[00:00:00] Speaker 00: Our next case is Netflix and Hulu versus Divx, or D-I-V-X, 2021, 1931. [00:00:11] Speaker 00: We're ready when you are, Mr. Batts. [00:00:18] Speaker 02: Thank you, Your Honor. [00:00:19] Speaker 02: May it please the Court, I have four separate issues that the Board erred on below. [00:00:24] Speaker 02: And I'd first like to address the issue of the claim construction of the term level of detail. [00:00:29] Speaker 02: The claim language itself clearly requires a determination of a region of pixels. [00:00:34] Speaker 02: At the institution stage, the board recognized this and adopted a construction of level of detail being a level of variation in visual elements across adjacent pixels. [00:00:45] Speaker 02: And this is a claim construction issue that should be reviewed de novo, and I don't believe Divix disputes that. [00:00:52] Speaker 02: But the board clearly changed its claim construction from the institution decision in the final written decision by injecting the requirement that there be a determination between adjacent pixels rather than across adjacent pixels. [00:01:06] Speaker 02: I think that's clear from the final written decision that we cited to at Appendix 2655. [00:01:12] Speaker 02: And furthermore, our brief also included the questions from the judge, Judge Ahmed, at the hearing about whether, if he adopted DIVIX's claim construction of between adjacent pixels. [00:01:24] Speaker 02: But claim one. [00:01:25] Speaker 03: Doesn't the question really then come down to the question of whether the term across in this context really means between or not? [00:01:35] Speaker 02: In some ways, yes. [00:01:36] Speaker 02: I do believe that's correct, Your Honor. [00:01:37] Speaker 03: The whole case really, in effect, the first part of the case at least, really turns on that question, right? [00:01:42] Speaker 02: I believe so, yes, your honor. [00:01:43] Speaker 02: So I think if we look at the claim language, the claim language is clearly not limiting to a particular calculation or how a calculation is performed for a level of detail. [00:01:54] Speaker 02: And the claim language is, so, and DIVIX has conceded that point in the red brief at pages 27 and 28. [00:02:01] Speaker 02: And furthermore, figure two, so there's various embodiments in the 651 patent. [00:02:05] Speaker 02: The figure two embodiment of the 651 patent does not limit a particular way of performing this calculation for a level of detail. [00:02:14] Speaker 02: And so the parties did both cite to the specification at column one, line 54, starting at line 54. [00:02:21] Speaker 02: The definition of detail. [00:02:24] Speaker 02: Sorry? [00:02:24] Speaker 03: The definition of detail is where you're going, right? [00:02:26] Speaker 02: The reference of detail, yes. [00:02:28] Speaker 02: And if we go to the... Well, definition may be an overstatement. [00:02:30] Speaker 02: the reference to detail. [00:02:31] Speaker 02: The explanation of detail, where it says the term detail, and it goes on to say, is typically used to describe visual elements of a video frame that vary significantly across adjacent pixels. [00:02:43] Speaker 02: So again, that language is also across. [00:02:45] Speaker 02: And so if we look at the dependent claims here. [00:02:49] Speaker 03: But does adjacent in that sense mean immediately adjacent? [00:02:54] Speaker 03: That is to say, as I am sitting adjacent to Judge Lurie, [00:03:00] Speaker 03: Arguably, at least, I'm not sitting adjacent to Judge Hughes. [00:03:05] Speaker 02: The problem with that, I guess the problem with Divick's argument on this, on trying to separate, on creating, adjacent being the same as between, is that adjacent, there's no dispute here that what we pointed to in our ground for invalidating here were figures five A, B, and C of Bavelin. [00:03:25] Speaker 02: All three of those figures include regions of adjacent pixels. [00:03:30] Speaker 03: Adjacent in which sense? [00:03:32] Speaker 03: The immediately adjacent or generally adjacent? [00:03:36] Speaker 03: You made the argument in your brief that that demonstrates that there can be instances in which they are immediately adjacent. [00:03:43] Speaker 03: But as I understand the thrust of your argument, it really is that you don't have to have a system that is testing for variation based on [00:03:55] Speaker 03: assessing differences between immediately adjacent pixels, which, of course, the Vivalinian does not do. [00:04:04] Speaker 02: I agree with you until the last point, Your Honor. [00:04:05] Speaker 02: Well, it doesn't do in all cases. [00:04:08] Speaker 02: It does not do it in all cases. [00:04:09] Speaker 02: Agreed. [00:04:10] Speaker 02: So, Weberlinen certainly teaches the min-max calculation that we pointed to in Weberlinen, and the Institution Decision credited on page 1712, it noted that Weberlinen does look at each and every pixel of all the pixels in the adjacent region. [00:04:25] Speaker 02: And there's been no change from that math or that calculation throughout the proceeding blow up to the final written decision. [00:04:31] Speaker 03: So is your position that adjacent, as is used both in the claims and in the discussion of detail, really means adjacent as in contiguous? [00:04:44] Speaker 