[00:00:02] Speaker 03:
This morning, ladies and gentlemen, we have four cases before the court this morning, one of which has been submitted on the briefs.

[00:00:11] Speaker 03:
The first case before the court is Sure Incorporated versus Clear One, Inc., an appeal from the Patent Trial and Appeal Board.

[00:00:18] Speaker 03:
Ms.

[00:00:19] Speaker 03:
Addy, it's my understanding that you want three minutes for rebuttal?

[00:00:22] Speaker 03:
Yes, Your Honor.

[00:00:23] Speaker 03:
Okay.

[00:00:25] Speaker 03:
You may proceed.

[00:00:26] Speaker 04:
Thank you, Your Honor.

[00:00:27] Speaker 04:
Good morning.

[00:00:28] Speaker 04:
In this obviousness case, the party's dispute

[00:00:31] Speaker 04:
just one element, fixed beams.

[00:00:34] Speaker 04:
And the board recognized that the Kellerman reference disclosed fixed beams.

[00:00:41] Speaker 04:
That's at A20 of the record.

[00:00:43] Speaker 04:
And that should have been the end of the story, because that's the only disputed element.

[00:00:47] Speaker 04:
However, the board legally erred by then relying on irrelevant teachings in Kellerman about fixed beam former design.

[00:00:59] Speaker 03:
found that Kellerman used the phrase fixed beams, but then the board expressly found that it didn't disclose fixed beams within the meaning of the patent that's being challenged, right?

[00:01:10] Speaker 03:
That's a difference.

[00:01:11] Speaker 04:
That is a difference, Your Honor, but when you look at the decision, two points here, when you look at the decision, the decision recognizes the fixed beam disclosure, but then in its analysis that starts on page 24,

[00:01:25] Speaker 04:
refers to figure 13.8, and then it says that the next section is fixed beam former design.

[00:01:31] Speaker 04:
And it doesn't refer to the paragraph below figure 13.8 that deals with fixed beams.

[00:01:39] Speaker 04:
And it doesn't refer to the record at 709 to 711 dealing with fixed beams.

[00:01:45] Speaker 04:
And there are other places in Kellerman that describe fixed beams that it doesn't refer to.

[00:01:49] Speaker 04:
It only refers to this fixed beam former design.

[00:01:54] Speaker 04:
One point is, because of the inconsistency in recognizing that they're fixed beams disclosed, but then not referring to them in the analysis and finding that they're not disclosed, that's a legal error.

[00:02:05] Speaker 01:
Well, as I understand what the board was saying was that it understood the claimed inventions referenced to a fixed beam to be referencing that the beams are fixed for the entirety of a teleconference.

[00:02:21] Speaker 01:
But for the prior art,

[00:02:23] Speaker 01:
When it refers to fixed beam, it's not talking about a beam being fixed for the entirety of a teleconference.

[00:02:29] Speaker 01:
It's only fixed with respect to a fairly short time interval in which, from the perspective of the echo cancellation module, it looks fixed.

[00:02:42] Speaker 01:
But over the course of the entire

[00:02:45] Speaker 01:
teleconference, it's not actually fixed.

[00:02:47] Speaker 01:
There's some kind of change going on.

[00:02:49] Speaker 01:
Either the individual beams themselves are somehow being adjusted according to some function or there's a different selection of a subset of different beams that are being used and employed during the teleconference.

[00:03:03] Speaker 01:
And so in either scenario, the beams are not fixed over the course of the entire teleconference.

[00:03:10] Speaker 01:
So fixed means

[00:03:13] Speaker 01:
Yes, it has a commonly understood meaning, but fixed as to what reference point.

[00:03:18] Speaker 01:
And for the board, they said the claimed invention is for the entirety of the teleconference.

[00:03:23] Speaker 01:
And for the reference, it's only perhaps with respect to how the echo cancellation module would see the beam.

[00:03:31] Speaker 01:
So could you speak with specificity to that?

[00:03:35] Speaker 04:
Yes, Your Honor.

[00:03:36] Speaker 04:
First of all, the claim construction out of the board was that it's fixed before a conference.

[00:03:42] Speaker 04:
And then the way the board applied that construction to Kellerman was that it had to be fixed during a conference.

[00:03:50] Speaker 04:
And in our claim construction section, we discussed that before a conference leaves ambiguity about what happens to the claim during the conference and what happens to the claim to the beam at any time other than before a conference.

[00:04:04] Speaker 03:
Well, I mean, Kellerman's own testimony, I mean, it was your expert, but Kellerman's own testimony

[00:04:10] Speaker 03:
said it could be fixed for milliseconds.

[00:04:14] Speaker 03:
In other words, it's just barely fixed, and then it's not fixed.

[00:04:19] Speaker 03:
I mean, the board said that to try to call that a fixed beam is really expanding the scope of the reference.

[00:04:28] Speaker 04:
Two points, Your Honor.

