01:07] Speaker 01: The first error was finding that the 761 patents claims directed to black box optimization recited patent eligible subject matter. [00:01:17] Speaker 01: The district court then compounded its error by granting summary judgment of eligibility for the 700, 509, and 922 patents, even though there is no optimization limitation in those three patents asserted claims. [00:01:31] Speaker 01: Now, at ALICE step one, all of the asserted patents claims are directed to abstract ideas. [00:01:39] Speaker 01: For the 761 patent, the abstract idea is using a non-uniform constellation optimized for parallel decode capacity. [00:01:47] Speaker 01: As in hawk technology, the 761 claims recite result-oriented language optimization. [00:01:54] Speaker 00: I understand the argument you're making with respect to the 761 claims and the 700 claims. [00:02:01] Speaker 00: I was wondering if I could ask you, what do you think [00:02:05] Speaker 00: The abstract idea is that the 509 and 922 claims are directed to. [00:02:12] Speaker 00: And I read the red brief, this is, I guess, a two-part question, as arguing these claims separately. [00:02:18] Speaker 00: So if you could answer whether you think the claims are argued separately, and also what you think. [00:02:26] Speaker 00: I know that you're arguing that those are ineligible as well. [00:02:29] Speaker 00: I'd like to know what you think the abstract idea those are directed to, because I think I see those claims as being perhaps different. [00:02:35] Speaker 01: Yes, Judge Stoll. [00:02:36] Speaker 01: So the 700 and 509 patents claims are similar. [00:02:42] Speaker 01: The 509 patent claims have the additional requirement about overlapping constellation points. [00:02:50] Speaker 01: And this ties into the 922 patent claims as well. [00:02:54] Speaker 01: So the 922 patent claims. [00:02:57] Speaker 01: The abstract idea is a representation of a signal as a constellation. [00:03:02] Speaker 01: And you can see even in the red brief that a constellation is really just a two-dimensional mathematical representation of a signal. [00:03:14] Speaker 01: And that's, for the 922 patent, what we believe the abstract idea is. [00:03:19] Speaker 01: It's a mathematical representation of a constellation and cases like [00:03:25] Speaker 01: Hawk Technology and Stanford confirmed that manipulation of data, mathematical representations, mathematical formulas, those are not patent-eligible ideas. [00:03:38] Speaker 01: Those are abstract ideas. [00:03:39] Speaker 00: What if it's, though, I would say that's probably true if that's all the claim is directed to. [00:03:47] Speaker 00: But what if it's an application of the mathematical [00:03:52] Speaker 00: formula, or this is an application of the constellation. [00:03:56] Speaker 00: It's a specific constellation that is intended to resolve a particular technical problem. [00:04:06] Speaker 00: Why is that not right? [00:04:08] Speaker 01: So we know from cases like Myriad that simply reciting the abstract idea here, a mathematical representation, and saying apply it is insufficient under patent eligibility case law. [00:04:25] Speaker 03: Well, Myriad was a natural product case. [00:04:28] Speaker 03: Different kind of case. [00:04:30] Speaker 01: I don't understand, Judge Laurie. [00:04:32] Speaker 01: But numerous cases from this court have also held that simply reciting an abstract idea and saying apply it is insufficient to confer patent eligibility. [00:04:45] Speaker 01: And here, that's what the 922 patent claims are doing. [00:04:52] Speaker 01: The other components in the claim are simply conventional receiver components. [00:04:58] Speaker 01: And so the claim is reciting the conventional receiver and then saying to apply or use this constellation. [00:05:08] Speaker 01: There's nothing additional within the claim that would take it beyond [00:05:18] Speaker 01: simply saying to apply the abstract idea. [00:05:22] Speaker 00: What if this claim said it gave the result? [00:05:28] Speaker 00: For example, it said wherein the result is that you have a lower reduced signal to noise ratio. [00:05:41] Speaker 00: Would that change your answer? [00:05:44] Speaker 01: If it were just that, I don't think so. [00:05:47] Speaker 00: What do you think the claim would require in order to make it not abstract? [00:05:54] Speaker 01: Perhaps specific specificity around the reduction in sigma noise ratio. [00:06:04] Speaker 01: also perhaps some description of how to apply the constellation, other than just conventionally. [00:06:16] Speaker 00: Are constellations, would a person ordinary? [00:06:18] Speaker 00: I didn't think that's what the invention was directed to, how to apply the constellation. [00:06:22] Speaker 00: I thought what the invention was directed to was the development of the constellations. [00:06:28] Speaker 00: Am I wrong? [00:06:29] Speaker 01: Well, I think you're right. [00:06:34] Speaker 01: trying to answer the hypothetical of what might cause this to be a patent eligible. [00:06:40] Speaker 00: So in your view, probably there's very little that a person could do with this invention to make it eligible. [00:06:47] Speaker 00: Nothing in the specification. [00:06:49] Speaker 00: If it's just the creation of a new constellation for purposes of the way constellations have always been used, then it's not going to be eligible because it's in your view. [00:07:01] Speaker 00: Do I understand that correctly? [00:07:04] Speaker 01: I don't know. [00:07:05] Speaker 01: I mean, they have dozens of patents, and perhaps there are other ways that they can claim these in a way that takes them outside of the realm of abstract ideas. [00:07:16] Speaker 01: But at least with the claims that we're dealing with here, [00:07:20] Speaker 01: They are squarely within the realm of abstract ideas. [00:07:24] Speaker 00: I did want to get back to the... Can I ask you about the 761 and the 700 patents real quick? [00:07:30] Speaker 00: For those doing... I think I read your argument and understand it as saying those claims are like the claims the Supreme Court rejected in Morse. [00:07:38] Speaker 00: in that Moore's claims were directed to the use of electromagnetism, however developed for purposes of marking or printing intelligible character signs or letters. [00:07:53] Speaker 00: Do you think that that's the claims in the 761 and the 700 are analogous to that? [00:08:00] Speaker 01: I think that's