03: That is part of a region of contiguous as opposed to segregated pixels? [00:04:53] Speaker 02: Well, I would say no, Your Honor. [00:04:54] Speaker 02: And I would say that if you look at, for example, what would not be an adjacent region of pixels, I would say we'd point to figure 5a of Babel Inon. [00:05:03] Speaker 02: And if you only took every other column of pixels. [00:05:06] Speaker 02: Right, but those wouldn't be contiguous either. [00:05:08] Speaker 02: They wouldn't be contiguous either. [00:05:09] Speaker 02: That's correct. [00:05:09] Speaker 03: I'm actually, this question, I thought you were going to give a different answer to this question, because it seems to me that your argument is at its strongest. [00:05:20] Speaker 03: If you can say that adjacent, [00:05:22] Speaker 03: refers to the region in which all of the pixels are contiguous. [00:05:29] Speaker 02: I think that's right, Your Honor. [00:05:30] Speaker 02: That was what I was trying to get at. [00:05:32] Speaker 02: Sorry if there's confusion on that point. [00:05:34] Speaker 02: But I do think there is an absolute difference between when you're talking about a calculation across pixels versus between pixels. [00:05:42] Speaker 02: Because a between pixel calculation that the board then used at the final written decision is requiring you looking at the space between two adjacent pixels having to do a calculation. [00:05:52] Speaker 03: But your opposing counsel is no doubt going to tell us that across adjacent pixels means from one pixel to the next. [00:06:00] Speaker 03: as long as they're next to one another. [00:06:02] Speaker 03: And that's a cross. [00:06:03] Speaker 03: And a cross doesn't refer to from pixel A to pixel J, even if there are pixels adjacent to each of the pixels in between. [00:06:13] Speaker 02: So I think that's what they will likely argue. [00:06:17] Speaker 03: But why is that wrong? [00:06:19] Speaker 02: So I think that what's clear is that there is no limitation in the specification or in the claim limiting to a particular type of calculation. [00:06:27] Speaker 02: And there's no limitation as to, it discusses a region of pixels, and it says the region of pixels has to include pixels adjacent to a block boundary. [00:06:36] Speaker 02: And if we look at the dependent claims, which we pointed out in our blue brief, dependent claims two and three are requiring a between adjacent pixel calculation. [00:06:46] Speaker 02: It's saying for, dependent claim two says for some pixels that are adjacent, and then it goes to the dependent claim three is requiring all [00:06:54] Speaker 02: pixels be utilized in that sum of absolute value determination. [00:06:58] Speaker 02: And then if we look at the file history, I do think that the Kim reference rejection is notable because during prosecution, when faced with their original claim before they amended it that included a level of detailed determination, [00:07:10] Speaker 02: The Kim reference was used as a rejection, and the Kim reference, I don't think there's any dispute, did not look at adjacent pixels. [00:07:17] Speaker 02: It was looking at separated pixels, and they didn't argue that that didn't disclose a level of detailed determination. [00:07:23] Speaker 03: But does Kim tell us what adjacent means in this context? [00:07:28] Speaker 03: That is to say, does it distinguish between [00:07:30] Speaker 03: Between and adjacent in the sense that you're using the term I think it does because it's saying you lay that out for me because I found the Discussion of Kim to be somewhat opaque. [00:07:39] Speaker 03: So help me out. [00:07:41] Speaker 02: Sure. [00:07:41] Speaker 02: So so Kim there I don't believe there's any dispute that when we look at the Kim Figure figure to a of Kim that the that the pixels that are be the eight pixels that are being analyzed There are not adjacent to one another. [00:07:55] Speaker 02: They're separated [00:07:56] Speaker 02: And yet, the patent owner did not argue that Kim was not determining a level of detail for the region. [00:08:03] Speaker 02: Instead, the patent owner amended its claims to add in further [00:08:07] Speaker 02: basically further amendments and further limitations to get past Kim. [00:08:11] Speaker 02: So if level of detail was limited to a determination between adjacent pixels, they could have simply said that to overcome Kim. [00:08:21] Speaker 02: So I don't think that's a binding argument, but it certainly is a persuasive argument in terms of consistent with the claim. [00:08:29] Speaker 03: So there's no disclaimer, but there is a choice of ways of getting around Kim that you say indicate or suggest [00:08:37] Speaker 03: that they had a more limited order. [00:08:44] Speaker 02: The way I would phrase it, Your Honor, is when does a patent owner choose to amend its claim and further narrow its claim when it could simply have overcome the claim rejection by arguing it didn't disclose the original claim's content? [00:08:57] Speaker 02: And so I do want to also hit the 5C because your questions touched upon the 5C issue as well. [00:09:04] Speaker 02: Because we contend that even under the board's final claim construction between adjacent pixels, they want in 5C satisfies that claim