[00:04:29] Speaker 04:
First of all, Kellerman is talking about this other embodiment that starts on 7-11 and goes to 7-12 about fixed beam former design.

[00:04:39] Speaker 04:
And the design is not in the claim.

[00:04:42] Speaker 04:
The claim only says fixed beam.

[00:04:44] Speaker 04:
If you look at 709 to 711, Kellerman is explicit in where he says that time varying beams cannot be successfully tracked by AEC.

[00:04:57] Speaker 04:
And he says in his invention, fixed beams, the AEC acts only on the output of the fixed beam former.

[00:05:05] Speaker 04:
That is the relevant disclosure.

[00:05:07] Speaker 04:
What he's talking about in his testimony is referring to fixed-beam design.

[00:05:12] Speaker 04:
And also, it's taken out of context, because when you read his testimony in context, he says, that's not what I was describing.

[00:05:19] Speaker 04:
Look at 709 to 711 to see what I was describing.

[00:05:25] Speaker 03:
It's pretty odd to have the petitioner ask for a narrower construction.

[00:05:30] Speaker 03:
And I'm not sure that I completely understand why your much narrower construction

[00:05:36] Speaker 03:
matters to the ultimate resolution of this case.

[00:05:41] Speaker 04:
Your Honor, there are two reasons.

[00:05:43] Speaker 04:
First, it doesn't matter to the ultimate resolution of this case because no matter what the construction is, fixed beams are the same in Kellerman and the disclosed in the 553 patent.

[00:05:56] Speaker 04:
However, you have to do a claim construction in order to make this decision.

[00:06:00] Speaker 04:
And there are downstream litigations dealing with both the 553

[00:06:05] Speaker 04:
and its progeny.

[00:06:07] Speaker 04:
And so to the extent that you have to rule on claim construction in making this decision, district court judges may consider your decision.

[00:06:16] Speaker 04:
Even though it's Brie, it still may be considered, and that's why it's important.

[00:06:21] Speaker 01:
But just to clarify, I think you said, and the other side also said, and so that makes it undisputed, that the actual claim construction dispute here doesn't matter as to the question on appeal as to whether Kellerman teaches this limitation, right?

[00:06:41] Speaker 04:
No, Your Honor.

[00:06:41] Speaker 04:
It doesn't matter.

[00:06:42] Speaker 04:
The limitation is taught.

[00:06:44] Speaker 04:
And that's because Kellerman and the 553 patent

[00:06:49] Speaker 04:
teach the same problem and teach the same solution of the problem.

[00:06:52] Speaker 04:
To reduce complexity and to maintain audio quality, both of them teach a fixed beam former connected to an AEC.

[00:07:03] Speaker 03:
So you'd be asking us, though, to basically give an advisory opinion that even that, which wouldn't be binding because it doesn't use the same standard as district courts.

[00:07:17] Speaker 04:
I disagree that it would be an advisory opinion simply because the court has to do a claim construction to reach invalidity.

[00:07:25] Speaker 04:
However, if the court doesn't have to do a claim construction to reach the decision on invalidity, then it doesn't have to happen.

[00:07:36] Speaker 04:
But I believe that the first step in invalidity is a claim construction.

[00:07:38] Speaker 03:
Well, generally that is always true.

[00:07:39] Speaker 03:
But if both parties agree that under either construction, the analysis is identical,

[00:07:45] Speaker 03:
we wouldn't necessarily have to pick between the two, right?

[00:07:48] Speaker 04:
I think that depends on the ultimate result.

[00:07:51] Speaker 04:
No, I don't even think it depends on that.

[00:07:56] Speaker 04:
Actually, Your Honor, simply because you render a Brie construction, other courts are going to look at that construction in dealing with the litigation.

[00:08:08] Speaker 04:
And while we all know that Brie doesn't have to be the same as a Phillips construction,

[00:08:17] Speaker 04:
At this point, in this case, at this time, the difference between the two is not at issue.

[00:08:29] Speaker 04:
We asked about the disclosure of time-invariant beams, and I believe you were concerned about whether the fixed beams are the same.

[00:08:42] Speaker 04:
Kellerman's disclosure of fixed beams, it can be found at 702 of the record in detail.

[00:08:48] Speaker 04:
It's also at 708, and as I've mentioned before, 709 to 711.

[00:08:53] Speaker 04:
But Kellerman is very specific that time-varying beam forming cannot be trapped satisfactorily with AEC in the first sentence.

[00:09:02] Speaker 04:
That's what Kellerman teaches.

[00:09:04] Speaker 04:
So it doesn't make sense that then the board would say, well, this doesn't teach fixed beam forming.

[00:09:12] Speaker 04:
Also, under figure 13.8, Kellerman refers to that figure and says that the AEC relies only on the time-invariant part.

[00:09:22] Speaker 04:
If you look at figure 13.8, Your Honor, you can see that from the fixed beam former, the M fixed beam signals go to the AEC.

[00:09:33] Speaker 04:
And there's nothing that discloses that anything else goes to the AEC.