fair. [00:08:01] Speaker 01: I think they're analogous to many other claims that have fallen under 101 as well. [00:08:09] Speaker 01: We recite the Hawk Technology case, the Stanford case, the electric power group, the two-way media. [00:08:23] Speaker 01: There are numerous cases that I think these claims are similar to. [00:08:27] Speaker 01: They recite result-oriented [00:08:30] Speaker 01: language without providing any details about how to obtain those results. [00:08:37] Speaker 01: I did want to get back to the 509 patent briefly, which you asked about. [00:08:43] Speaker 01: So the 509 patent, it recites the same abstract idea of selection of a paired code grade in a non-eiform constellation. [00:08:54] Speaker 01: It adds the [00:08:57] Speaker 01: overlapping constellation points. [00:08:59] Speaker 01: But in our view, that's, again, similar to the 922 pattern. [00:09:04] Speaker 01: That's a mathematical representation of a signal. [00:09:07] Speaker 01: It's just saying there are overlapping constellation points. [00:09:10] Speaker 01: And we know that layering one abstract idea on top of another abstract idea is not sufficient to confer eligibility. [00:09:20] Speaker 00: As a technical matter, what am I to understand from the locations if at least two of the constellation points are the same? [00:09:27] Speaker 00: Does that mean it's non-uniform? [00:09:30] Speaker 01: It is non-uniform. [00:09:31] Speaker 00: I'm just trying to understand technically. [00:09:34] Speaker 00: Is it saying anything more than that? [00:09:36] Speaker 01: It's non-uniform? [00:09:37] Speaker 01: It's saying that there are constellation points that are in the same location. [00:09:42] Speaker 03: OK. [00:09:43] Speaker 01: But that's it. [00:09:44] Speaker 02: Could I ask you a little more about the 761 patent? [00:09:48] Speaker 02: In the district court, the argument was essentially, this is optimization. [00:09:52] Speaker 02: That goes to an abstract idea. [00:09:56] Speaker 02: But the emphasis of your friend on the other side seems to be that this really is optimization specifically by reference to parallel decode capacity. [00:10:07] Speaker 02: And I guess, why isn't that? [00:10:09] Speaker 02: What is it about parallel decode capacity that [00:10:12] Speaker 02: isn't an innovative concept or isn't something specific enough to be not going to an abstract idea. [00:10:21] Speaker 01: So parallel to code capacity, [00:10:24] Speaker 01: is a concept that already existed in the prior art. [00:10:28] Speaker 01: We know that from the summer reference. [00:10:31] Speaker 01: And it's just a measure of the amount of information from the coder to the decoder. [00:10:42] Speaker 01: And the 7.61 patent claims are just saying to optimize for this measure without telling you how to do so. [00:10:53] Speaker 00: Are there different ways to optimize for parallel decode capacity? [00:11:00] Speaker 01: Yeah, and that's part of the problem here. [00:11:02] Speaker 01: The claim is attempting to preempt the entire endeavor of optimizing for parallel decode capacity simply by reciting that desired result in the claim without providing any mechanism for doing so. [00:11:22] Speaker 01: If there are no further questions on patent eligibility, I wanted to turn to non-infringement. [00:11:29] Speaker 01: There are two legal flaws with Constellation Design's standards-based infringement theory at trial. [00:11:39] Speaker 01: These flow directly from Fujitsu and its progeny, which [00:11:43] Speaker 01: permit a narrow exception if the reach of the claims includes any device that practices the standard. [00:11:50] Speaker 01: And there's two requirements built in there. [00:11:53] Speaker 01: One is to compare the claims to the standard. [00:11:57] Speaker 01: That's what Fujitsu and INVT [00:11:59] Speaker 01: in those cases require. [00:12:03] Speaker 01: And if the standard does not provide enough specificity, then those cases require the normal comparison between the claims and the accused product. [00:12:15] Speaker 01: The other aspect of this narrow exception in Fujitsu and INVT is that the product has to practice the standard. [00:12:25] Speaker 01: And here the... [00:12:29] Speaker 01: Neither one of these requirements is met here, or even arguably met here. [00:12:35] Speaker 00: You're saying that your product doesn't practice the standard? [00:12:39] Speaker 01: The product must practice the standard. [00:12:41] Speaker 01: So Fujitsu provides a narrow exception if the reach of the claims includes any device that practices the standard. [00:12:48] Speaker 01: And the standard that's at issue here [00:12:53] Speaker 01: The A322 standard is a transmitter standard, and so it's not a standard that the accused television receivers could even arguably practice. [00:13:05] Speaker 00: How do you respond to Judge Gilstrap's interpretation of Fujitsu, saying that [00:13:12] Speaker 00: Fujitsu doesn't mean that the only way that you could rely on the standard for purposes of showing infringement is when the standard covers every limitation of the claim. [00:13:24] Speaker 00: But instead, it is providing certain protections to make sure that someone's not using the standard in an improper or unreliable way. [00:13:36] Speaker 00: And so the protections that come with Fujitsu with respect to the whole claim should just apply when you're looking at a particular single limitation. [00:13:45] Speaker 00: What is wrong with that interpretation? [00:13:48] Speaker 00: It seems pretty commonsensical for patent law to me. [00:13:52] Speaker 01: A couple of responses Judge Stoll. [00:13:53] Speaker 01: So first of all, Fujitsu and INVT, I believe, foreclose that path because they require that if the standard does not provide the level of specificity required to show that a claim is standard essential, then the patentee has to follow the ordinary course of comparing the claim to the accused product. [00:14:22] Speaker 01: The other aspect here is by allowing less than a full claimed invention compared to less than a [00:14:38] Speaker 01: Compared to a standard that's not even arguably directed to the accused product, it's eviscerating the guardrails around the standard essentiality requirement in Fujitsu and INVT in those cases. [00:14:57] Speaker 00: Was the jury instructed on Fujitsu, or was there any objection made to the expert's reliance on this methodology for showing infringement? [00:15:12] Speaker 01: We