construction as well. [00:09:15] Speaker 02: And that's because if we look at the figure 5C, which is on appendix page 28 for example, you have [00:09:21] Speaker 02: to a region of pixels 57, and if we look on the left-hand side of that figure, there's a total of six pixels in a two-by-three layout. [00:09:32] Speaker 02: And what we know from the min-max calculations of vaveline is that the pixels that are going to be analyzed at the final analysis, not looking at all the pixels, but the two pixels that are used for the min-max calculation are going to be adjacent to each other in the majority of cases. [00:09:47] Speaker 02: So in terms of the pixel pairs, if you're looking at the middle pixel, either of the middle pixels on the left-hand side, that middle pixel is going to be looking at all the pixels that it's analyzing are adjacent to it. [00:09:59] Speaker 02: So in 11 out of the 15 possible combinations that would be used for the min number and the max number for the calculation, [00:10:08] Speaker 02: those pixels are going to be adjacent, and the calculation is going to be between adjacent pixels. [00:10:13] Speaker 02: So even if you don't agree on our issue of claim construction, we certainly showed, and we raised this issue in our briefing, in the reply brief after we saw patent owner's argument about between pixels and its patent owner response, we pointed to figure 5C and said, even under your construction, this renders the claims obvious. [00:10:33] Speaker 02: And I don't think DIVIX has contested that math. [00:10:36] Speaker 02: It's math. [00:10:37] Speaker 02: You look and you say, how many permutations can there be? [00:10:39] Speaker 02: It's 11 out of 15. [00:10:41] Speaker 02: And this court has repeatedly held that a reference need not always teach a claim limitation. [00:10:47] Speaker 02: So we cited the Hewlett-Packard-Mustic decision for that issue. [00:10:51] Speaker 02: And also, I think the CRFD decision is quite instructive, because in CRFD, this court said that if there was a 50% likelihood, it was equally likely. [00:11:01] Speaker 02: that something was obvious. [00:11:03] Speaker 02: That was enough. [00:11:04] Speaker 02: Here we're talking about 11 out of 15 permutations. [00:11:06] Speaker 02: That's more than two-thirds. [00:11:08] Speaker 04: Can I move you on, because you're into your rebuttal, to the motivation to combine issue? [00:11:14] Speaker 02: Yes. [00:11:14] Speaker 02: And I do want to hit the Codono threshold real fast. [00:11:16] Speaker 04: Yeah. [00:11:17] Speaker 04: Yes. [00:11:17] Speaker 04: Well, that's what I want to ask you about. [00:11:20] Speaker 04: So this stuff is very difficult, but the board found that [00:11:26] Speaker 04: You know, there's no dispute that Vivalainen doesn't teach the SAD method. [00:11:31] Speaker 04: You have to get to Kadono for that. [00:11:32] Speaker 04: And they found that there was no reason to substitute Kadono into Vivalainen because it was focused on variance. [00:11:38] Speaker 04: That's absolutely correct. [00:11:39] Speaker 04: So what's wrong with that? [00:11:40] Speaker 02: So the issue that we have... Are you talking about principal operation or the mega threshold issue? [00:11:46] Speaker 04: Oh, I think you just went way too deep. [00:11:50] Speaker 04: So here's the thing I'm asking and I do see some issues here. [00:11:54] Speaker 04: It seems to me that there's two steps, right? [00:11:56] Speaker 04: There's determining the level of detail and then there's filtering. [00:12:01] Speaker 04: And Vivalenin clearly focuses on using variance for filtering because whatever, it makes it work better in its view. [00:12:08] Speaker 04: I'm very high level here, so try to stick with me. [00:12:11] Speaker 04: I take it your argument is that even though Vivalenin focuses on variance for filtering and it uses variance for detecting, you could still use Codono for detecting and then move to variance for filtering. [00:12:25] Speaker 02: Yes, but I think you're actually addressing two of our separate issues, the principle of operation error and the error about not considering a separate ground on condono. [00:12:34] Speaker 02: So I want to address those. [00:12:35] Speaker 02: Talk about the principle of operation. [00:12:37] Speaker 02: Sure. [00:12:38] Speaker 02: So for principle of operation, I think the record is clear. [00:12:40] Speaker 02: They argued a principle of operation for the selecting step. [00:12:43] Speaker 02: They never raised a principle of operation argument for what we're discussing here, the level of detail. [00:12:48] Speaker 02: There was no briefing on that. [00:12:50] Speaker 02: There was no argument. [00:12:51] Speaker 02: The word principle of operations never came up in the oral hearing. [00:12:55] Speaker 02: The first time we saw a determination on, so we had no time opportunity to brief what would be the principle for that claim limitation. [00:13:03] Speaker 02: It first popped up in the final written decision. [00:13:05] Speaker 02: So that's a clear error where the board raised a new argument that had never been briefed or argued. [00:13:10] Speaker 03: I wonder whether that's really a new argument or simply a characterization of an argument that was already made, albeit not with the use of the term principle of operation. [00:13:20] Speaker 02: Well, I don't think it does. [00:13:21] Speaker 02: Because I think if you look at the citations that Divics relied upon, they have a scattering of citations in their brief that really don't address. [00:13:30] Speaker 02: A lot of them aren't even for motivation to combine. [00:13:33] Speaker 02: So motivation to combine is still separate from principle of operation. [00:13:37] Speaker 02: And if we look at the Henry Moutet case from this court, you have to make a determination of what is that principle of operation before you can decide whether that principle of operation would be changed. [00:13:46] Speaker 02: And that was never addressed below. [00:13:48] Speaker 02: And I would like to get back to the Codono threshold issue, because I think it is an important one for a reversal here. [00:13:53] Speaker 02: Because we had two separate reasons on why you combine Babel-Linen with Codono. [00:13:59] Speaker 02: One was the substituting of the SAT calculation for the variance calculations in Babel-Linen. [00:14:05] Speaker 02: But we had a separate one, and we included it on, I will just quickly give it to you, on pages [00:14:09] Speaker 02: appendix pages 130 through 134, we included a separate entire basis of which you would add. [00:14:16] Speaker 02: So you're not substituting anything out. [00:14:18] Speaker 02: You're adding a sad calculation, omega threshold calculation, to Babeline's existing structure and then modifying the filter for Babeline and to be able to utilize that. [00:14:29] Speaker 02: And that was addressed, credited by the board in terms of we made that argument in the final written decision, but they never addressed that argument or rejected that argument. [00:14:36] Speaker 02: And that alone is a basis for remand. [00:14:38] Speaker 00: Counsel, as you've noted, your time is almost expired. [00:14:41] Speaker 00: We'll give you two minutes for a bottle. [00:14:44] Speaker 01: Thank you, Your Honor. [00:14:45] Speaker 01: To Weather Wax. [00:14:47] Speaker 01: Good morning, Your Honors. [00:14:49] Speaker 01: I'd like to immediately address some of the issues raised by my friend on the other side. [00:14:56] Speaker 01: First of all, with respect to across versus between, I would like to direct Your Honor's attention [00:15:04] Speaker 01: to where they admitted that they were the same between two pixels. [00:15:09] Speaker 01: If you would please look at Appendix 3403, Judge Auchman, who wrote the decision below, straight up asked Netflix Hulu Council, is there a difference between those two? [00:15:25] Speaker 01: And Petitioner's Council said that between two adjacent pixels, and we mean immediately adjacent, that was never in doubt throughout the case. [00:15:32] Speaker 01: It was always immediately adjacent. [00:15:34] Speaker 01: But he said, between two adjacent pixels. [00:15:38] Speaker 03: You're on 3403, did you say? [00:15:40] Speaker 01: Yes, that's correct. [00:15:43] Speaker 01: There's no difference between across and between. [00:15:47] Speaker 01: And they never articulated how that somehow changes in some other context. [00:15:52] Speaker 01: They never offered anything else for that. [00:15:54] Speaker 01: And they've never mentioned that here. [00:15:58] Speaker 03: Are you talking about the response by Mr. Liang at line 19? [00:16:03] Speaker 01: Yes, he says. [00:16:05] Speaker 01: I suppose. [00:16:05] Speaker 03: But the question was, if you're looking at just two pixels, is there a difference between and across? [00:16:13] Speaker 03: He says, I suppose if you're just looking at just two pixels, then it would seem to do the same. [00:16:17] Speaker 01: That's right. [00:16:18] Speaker 01: But he never go on. [00:16:19] Speaker 01: to explain how it could be different otherwise. [00:16:22] Speaker 03: Oh, well, that's not... I wouldn't regard that as a concession, that it couldn't be otherwise if you had multiple pixels in the picture. [00:16:30] Speaker 03: Your Honor, if you go back to the... Do you have anything else by way of acknowledgement on their part that the between and across mean the same? [00:16:37] Speaker 03: Because I'm not persuaded by what you've just pointed to. [00:16:40] Speaker 01: Judge Bryson, they never said what the difference [00:16:45] Speaker 01: with respect to some other context besides between two adjacent pixels, which, by the way, was always the big focus here, whether this had to be reflecting the difference between how adjacent pixels, immediately adjacent pixels, vary compared to each other. [00:17:04] Speaker 01: If you go back and look at the statements in the patent that everyone always was looking at with respect [00:17:13] Speaker 01: to what detail meant. [00:17:16] Speaker 01: It doesn't say between, it says across adjacent pixels. [00:17:21] Speaker 01: And there's no reason why there should be any difference between across and between without it somehow being articulated. [00:17:29] Speaker 01: And no one's ever really explained what the difference is supposed to be. [00:17:34] Speaker 01: After all, detail is supposed to be reflecting how they change across adjacent pixels. [00:17:41] Speaker 01: What could it mean other than [00:17:43] Speaker 