objected. [00:15:16] Speaker 01: We raised this in a pretrial motion challenging the reliance on the standard here. [00:15:28] Speaker 00: I'm just thinking about it in terms of like JMAW. [00:15:32] Speaker 00: For JMAW, what we're doing is we're looking at whether there's substantial evidence to support the jury's verdict. [00:15:38] Speaker 00: So I guess your view is that there shouldn't be any reliance on the standards. [00:15:44] Speaker 00: So that means that one limitation, the jury, there's no evidence to support it. [00:15:49] Speaker 01: As a matter of law, what Constellation Designs and Dr. Jones did here was flawed. [00:16:00] Speaker 01: They relied on a standard without comparing the full scope of the claims to the standard. [00:16:09] Speaker 01: And they also relied on a standard that is indisputably not directed to or practiced by LG's televisions. [00:16:18] Speaker 01: It's a transmitter standard. [00:16:20] Speaker 01: And if there's a... So reliance on a standard [00:16:33] Speaker 01: as a proxy for infringement means that the mapping has to be 100% dead on. [00:16:40] Speaker 01: And there's no dispute that that wasn't the case here. [00:16:44] Speaker 01: And so that, as a matter of law, is a failure in their presentation. [00:16:49] Speaker 00: Can I ask you a quick question on damages? [00:16:52] Speaker 00: Yes. [00:16:52] Speaker 00: If this court were to determine that some of the patent claims are eligible and some of them are ineligible, what happens to damages? [00:17:04] Speaker 01: Obviously, as we've laid out in our brief, we think that the damages theory presented by Constellation Designs is fundamentally flawed. [00:17:11] Speaker 01: But even setting that aside, if this court were to reverse on patent eligibility, then the damages verdict should also be vacated for that reason as well. [00:17:26] Speaker 01: Because the way this case was tried [00:17:29] Speaker 01: All of the claims were just kind of lumped together. [00:17:32] Speaker 01: We don't know how the jury arrived at its decision, whether it placed more value on the 761 patent claims, for example, than other patents. [00:17:46] Speaker 01: Certainly, it would have been a very different trial if the 761 patent had not been at issue, because that's the one that claims optimization just kind of in the [00:17:58] Speaker 01: in the abstract, and the other claims don't recite the optimization. [00:18:03] Speaker 00: And also, do you recall by any chance how the damages evidence was presented? [00:18:13] Speaker 00: apportionment for the different patents, or consideration of if certain patents were found infringed in certain work, or they were all lumped together by your friend on the other side's expert witness? [00:18:24] Speaker 01: Well, they were all lumped together by Consolation Design's expert, and that's part of the fundamental problem with their damages theory. [00:18:34] Speaker 01: that they relied on a one price for all patent damages model, which this court's cases like Omega Patents preclude. [00:18:46] Speaker 01: So that was how it was presented to the jury. [00:18:51] Speaker 01: But we also don't know what happened in the jury deliberation room and how they [00:18:58] Speaker 01: We know that they reached the royalty rate espoused by Consolation Design's expert, but we don't know if the 761 patent, for example, supported most of that value or what have you. [00:19:14] Speaker 01: Long story short, if this [00:19:19] Speaker 01: If this court does anything on patent ineligibility, as we think it should, then the damned is verdict. [00:19:27] Speaker 01: For that reason, it should also be vacated. [00:19:30] Speaker 03: Counsel, you've exceeded your time for answering questions. [00:19:34] Speaker 03: We'll give you your rebuttal time back. [00:19:37] Speaker 03: Let's hear from Constellation. [00:19:39] Speaker 03: Mr. Wall. [00:19:44] Speaker 04: Judge Lurian may please the court. [00:19:47] Speaker 04: Judge Stoloff, I can go to your very first question, because I think it's the only real legal question in the case. [00:19:52] Speaker 04: At bottom, infringement and damages are all subject to a substantial evidence standard. [00:19:57] Speaker 04: Set aside for just a moment the 509 and the 922 patents, because I think the analysis on those falls out pretty easily once you have the 761 and the 700. [00:20:06] Speaker 04: And I want to try to convince the court, not just that they're patent eligible, that it's not a closed case. [00:20:12] Speaker 04: If you look at the language in 761, it's on the very first page of our brief, Romanette I, right after the cover. [00:20:18] Speaker 04: And you read it in light of the specification, which appears at page 135 of the appendix. [00:20:24] Speaker 04: Here's what's going on. [00:20:25] Speaker 04: When Drs. [00:20:26] Speaker 04: Jones and Barsoom are sitting there in the NASA lab, and he explains all of this starting at page 20172 of the appendix, everybody's focused on the output from the demodulator. [00:20:36] Speaker 04: You translate the waves into these symbols or points on the map. [00:20:39] Speaker 04: There's one of them at page six of our brief. [00:20:41] Speaker 04: And everybody's focused on trying to get those symbols closer to the grid points. [00:20:46] Speaker 04: And what Jones and Barsoom say is, well, wait a minute. [00:20:49] Speaker 04: You're looking at the wrong thing. [00:20:50] Speaker 04: You ought to look at the result not of the demodulator, but the demapper. [00:20:54] Speaker 04: Compare the bits going into the mapper on one end with the bits coming out of the demapper on the other. [00:21:00] Speaker 04: That's parallel decode capacity, which people were not focused on. [00:21:04] Speaker 04: And see if by moving the points around in the constellation, you can get better bit quality on the back end. [00:21:10] Speaker 04: Turns out you can. [00:21:11] Speaker 04: Counterintuitively, if you move the points closer together, instead of as everybody else has been trying to do, further apart, you can produce really substantial gains approaching what's called the Shannon limit. [00:21:23] Speaker 04: That's the graph that's on page 11 of our brief. [00:21:25] Speaker 04: So when they draft up the claim, it is [00:21:28] Speaker 04: a receiver using non-uniform constellations optimized for parallel decode capacity that at a given capacity, a signal noise ratio outperforms your uniform capacity. [00:21:40] Speaker 04: That's the demon part at the end of that claim language there. [00:21:43] Speaker 04: And so I look at that and I say, wait a minute, that's concrete all the way down the line. [00:21:47] Speaker 04: That's a different output. [00:21:48] Speaker 04: It's assessing it based on a different means. [00:21:52] Speaker 04: and it's producing different constellations than in the real world of digital communications give you a much more efficient over-the-air broadcast. [00:22:00] Speaker 04: Everything about that indicates that it ought to be eligible for patentability under the Patent Act. [00:22:07] Speaker 04: So whether the court looks at the claim language, reads it in light of the specification, as I say, I think the testimony from the inventor on this is very helpful, because Dr. Chris Jones testified at trial. [00:22:17] Speaker 00: One of the things I'm concerned about is this language optimized for capacity. [00:22:23] Speaker 00: And it does say how it's optimized. [00:22:25] Speaker 00: It's using something, the parallel decode capacity. [00:22:28] Speaker 00: But that's pretty broad language. [00:22:31] Speaker 00: It feels abstract because it's not saying how you do it, which I agree, your patent specification, I think it's figure five, has some good detail in that on how you do it. [00:22:45] Speaker 00: But it's not in your claim. [00:22:47] Speaker 00: And I understand the desire of having a claim that recites every single way of optimizing for capacity using parallel decode capacity. [00:22:55] Speaker 00: But I don't know that that's what your invention is. [00:22:58] Speaker 00: And so I just want to give you my feedback so that you can respond to it. [00:23:02] Speaker 04: Yeah, totally fair. [00:23:03] Speaker 04: So I think it goes to one of the questions you asked earlier, right? [00:23:05] Speaker 04: There are different ways to optimize. [00:23:07] Speaker 04: You can optimize for things other than parallel decode capacity. [00:23:11] Speaker 00: How many other things? [00:23:11] Speaker 04: Well, you can optimize for joint capacity. [00:23:14] Speaker 04: There was some testimony about that. [00:23:16] Speaker 04: People were using Gaussian distributions before Jones and Barsoon came along. [00:23:20] Speaker 04: But there's no testimony in this record that you can optimize using parallel decode capacity in multiple ways. [00:23:26] Speaker 04: Because parallel decode capacity is the metric. [00:23:29] Speaker 04: You start moving around the points on the constellation. [00:23:32] Speaker 04: And the comparison between bit quality on the transmitter end and on the receiver end [00:23:37] Speaker 04: That metric, that's parallel decode capacity. [00:23:39] Speaker 04: So you can do it other ways. [00:23:41] Speaker 04: Other people were. [00:23:42] Speaker 04: You were looking at the stuff coming out of the demodulator and trying to get the points closer to the grid points, the symbols closer to the grid points. [00:23:48] Speaker 04: But looking at the demapper results and the bit quality, that's parallel decode capacity. [00:23:55] Speaker 04: There's nothing that's in this. [00:23:57] Speaker 00: Has parallel decode capacity been defined in this case? [00:24:00] Speaker 00: Was it construed? [00:24:02] Speaker 04: I don't think anyone asked Judge Gilstrepp to construe that specific term in the claim. [00:24:09] Speaker 04: But there was a lot of testimony at trial about what parallel decode capacity is, why people weren't using it before Jones and Barsoon came along, and why what they brought to the table was a really inventive concept. [00:24:22] Speaker 00: This is summary judgment, so it doesn't really matter, right? [00:24:25] Speaker 04: That's right. [00:24:26] Speaker 04: My point is just they didn't walk in and say, [00:24:31] Speaker 04: There's a real lack of clarity around parallel decode capacity. [00:24:34] Speaker 04: We can't understand what that is from the claim language in light of the specification. [00:24:38] Speaker 04: Everyone, I think, agreed. [00:24:40] Speaker 04: And I don't see anything in their brief in this court to suggest that everyone didn't agree. [00:24:45] Speaker 04: that parallel decode capacity was a specific way, a metric, of comparing the bit quality on one end with the bit quality on the other. [00:24:53] Speaker 02: But is it more than a measurement or an aspirational thing, a goal? [00:24:59] Speaker 02: Is it just a measure of parallel decode capacity, or is it something specific? [00:25:04] Speaker 04: I would say it's a technique. [00:25:05] Speaker 04: So, for instance, if you look at a case like McRoe, this court says, [00:25:10] Speaker 04: All right, look, that's about sort of animation and trying to match up facial expressions with sounds. [00:25:17] Speaker 04: And everybody was doing one facial expression per sound. [00:25:21] Speaker 00: And then the patent there said, no, you should... Can you help us, instead of talking about our prior cases, could you help us? [00:25:27] Speaker 00: We're struggling here, I think, both of us, with what is the meaning of parallel decode capacity in the way that you're talking about it. [00:25:35] Speaker 00: You're trying to say that this is something concrete. [00:25:37] Speaker 00: It's not just a measurement. [00:25:40] Speaker 00: So maybe turning to the specification and helping us understand in that sense might be more helpful than talking about McRoe. [00:25:46] Speaker 04: Sure. [00:25:46] Speaker 04: So I'll pull up the specification at page 135, but I was going to pull around McRoe to say there it was a technique two because there it was assigning the facial expression to one sound and the patent was assigning it to a set of sounds. [00:26:01] Speaker 04: And so the claim there was the same as here, Judge Jack. [00:26:03] Speaker 04: And it was like, look, that's a technique that's not patentable. [00:26:06] Speaker 04: And what this court said was, but it's a technique that people hadn't been using. [00:26:10] Speaker 04: And it was a new and inventive way of using that technique to achieve a real concrete gain in the real world. [00:26:16] Speaker 04: They're better animation. [00:26:18] Speaker 04: Here, better over-the-air broadcasts. [00:26:20] Speaker 04: So once they come up with optimizing for parallel decode capacity, once you start using that technique, [00:26:27] Speaker 04: You produce these non-uniform constellations, which are the opposite of what