01: You have two adjacent pixels. [00:17:44] Speaker 01: How do they change? [00:17:46] Speaker 01: What is supposed to be the difference between two adjacent pixels or the difference across adjacent pixels? [00:17:54] Speaker 01: Because the difference is the difference in prominence or luminance, in value, visual elements. [00:18:01] Speaker 03: What would you say as to the fact that the patent sometimes refers to immediate [00:18:13] Speaker 03: be adjacent and sometimes merely too adjacent. [00:18:16] Speaker 03: Do you think there's a difference between those two? [00:18:19] Speaker 01: No one has ever argued there was a difference. [00:18:21] Speaker 03: Well, I'm asking. [00:18:22] Speaker 01: I don't think there is a difference. [00:18:24] Speaker 01: Is this careless drafting in the patent? [00:18:27] Speaker 01: I would think so. [00:18:27] Speaker 01: And furthermore, I would also say that when it was being discussed in the case, sometimes people would emphasize what they meant. [00:18:38] Speaker 01: But I don't think there was ever any argument that there was somehow [00:18:41] Speaker 01: adjacent could mean pixels that weren't contiguous to one or two pixels that weren't touching one another, so to speak. [00:18:50] Speaker 01: I've never been pointed to anything in the record about that. [00:18:53] Speaker 01: And I don't believe there was one. [00:18:54] Speaker 01: And it was in the entire case. [00:18:57] Speaker 03: So if you were just looking at two pixels that are adjacent testing [00:19:08] Speaker 03: the relationship between those two, as the SAD system does. [00:19:14] Speaker 03: How would that tell you anything about whether chrominance or luminance vary gradually across adjacent pixels? [00:19:26] Speaker 01: Just looking at two adjacent, in the abstract, it wouldn't tell you much. [00:19:33] Speaker 01: It wouldn't tell you much. [00:19:34] Speaker 01: There might be instances in which [00:19:37] Speaker 01: you come to the conclusion that, under certain circumstances, it might tell you something from an approximation standpoint. [00:19:45] Speaker 01: This is not something that's been discussed. [00:19:47] Speaker 01: I don't understand that. [00:19:48] Speaker 01: OK. [00:19:48] Speaker 01: Well, for example, if you look at Vevelanen, the board made fact findings that it had a three branch strategy. [00:19:59] Speaker 01: And it wasn't flexible about it. [00:20:01] Speaker 01: It thought that that was what it should be doing. [00:20:04] Speaker 01: Many of these fact findings actually had to do [00:20:06] Speaker 01: with the selecting step issue, which the board didn't need to reach. [00:20:10] Speaker 01: But they had three branches. [00:20:12] Speaker 01: They had three carefully selected filters for three carefully selected regions. [00:20:17] Speaker 01: And they were looking at them separately. [00:20:19] Speaker 01: And the question was what you were comparing. [00:20:22] Speaker 01: And then they were doing a min-max approximation. [00:20:24] Speaker 01: They were comparing two pixels. [00:20:27] Speaker 01: They didn't care whether they were next to each other. [00:20:29] Speaker 01: They just had to be in the same region. [00:20:31] Speaker 01: But you could conclude that you might learn something [00:20:35] Speaker 01: approximately about a property. [00:20:40] Speaker 01: A property like elevation and density in other contexts here would be variance. [00:20:47] Speaker 01: Statistical variance is a property that you might be measuring or determining, as the patent says, or level of detail, which is another property, which has to do with [00:21:00] Speaker 01: how many little details there are in the picture. [00:21:03] Speaker 01: Because the whole point of the patent, just take a step back. [00:21:07] Speaker 01: When you are compressing and decompressing each frame, which you're doing millions of calculations, you're doing this very quickly, on very cheap machines. [00:21:17] Speaker 01: So you compress it block by block. [00:21:20] Speaker 01: And then you decompress them block by block. [00:21:22] Speaker 01: And what happens is, because of your quick and dirty calculations, the blocks end up being done a little differently. [00:21:29] Speaker 01: But the point is that the block difference represents a spurious detail, is one way to put it. [00:21:41] Speaker 01: It represents, it looks like it's part of the picture, but it's not. [00:21:44] Speaker 01: So how do you remove it? [00:21:46] Speaker 01: Vevelenian's idea was, oh, I'm going to try very hard to not change any of the differences between [00:21:54] Speaker 01: pixels near that boundary because I think that's going to phase out the details. [00:22:00] Speaker 01: So I'm going to refrain from basing my filtering decisions on how adjacent pixels change relative to each other. [00:22:09] Speaker 01: In fact, avoiding basing it on the level of detail. [00:22:12] Speaker 01: And what they did was they took a reference line from one end of each region to the other, and then if one [00:22:22] Speaker 01: block was higher in luminance than the other, they would just sort of move them like this. [00:22:27] Speaker 01: Everything's sort of trying to be in unison. [00:22:29] Speaker 01: We did it