everybody had been doing. [00:26:32] Speaker 04: They were doing uniform. [00:26:33] Speaker 04: And so they produce these non-uniform constellations and it turns out they work far better. [00:26:38] Speaker 04: You get a much more efficient broadcast. [00:26:40] Speaker 04: You get like that same kinds of gains you had seen collectively in the previous 40 years. [00:26:45] Speaker 04: So to my mind, yes, you are looking at a different output. [00:26:50] Speaker 04: You're applying a different technique. [00:26:52] Speaker 04: But what you're doing then is you're producing these different constellations that work a lot better in the real world. [00:26:58] Speaker 04: And all of that together is a concrete, specific advance. [00:27:02] Speaker 04: It's not just optimization. [00:27:04] Speaker 04: It's not just go do something with parallel deco capacity. [00:27:07] Speaker 04: It's take a receiver that uses non-uniform constellations that have been so optimized and that will, compared to uniform constellations at a given signal noise ratio, produce a better channel capacity, a better over-the-air broadcast. [00:27:21] Speaker 02: Mr. Schwinker says in his brief that you can't get around the summer reference and there's earlier reference in summer to parallel decode capacity and how do you respond to that? [00:27:33] Speaker 04: So it's not a reference in summer to parallel decode capacity, it's a reference to shaping gain and it doesn't have anything I think to do with 101 with all respect, Judge Ecken. [00:27:42] Speaker 04: It can be relevant to either 102 or 103 [00:27:45] Speaker 04: They brought obvious in the sandwich as a trial. [00:27:47] Speaker 04: They lost all of those in front of the jury. [00:27:49] Speaker 04: They haven't brought them again here. [00:27:51] Speaker 04: So that whether or not this was obvious in light of the prior arc or novel, those are other claims. [00:27:58] Speaker 04: They don't go to whether what these fellows did in 2006 was an abstract idea. [00:28:05] Speaker 04: or a concrete and specific improvement in the field of digital communication technology. [00:28:09] Speaker 00: Just to be clear, it's not talking about what these fellows did. [00:28:12] Speaker 00: It's talking about claim 17 of 761 PADS. [00:28:15] Speaker 04: Absolutely right, Judge Phil. [00:28:17] Speaker 04: So when I say what they did, what I mean is... I think they did something concrete. [00:28:20] Speaker 00: I'm worried about the claim. [00:28:22] Speaker 04: So I think they, sort of to your questions earlier, I think they wrote the claim in as concrete a way as they could have, which is to say, [00:28:33] Speaker 04: Obviously, the patents 509 and 922 are claiming specific kinds of constellations. [00:28:39] Speaker 04: So I think those are even more concrete. [00:28:41] Speaker 04: But what they were doing in 2006 was saying, look, we are designing our constellations [00:28:48] Speaker 04: by optimizing for parallel decode capacity, which is to say, we're moving the points on the constellation closer together rather than farther apart to increase the bit quality on the back end so that we have better bits being fed into the decoder, and then ultimately on the back end, a better broadcast. [00:29:05] Speaker 00: Some of the things you're talking about, like looking at the demapper and doing this, they're just not in the claim. [00:29:12] Speaker 00: That's one of the concerns. [00:29:14] Speaker 00: And instead, you've got [00:29:16] Speaker 00: optimize for capacity using parallel decode capacity without saying how you do that. [00:29:23] Speaker 00: It's any optimization whatsoever, any technique that would optimize parallel decode capacity. [00:29:30] Speaker 04: So sorry, that's the point I was trying to make earlier. [00:29:32] Speaker 04: It's not any technique for optimization, which is to say it's not doing optimization the way people were doing it pre-2006. [00:29:40] Speaker 00: I understand. [00:29:41] Speaker 00: It's optimizing using parallel decode capacity. [00:29:43] Speaker 04: That's right. [00:29:44] Speaker 00: But it's any way to optimize. [00:29:48] Speaker 00: So help me understand, because my understanding is that there's different ways that you could optimize for capacity using parallel decode capacity. [00:29:56] Speaker 04: Sorry. [00:29:57] Speaker 00: So that's not right. [00:29:58] Speaker 04: That right there, maybe that's the nub of the dispute. [00:30:01] Speaker 04: There is no testimony in this record that says there are different ways of optimizing for parallel decode capacity. [00:30:08] Speaker 04: Parallel decode capacity is a way of optimizing, which is to say, I'm moving the points around, and I'm just looking to see, are the bits getting closer on the back end to what got fed into the mapper here? [00:30:20] Speaker 04: So it's one metric. [00:30:21] Speaker 04: I'm trying to improve that one metric. [00:30:24] Speaker 04: And if I can get improvement along that one metric, [00:30:27] Speaker 04: then I've got a better broadcast on the back end. [00:30:29] Speaker 04: And so that's why I say all of this other stuff about the demapper and everything else, that belongs in the specification. [00:30:35] Speaker 04: The claim itself is a receiver using non-uniform constellations that have been improved with this technique. [00:30:42] Speaker 04: In other words, I've moved the points to get the bit quality on the back end the best that I can get it. [00:30:49] Speaker 04: And then that non-uniform constellation gives me a concrete gain in the here and now over a uniform constellation. [00:30:58] Speaker 04: So I'm not trying to fight the question, Judge. [00:31:01] Speaker 04: It's more, I don't know if I've been sitting there and I have designed this. [00:31:06] Speaker 04: What I would have done differently in terms of writing this claim other than to say, everybody else has been using uniform constellations that have been optimized a different way. [00:31:16] Speaker 04: I want a non-uniform constellation that I have made the best I can by using parallel decode capacity. [00:31:23] Speaker 04: I'm looking at a different thing in the process than what everybody else has looked at. [00:31:27] Speaker 04: And I would end up with a claim exactly like this. [00:31:30] Speaker 02: If I could just clarify, the 2007 article was the breakthrough article by the inventors. [00:31:35] Speaker 02: Does the patent date back to 2007 or 2008 somehow? [00:31:38] Speaker 