differently. [00:22:30] Speaker 01: We had, after hundreds of tests, we came up with a different strategy. [00:22:35] Speaker 01: We thought you should look, and you should base your calculations on the level of detail on how the things change, and not just on comparing each pixel to the average. [00:22:48] Speaker 01: across all pixels. [00:22:49] Speaker 03: Well, I understand. [00:22:50] Speaker 03: And that certainly is an explanation for the difference between beveling and, let's say, in the SAD system, which you embody in claims two and four. [00:23:01] Speaker 03: But the claim language of claim one of your patent is very broad. [00:23:05] Speaker 03: It talks about determining the level of detail across a region that includes the black boundary. [00:23:14] Speaker 03: A carefully selected region, yes. [00:23:17] Speaker 03: Taking that language just by itself, that would seem clearly to cover Vevolina. [00:23:23] Speaker 03: And the only reason, it seems to me, that you have an argument, which the board adopted, that isn't as broad as it looks on its face, is because you've incorporated your interpretation of level of detail as meaning between including [00:23:40] Speaker 03: analysis between two adjacent, immediately adjacent pixels. [00:23:43] Speaker 01: Am I right so far? [00:23:44] Speaker 01: I disagree with that, Your Honor, in the following sense. [00:23:47] Speaker 03: Which part of that did you disagree with? [00:23:49] Speaker 03: I had several different compound questions. [00:23:51] Speaker 03: OK, well, I apologize. [00:23:53] Speaker 01: Well, it's my fault. [00:23:54] Speaker 01: What I agree with so far is we certainly are incorporating what we mean by looking at adjacent pixels. [00:24:02] Speaker 01: But the detail, the patent is very clear of what it thinks the typical, which the board found as a fact, is the ordinary. [00:24:09] Speaker 01: at undisputed ordinary meaning is that you're looking at how adjacent pixels change relative to one another. [00:24:17] Speaker 01: And Vevelenian absolutely isn't doing that. [00:24:20] Speaker 03: Yeah, but my basic point is that that only comes in by virtue of your incorporation of the term level of detail and the description of level of detail in the specification, not in anything in the claim. [00:24:34] Speaker 03: You would agree with me, I think, that the claim itself [00:24:37] Speaker 03: setting aside the definition of if it's a definition of level of detail is very broad, and broad enough to include double-layment. [00:24:44] Speaker 01: Well, I disagree when you say it's very broad. [00:24:46] Speaker 01: I actually think this is a quite narrow claim. [00:24:49] Speaker 04: What specific language in the claim do you think narrows it to this adjacent pixel? [00:24:56] Speaker 01: Level of detail, Your Honor. [00:24:57] Speaker 04: But that's only because you're incorporating those two sentences that define it further. [00:25:06] Speaker 04: That's not the normal [00:25:07] Speaker 04: That's not the only plain language definition of level of detail, right? [00:25:13] Speaker 04: That it's adjacent pixels. [00:25:14] Speaker 04: No, no, no. [00:25:16] Speaker 04: So let's just look at the claim language, not the other stuff. [00:25:20] Speaker 04: Is there anything in the claim language that incorporates your adjacent pixel definition? [00:25:27] Speaker 01: As far as requiring, it doesn't have the words adjacent pixels in it, of course. [00:25:32] Speaker 04: So you have to look at something in the specification [00:25:35] Speaker 04: to give you some further limitation of that claim? [00:25:38] Speaker 01: Well, we also have the board's finding a fact that a person of ordinary skill in the art would have found this to be the ordinary meaning of the claim, just as the patent says it is. [00:25:49] Speaker 01: The board said that this is the ordinary meaning, i.e., the typical meaning, i.e., the ordinary meaning which no part of the dispute of level of detail. [00:26:00] Speaker 04: So you don't think that the min-max invariance [00:26:02] Speaker 04: formulas are determining a level of detail? [00:26:05] Speaker 01: Absolutely not, your honor. [00:26:06] Speaker 01: And Revolanian says they're not. [00:26:08] Speaker 01: And the board found as a fact that they're not. [00:26:11] Speaker 01: The board specifically found that just because the result, even considering the min-max calculation, the result of that variation does not reflect variation between adjacent pixels. [00:26:24] Speaker 01: No, that's not the same thing. [00:26:27] Speaker 04: You're assuming that level of detail only refers to adjacent pixels. [00:26:31] Speaker 01: True. [00:26:31] Speaker 04: But if it doesn't, then why isn't MinMax and Variance determining the level of detail across the region too? [00:26:38] Speaker 01: Your Honor, okay, but from the very outset in the institution decision, it must assess [00:26:47] Speaker 04: of the difference across adjacent because that's what you know i don't think we're talking and uh... we're not talking to express a lot i'm trying to get you to look at the claim language without the boards claim construction relying on those two cents [00:27:03] Speaker 04: And there's nothing in the claim language itself in claim one that confines it to level of detail determined by adjacent pixels, is there? [00:27:13] Speaker 04: To get that, you have to adopt a definition of level of detail that's based upon the specification. [00:27:20] Speaker 01: My difficulty with your statement, Your Honor, is that it has the word detail. [00:27:25] Speaker 01: It does not have the word detail in it. [00:27:27] Speaker 01: If it didn't have the word detail, we would not be having this conversation. [00:27:30] Speaker 04: Do you think that when you use the phrase level of detail across the region, which is what that claim, it's always talking about between two pixels? [00:27:41] Speaker 01: It's always talking about detail, which means how the pixels in the region vary across adjacent pixels. [00:27:53] Speaker 01: And I would point out, [00:27:55] Speaker 01: Vevelenian specifically says it's avoiding that calculation. [00:27:59] Speaker 01: It doesn't want to know. [00:28:01] Speaker 01: It doesn't want to vary to basis filtering decisions on the difference between adjacent pixels. [00:28:09] Speaker 01: Everyone knew what a detail was. [00:28:11] Speaker 01: Again, there's no dispute that this is the ordinary meaning in the context of the pattern of what a detail means. [00:28:17] Speaker 01: Now detail outside of the context of video, of deblocking, [00:28:22] Speaker 01: Well, obviously that could mean a whole bunch of very broad things. [00:28:27] Speaker 01: But your honor, the fact finding of the board and just logic as a matter of logic, that's not what it means in this context. [00:28:35] Speaker 01: The patent said that's not what it means in this context. [00:28:39] Speaker 01: You have to be looking at adjacent pixels and Bevelenian specifically said, I don't want to, I'm not going to. [00:28:45] Speaker 01: I am not making that determination and I'm doing it for a reason. [00:28:49] Speaker 01: So I would say that [00:28:52] Speaker 01: The word detail is there for a reason. [00:28:56] Speaker 01: And if it wasn't there, it would be calculating a different factor. [00:29:02] Speaker 01: Now, they asked you to say that you should look at the prosecution history. [00:29:08] Speaker 01: They didn't mention the prosecution history for this below. [00:29:12] Speaker 01: They didn't talk about Kim. [00:29:14] Speaker 01: You asked him to lay out Kim for you. [00:29:16] Speaker 01: If only an expert had ever done that. [00:29:19] Speaker 01: If only it had been mentioned below. [00:29:21] Speaker 01: But it wasn't. [00:29:22] Speaker 01: Nobody talked about this. [00:29:23] Speaker 01: Nobody tried to reconcile what Kim was actually measuring, whether Kim was doing the calculations that's in the claim, that's anything like what's in the specification. [00:29:35] Speaker 01: Kim is in the frequency domain. [00:29:38] Speaker 01: You may have noticed from this case that there was quite sophisticated expert testimony on what Vevelanian is doing and what the patent is doing. [00:29:46] Speaker 01: None of that is there for Kim, and that's because it was waived below. [00:29:50] Speaker 01: And it was no evidence on it, and it certainly was disputed. [00:29:53] Speaker 01: So we don't think that that should be a basis for any determination here. [00:29:57] Speaker 01: Now, with respect to figure 5C and Hewlett-Packard and the... Your time has expired. [00:30:13] Speaker 00: Will you just finish your thought, please? [00:30:16] Speaker 01: I would say Hewlett-Packard was only raised now [00:30:20] Speaker 01: on appeal, in fact, in the reply. [00:30:23] Speaker 01: And in the park revision case, Your Honors found that raising Hewlett-Packard in the reply brief before the board was too late, that that kind of sometimes argument doesn't work. [00:30:35] Speaker 01: And it also said that what matters is whether you meet the method, not whether you happen to sometimes meet the same result. [00:30:41] Speaker 00: Thank you, Your Honor. [00:30:43] Speaker 00: Thank you, Mr. Reddewark. [00:30:45] Speaker 00: Mr. Batts has two minutes for a follow-up. [00:30:49] Speaker 04: First respond to that argument, the level of detail is understood by all skilled artisans to mean looking at adjacent pixels. [00:31:00] Speaker 02: That was issue number five. [00:31:01] Speaker 02: So I think we have a huge dispute. [00:31:03] Speaker 02: He's saying there is no dispute on level of detail or what the meaning of level of detail is. [00:31:07] Speaker 02: I mean, obviously, that's one of the primary disputes we do have here, is what is that meaning and what should the construction be? [00:31:13] Speaker 04: But I guess his argument is that with the lane [00:31:16] Speaker 04: doesn't want you to look at level of detail at all somehow. [00:31:21] Speaker 04: Frankly, I had a hard time following it. [00:31:23] Speaker 04: But because of the methods it uses, the variance or the min-max, those aren't level of detail calculations or whatever is the right word. [00:31:35] Speaker 02: That's my point number two. [00:31:36] Speaker 02: So I think it's kind of incredible to me that he's arguing that the calculations of Babel line [00:31:43] Speaker 02: can't be level of detail. [00:31:45] Speaker 02: Like he said, variance is not equal level of detail. [00:31:48] Speaker 02: If we go back to