02: I believe it does. [00:31:39] Speaker 02: OK. [00:31:43] Speaker 04: I don't know if the court has any questions on infringement or on damages. [00:31:48] Speaker 04: Judge Stoll, on damages, to your question, in many cases, where you knock out particular claims, it does affect the damages. [00:31:56] Speaker 04: On the specific record here, it wouldn't. [00:31:59] Speaker 04: Mr. Schwinkert's right, our expert testified, this is page 20280 of the appendix, that the licensing fee would have been the same with respect to any of the PEDs, but it wasn't just our expert. [00:32:12] Speaker 04: Their corporate representative, Richard Lewis, testified at 20333. [00:32:18] Speaker 04: He's an LG employee now, but he actually was at Zenith and helped negotiate the licenses that were at issue for comparability purposes. [00:32:26] Speaker 04: And he testified that the earlier $5 fee for the Zenith licenses was also not tied to any particular family. [00:32:33] Speaker 04: It was, as here, about a technology, the physical layer of this ATSC standard. [00:32:39] Speaker 04: So on the particular facts of this case, [00:32:41] Speaker 04: If you found that some of the claims were not patent eligible, it wouldn't affect the damages award in this particular case because of the testimony. [00:32:50] Speaker 00: How do you respond to Mr. Schwinker's point that we don't know how the jury valued [00:32:57] Speaker 00: the different patents. [00:32:59] Speaker 00: And if some of them were to survive and some of them weren't, how can we know how the jury valued it? [00:33:07] Speaker 00: I mean, I think that's a fair point. [00:33:09] Speaker 04: Well, except that the testimony from both of the experts was that the value of the licenses that were what the jury used for comparability, the zenith licenses, the testimony from both sides was that that value [00:33:20] Speaker 04: wasn't tied to any particular patent within the group. [00:33:23] Speaker 04: It was about a group of patents and a particular technology. [00:33:27] Speaker 04: Here, obviously, it's the technology that goes into what we call the physical layer of the ATSC standard, which is the constellations used by the receivers to pick up the broadcast. [00:33:37] Speaker 04: So I would say on this particular record, [00:33:39] Speaker 04: there would not be a basis to remand it back to the district court and have a new trial on damages. [00:33:44] Speaker 04: Because if they had tried to argue that the license value was in some way tied to particular patents in the group, then maybe. [00:33:53] Speaker 04: But here, they didn't make that argument below. [00:33:56] Speaker 04: And their own corporate witness said, no, that's not how I negotiated this license. [00:34:00] Speaker 02: But isn't that approach inconsistent with the Omega case, which rejected the one size fits all royalty approach? [00:34:07] Speaker 04: I don't think so, Judge Atkins, because in a case like Omega, the previous license was based on a whole group of patents, and then only one of those, one or two, had been asserted in the follow-on case. [00:34:19] Speaker 04: And the expert hadn't done anything to show that the value of the former license [00:34:23] Speaker 04: was actually tied to the now asserted patents as opposed to some of the other patents that had been in the group that were earlier licensed. [00:34:31] Speaker 04: Here it's a very different situation. [00:34:33] Speaker 04: It's a group of patents around a technology that gets licensed, and then the expert comes in and says, it's a corresponding group of patents now around the same kind of technology. [00:34:44] Speaker 04: And you have both sides saying that the earlier license value wasn't tied to any of the particular patents. [00:34:50] Speaker 04: So to me, there's a lot more correspondence in a case like this than in a case like Omega or Rex Medical, where there really was a concern that that earlier value was tied to some of the unasserted or uninvolved patents. [00:35:03] Speaker 04: And that's just not on the table here. [00:35:05] Speaker 00: It seems to me that if [00:35:06] Speaker 00: that the 707 patent claims, or sorry, 761 patent claims are extremely broad as compared to, say, the later claims in a 922 patent. [00:35:26] Speaker 00: Why would it be reasonable to think that the reasonable royalty might change based on what that claim is related to? [00:35:34] Speaker 04: I think for the reason that our expert gave below and that they did not meaningfully dispute the trial, which was this is really just about the technology, as I say, that sort of physical layer, the constellations that the receivers are using. [00:35:49] Speaker 04: And so the earlier licenses went to that group of patents that covered that technology, and so too here. [00:35:56] Speaker 04: And the testimony of trial was unequivocal that the standard, the constellations covered by all the patents are practiced by the standard. [00:36:06] Speaker 04: So if any of these patents is patent eligible, our experts said, they would have paid the same fee because they're [00:36:16] Speaker 04: practicing the standard and the standard uses, the constellations covered by the patent. [00:36:20] Speaker 04: I do hope that, though, I know that if you just looked at it, you might think it would be broad if there were lots of different ways of optimizing for parallel decode capacity, and you thought we were trying to claim and preempt the entire field, but that really [00:36:36] Speaker 04: Isn't what went on with this particular patent and it isn't as broad as it seems because it is a technique or a metric for optimizing But it's but it could result in a lot of different constellations, right? [00:36:48] Speaker 00: I mean, there's a lot of different constellations that would satisfy that claim Right. [00:36:54] Speaker 04: I mean yes judge stole you know in well yes and no keep in mind this is talking about at a given code rate and [00:37:04] Speaker 04: at a given signal to noise ratio. [00:37:07] Speaker 04: And there are other ways of optimizing and other ways of doing this. [00:37:10] Speaker 04: But it is right that if you are optimizing for parallel decode capacity, yes, it does claim those non-uniform constellations. [00:37:17] Speaker 04: I just don't think that that goes to the Alice question, which is not sort of, did you come up with an inventive concept that has a number of concrete applications? [00:37:30] Speaker 04: Did you come up with something concrete as compared to something abstract? [00:37:34] Speaker 04: And here, if they had just claimed optimization, if it were a case like that where this were a vague or not very good pattern, I'd understand it. [00:37:45] Speaker 04: But here, they were very specific about how you optimize and what it produces. [00:37:50] Speaker 04: And particularly right in light of the specification, I just don't see any abstractness there. [00:37:54] Speaker 00: So let me try to say that I think this case is a little bit different, that I see it a little bit more like a Morse case, where it's that the claim is so broad in claiming an end result with functional language as opposed to having a specific technical solution. [00:38:18] Speaker 00: You're familiar with the Morse case, right? [00:38:20] Speaker 04: Yes. [00:38:21] Speaker 04: So to me, it is far more specific than a case like that, again, because it's not like it's just claiming the result. [00:38:28] Speaker 04: We're just claiming the technique. [00:38:30] Speaker 04: It's sort of saying we're looking at a different output. [00:38:33] Speaker 04: We're judging it by a different means or metric. [00:38:35] Speaker 04: And we're getting an entirely different set of constellations that are sort of counterintuitive from what the rest of the field had been doing. [00:38:43] Speaker 04: And so to me, it's the combination of all of those that gives you real concreteness. [00:38:49] Speaker 04: We're telling somebody exactly what you get and how you get there. [00:38:54] Speaker 04: You use this particular technique. [00:38:55] Speaker 04: There's no testimony here that they're... The technique is to... Optimize using parallel depot capacity. [00:39:00] Speaker 04: That's right. [00:39:01] Speaker 04: And I know after the fact... How do you optimize? [00:39:05] Speaker 04: Is this figure five? [00:39:08] Speaker 04: So it is part figure five, but I think the best explanation actually is if you read Dr. Jones' testimony at pages 21, 72 to 175 of the appendix, he says, look, we had this kind of aha moment where we thought we could start looking at bit quality coming out of the demapper. [00:39:26] Speaker 04: So we wrote this software to optimize, which is to say we'd start moving the points on the constellations and seeing [00:39:32] Speaker 04: whether we got better bit quality that we could feed into the decoder. [00:39:36] Speaker 04: And right away, we saw massive gains, like on the order of 25%. [00:39:39] Speaker 04: We'd been laboring with the whole field for decades to try to achieve a one or a 2% gain. [00:39:46] Speaker 04: And all of a sudden, we start operating the software, moving the points, looking at the bit quality, and we get like 25% gains. [00:39:52] Speaker 04: The staggering, the sort of stuff like I couldn't have achieved in a career, basically, he says. [00:39:58] Speaker 04: In terms of how you actually implement it, they designed some software in order to move the points, look at what it did to the bit quality, and try to come up with an algorithm in order to optimize. [00:40:08] Speaker 04: And what they discovered was not just moving the points closer together, [00:40:13] Speaker 04: better, but in some cases, putting the points right on top of each other. [00:40:16] Speaker 04: That's the 509 pattern, a constellation with two identical points. [00:40:21] Speaker 04: And nobody would have thought to do that before Jones and Barcine came along. [00:40:24] Speaker 04: And that's as concrete as it gets. [00:40:26] Speaker 04: Yes, 509 and 922 are sort of the more concrete examples of that. [00:40:32] Speaker 00: It's like an application. [00:40:33] Speaker 04: An application almost. [00:40:35] Speaker 04: But I still think that the manner in which you get there [00:40:40] Speaker 04: is equally concrete. [00:40:42] Speaker 04: It produces different sets of constellations, some with two identical points, the hundreds that are claimed in 922. [00:40:49] Speaker 04: But it is still a very concrete way of doing this that is a real advance in the field of digital communication. [00:40:57] Speaker 02: In the industry, they were already working with non-uniform constellations. [00:41:03] Speaker 02: Is it possible? [00:41:05] Speaker 02: plausible that they actually were, and they realized that was better in some ways, people in the industry. [00:41:11] Speaker 02: Is it possible they were actually optimizing for PD capacity in doing that, maybe without naming it? [00:41:18] Speaker 04: So there's no testimony about that in this record, Judge Eken. [00:41:23] Speaker 04: The only thing we have in the record is at page 6097 there is an article that was in the papers and was referred to a trial. [00:41:31] Speaker 04: One of the co-authors is an algae employee which credits this using parallel optimization using parallel decode capacity as an advance in the field which had been focused on other things like Gaussian distribution, bell curves, that sort of thing. [00:41:45] Speaker 04: I will say not only is there nothing in the record about it but having now learned more about the science here than I [00:41:52] Speaker 04: ever would have believed possible. [00:41:55] Speaker 04: I don't think there is anything outside the record that would suggest that what Dr. Jones and Barsoom did was not a real advance in the field. [00:42:02] Speaker 04: That graph at page 11 on the Shannon limit. [00:42:05] Speaker 04: I don't take the other side to be contesting that, which is to say, once they come along in 2007, they publish, they do the conference, all the things Judge Stoll talked about. [00:42:13] Speaker 04: This is a real advance. [00:42:15] Speaker 04: I mean, almost overnight, these broadcasts get more efficient, and you see huge gains toward the Shannon limit. [00:42:21] Speaker 04: So did people know about non-uniform constellations? [00:42:24] Speaker 04: Yes, but. [00:42:25] Speaker 03: You have exceeded your time. [00:42:27] Speaker 03: I answered questions. [00:42:29] Speaker 03: Thank you very much. [00:42:30] Speaker 03: Mr. Wall, you have five minutes for the public to leave it. [00:42:40] Speaker 01: A few quick points. [00:42:48] Speaker 01: Regarding damages, so the testimony from LG's corporate witness, Mr. Lewis, that Mr. Ball referred to, says nothing about the value of Constellation Design's patents. [00:43:06] Speaker 01: The notion that his testimony is somehow directed to how the constellation