the claim construction that the board adopted, what was that claim construction? [00:31:52] Speaker 02: You look at the level of variation. [00:31:54] Speaker 02: So we're really talking about variance and looking at variation. [00:31:57] Speaker 02: And he's saying those are not related. [00:31:58] Speaker 03: Well, he says well, but it's variation between adjacent pixels. [00:32:05] Speaker 03: Let me make sure that you have an opportunity to answer the first part of Judge Hughes' question, which I'm also interested in. [00:32:13] Speaker 03: is in the skill in the art, or in Vevelinen, that, according to Mr. Weatherwax, disclaims any understanding of level of detail other than level of variation between immediately adjacent pixels. [00:32:28] Speaker 02: Nothing. [00:32:30] Speaker 02: And then I still think 3D and 5D still satisfies that. [00:32:32] Speaker 03: What is there that is to the contrary of that, then? [00:32:37] Speaker 02: I think if you look at Babeline and the 651PAT, they are both looking at the same exact, you're doing the same exact issue. [00:32:45] Speaker 02: You're looking at a mathematical calculation to determine whether a region is low activity, medium activity, or high activity. [00:32:53] Speaker 03: But you're doing it in a different way, as everybody agrees. [00:32:56] Speaker 03: You're doing it in the case of the SAD, which is the basic [00:33:01] Speaker 02: I would disagree. [00:33:02] Speaker 03: This is comparing pixel 1 with pixel 2, pixel 2 with pixel 3, pixel 3 with pixel 4. [00:33:08] Speaker 03: If pixel 1 to pixel 2 is 2 higher, you put in a 2. [00:33:12] Speaker 03: If 2 to 3 is 3 higher, you put in a 3. [00:33:15] Speaker 03: And you add them all up, and that's what SAD is. [00:33:17] Speaker 03: I understand it. [00:33:19] Speaker 03: But what is it that you can point to that says that is not the limit of what detail means in the context [00:33:31] Speaker 03: of a person who's going out for this patent. [00:33:33] Speaker 02: The problem is that your assumption there is what they don't want to say. [00:33:36] Speaker 02: And they've repeatedly said that it's not limited to sad. [00:33:39] Speaker 02: It's not limited to sad. [00:33:40] Speaker 02: They keep on saying this claim is not limited to sad. [00:33:42] Speaker 02: That's the dependent claims. [00:33:45] Speaker 02: But your question is predicated on it being limited to sad. [00:33:48] Speaker 02: If it is limited to sad, then we're in a different world. [00:33:51] Speaker 03: But what I'm asking, what is it that you can point to that shows that [00:33:56] Speaker 03: The level of detail is not in fact limited to SAD. [00:34:02] Speaker 02: Because the dependent claims two and three are the claims that are covering SAT. [00:34:08] Speaker 02: So by, I guess, controlling case law here, [00:34:12] Speaker 02: then the independent claim must be broader than those dependent claims. [00:34:18] Speaker 03: But it seems to me not clear, at least, that this is a straightforward claim differentiation point, because it seems to me that even if you assume the definition of level of detail, as the appellees have presented it, that there is still room to say that there are alternatives to the SAD that still [00:34:41] Speaker 03: go pixel by pixel. [00:34:44] Speaker 02: That's the argument that he would make, I think, in response to your point. [00:34:48] Speaker 02: Yes, and the problem with that argument is there is absolutely nothing in the record below. [00:34:52] Speaker 02: No expert provided anything, no reference, nothing of evidence in the record below. [00:34:58] Speaker 02: that there's some other calculation other than SAT for a between pixel calculation. [00:35:04] Speaker 02: So that's, and when we look at figure 2 of the 651 patent, they're not limiting it to a SAT. [00:35:10] Speaker 02: So they want to say it's not limited to SAT. [00:35:12] Speaker 02: They don't want it to be limited to SAT. [00:35:14] Speaker 02: Well, that would be suicide. [00:35:18] Speaker 02: Yes. [00:35:20] Speaker 02: But yet they're saying it has to be between pixel calculation. [00:35:24] Speaker 02: And I'm sure I'm out of time, so I'm going to be very quick on this. [00:35:27] Speaker 02: So I do think also that they were conflating the concepts in Weiblinen that we discussed earlier. [00:35:31] Speaker 02: There's a mathematical equation for determining a region, whether it's a low, medium, or high region. [00:35:36] Speaker 02: That's one mathematical equation that Weiblinen is teaching. [00:35:40] Speaker 02: The separate, entirely separate process [00:35:42] Speaker 02: is what you do to basically apply a filter on pixels. [00:35:46] Speaker 02: What is the mathematical change that you're doing to those pixels? [00:35:49] Speaker 02: What they were pointing at for Weible-Einand is when you're applying a filter, the mathematical changes that you're making to pixels, rather than what Weible-Einand does for teaching a determination of detail. [00:36:00] Speaker 00: Thank you, counsel. [00:36:01] Speaker 00: As you can see, you're well over your time. [00:36:03] Speaker 00: Thank you, Your Honors. [00:36:04] Speaker 00: You both argued this case at a very high level of detail. [00:36:08] Speaker 00: The case is submitted.