designs patents should be valued vis-a-vis each other, that's simply not the case. [00:43:21] Speaker 01: He was discussing the Zenith patent portfolio. [00:43:25] Speaker 01: The Zenith patents, as we discussed in our briefing, [00:43:32] Speaker 01: The Zenith patent portfolio is an entirely different set of patents than the Constellation Design patents, directed to an entirely different technology, technological standard. [00:43:45] Speaker 01: The Zenith patents were directed to VSB modulation in the original digital TV broadcast standard. [00:43:54] Speaker 01: The Constellation Design patents are allegedly [00:43:58] Speaker 01: directed to ATSC 3.0, which ATSC 3.0 involves a different type of modulation technique, OFDM. [00:44:08] Speaker 01: And there's just a number of differences as we've laid out in the briefing. [00:44:18] Speaker 01: As for what Dr. Sullivan testified regarding on appendix page 20280, [00:44:28] Speaker 01: key. [00:44:29] Speaker 01: He testified that it should be one price for all, that regardless of the number of patents are infringed, that the rate should be $6.75 per unit. [00:44:43] Speaker 01: And that's exactly the type of legally flawed testimony that this court rejected in Omega patents. [00:44:50] Speaker 01: In fact, the situation is even worse here than in Omega patents, because at least in Omega patents, [00:44:58] Speaker 01: Two of the licenses included the asserted patent in that case. [00:45:04] Speaker 01: And even in that case, the testimony was insufficient to apply a one price, $5 per unit, regardless of the number of patents that were infringed. [00:45:19] Speaker 01: Here, we don't even have the assertive patents in the agreements that Constellation Designs and Dr. Sullivan were relying on to try to establish a royalty. [00:45:32] Speaker 00: Before you run out of time, can I ask you a really quick eligibility question? [00:45:36] Speaker 00: I would like to hear your response to the argument that I understand is being made that optimizing parallel decode capacity is specifically looking at comparing the information sent to the mapper on one end of the channel and the information coming out of the new mapper on the other end. [00:45:55] Speaker 00: And so it's much more specific than, it's not an abstract idea, it's a specific concrete idea. [00:46:02] Speaker 01: So we know that, at best I would say that that is a narrowing of the abstract idea. [00:46:11] Speaker 01: And we know from cases like BSG technology that simply a narrower abstract idea is still an abstract idea. [00:46:23] Speaker 01: get you out of the issues under ALICE step one and ALICE step two. [00:46:31] Speaker 01: And the parallel to code capacity, it is just a measure of [00:46:38] Speaker 01: of the information between the coder and the decoder, and that this is supported by the patent specification. [00:46:51] Speaker 01: There was also a suggestion that summer does not address parallel decoded capacity, and I would point the court to appendix page 1118 to 1119. [00:47:02] Speaker 01: that's within Dr. Ackle's report. [00:47:05] Speaker 01: Dr. Ackle, in his expert report, went through all of these patent claims and limitations and provided his analysis for why the patent claims are directed to abstract ideas and also why the claimed elements are conventional. [00:47:26] Speaker 01: And he tied the summer reference to parallel decode capacity. [00:47:32] Speaker 01: And I did not understand there to be any dispute that the sum of references was about parallel to code capacity, but certainly, Dr. Atkinsford had very well suggested. [00:47:45] Speaker 00: Yeah, but we just have to be careful not to confuse eligibility and whether something's directed to an abstract idea. [00:47:52] Speaker 00: That might matter for Alice step two, but not for Alice step one, right? [00:47:56] Speaker 00: Whether something's conventional learning the prior art doesn't matter for Alice step one. [00:48:02] Speaker 01: two responses to that, Judge Stoll. [00:48:04] Speaker 01: So one, the point I was trying to make is that the summer reference is tied to parallel decode capacity. [00:48:10] Speaker 01: It is talking about parallel decode capacity, and you can see that in, at the very least, Dr. Ackles' report. [00:48:17] Speaker 01: But the other point I would make in response is that the [00:48:24] Speaker 01: You are certainly correct that the ALICE Step 2 looks at what is conventional. [00:48:32] Speaker 01: However, for ALICE Step 1, we do at least look, and this is supported by, for example, the semi-OKs [00:48:43] Speaker 00: Well, what I hear your friend on the other side saying is that this is a technological solution to a technical problem. [00:48:51] Speaker 00: And the technological solution is looking at the information coming out of the mapper or going in and then coming out and just trying to optimize that they be as similar as possible. [00:49:05] Speaker 00: Is that right? [00:49:08] Speaker 00: And so he was saying that's concrete. [00:49:10] Speaker 00: It's a technological solution. [00:49:11] Speaker 00: Nobody ever did it before. [00:49:14] Speaker 01: That is just the idea, though, of using a measure to optimize a constellation for that measure. [00:49:44] Speaker 01: The other point that I wanted to make was that we know from Dr. Jones' deposition testimony, he conceded that any method of optimization [00:50:02] Speaker 01: would meet the claim language. [00:50:05] Speaker 01: So any method of optimizing for parallel code capacity, that would meet the claim language. [00:50:12] Speaker 02: And do you believe there are more than one way to do that, to optimize for PD capacity? [00:50:17] Speaker 01: Yeah, and the pad specification, I think, supports this. [00:50:21] Speaker 01: It talks about how the user can specify different threshold values, for example, in the optimization process. [00:50:32] Speaker 01: So as a result, if it's just up to the user to set threshold values during the optimization process, that means there are more than one way to optimize a constellation for parallel to code capacity. [00:50:49] Speaker 01: Because we could have two different people setting different threshold values, and that by definition results in two different optimization processes. [00:50:59] Speaker 03: Thank you, counsel. [00:51:00] Speaker 03: I think we have both arguments. [00:51:02] Speaker 03: We thank both of you. [00:51:04] Speaker 03: The case is submitted.