[00:09:37] Speaker 01:
Finally, I want to address the... Beam design and control module affect what kind of beams go into the AEC?

[00:09:49] Speaker 04:
No, it affects the coefficients of the beams, and the testimony of the experts was... When you change the coefficients, though, of the beams, aren't you going to actually end up changing the character of the beams?

[00:10:02] Speaker 04:
Yes, but in talking about that in the beam design and control section,

[00:10:06] Speaker 04:
The Kellerman always uses permissive language, may.

[00:10:10] Speaker 04:
And at one point, it says it may occur during the initial training phase only.

[00:10:18] Speaker 04:
And the testimony of the experts is in agreement that during the initial training phase is before a conference.

[00:10:24] Speaker 04:
You can see Dr. Loy's testimony at A2420 to 2421.

[00:10:34] Speaker 01:
Do you agree that if a different subset of beams get selected during the course of the teleconference from, say, the initial number of subset of beams that are selected at the beginning of the teleconference, then if that is how Kellerman should be understood, then Kellerman isn't teaching fixed beams as the term fixed beams has been construed by the board?

[00:11:00] Speaker 04:
I disagree, because I think Kellerman is teaching two different things.

[00:11:04] Speaker 01:
Yeah, this is a hypothetical.

[00:11:06] Speaker 01:
If we understand Kellerman to be disclosing that, switching up the subset of beams that are being used and relied on and echo canceled, then is it your view that that still nevertheless meets the claim limitation of fixed beams as that term fixed beams has been construed by the patent board?

[00:11:30] Speaker 04:
No, Your Honor.

[00:11:32] Speaker 04:
Actually, I think that's where the construction of fixed beams is hard to understand and ambiguous, because it just says fixed before a conference.

[00:11:42] Speaker 04:
And if you rule that the beams can be changing during a conference, then that still fits under the claim construction, because the claim construction provided by the board says nothing about

[00:11:55] Speaker 04:
what happens after before a conference.

[00:11:57] Speaker 04:
It just says before conference.

[00:11:59] Speaker 04:
And so in an adaptive beam situation, Your Honor, it is fixed before a conference.

[00:12:05] Speaker 04:
But then during a conference, the adaptive beam moves around and adapts to the signal situation.

[00:12:12] Speaker 04:
And that was explicitly excluded in the prosecution history.

[00:12:17] Speaker 04:
The prosecution history at 8, 10, 12 of the record

[00:12:22] Speaker 04:
clear one said, fixed beams does not mean adaptive and does not mean adjustable.

[00:12:28] Speaker 03:
The disclosure that's in that prosecution history says they're not steered beams.

[00:12:36] Speaker 03:
That's different than the type of changeable beams that Kellerman disclosed.

[00:12:44] Speaker 04:
That's correct.

[00:12:47] Speaker 04:
The parties, I think, agree that steered beams is the adaptable, adaptive type of beams.

[00:12:53] Speaker 04:
But if you read the description or the statement that the patentee makes at A1012, he says they're not adjustable either.

[00:13:02] Speaker 04:
And adjustable beams are when you're manually changing the beams instead of adaptive, where it's moving around to detect who's talking at a certain time.

[00:13:14] Speaker 00:
So what's the difference between a design and the operation of an adaptive system?

[00:13:20] Speaker 04:
Well, there are two different embodiments in Kellerman.

[00:13:22] Speaker 04:
And Kellerman says in the first one, we're talking about fixed beams.

[00:13:28] Speaker 04:
And then Kellerman says in the second one, we're talking about ways that you can optimally adjust these fixed beams.

[00:13:35] Speaker 04:
You don't have to do it during a conference.

[00:13:38] Speaker 04:
And that doesn't change anything in the earlier discussion.

[00:13:41] Speaker 04:
He says you can do it only

[00:13:43] Speaker 04:
during the initial training phase.

[00:13:45] Speaker 04:
And so that's the difference.

[00:13:48] Speaker 04:
And I think it's in the discussion in Kellerman.

[00:13:51] Speaker 00:
So in figure 13.8, when I'm looking at being designed and controlled, am I also considering that that's the adaptive system?

[00:14:01] Speaker 04:
It's interesting, because Kellerman uses adaptive, but he uses adaptive in a different way than Kajala, which was the prior art reference in the prosecution history.

[00:14:11] Speaker 04:
So Kellerman's adaptive

[00:14:12] Speaker 04:
is really talking about adjusting the beam coefficients.

[00:14:17] Speaker 04:
But again, I want to reiterate that it says that that can be done optionally.

[00:14:23] Speaker 04:
And Kellerman, it's clear that the less that you adjust it, the better.

[00:14:28] Speaker 04:
The first part saying no adjustment.

[00:14:30] Speaker 04:
If you have no more questions, I'll reserve my time.

[00:14:33] Speaker 04:
OK, thank you.

[00:14:51] Speaker 02:
Good morning, your honors, and may it please the court.

[00:14:56] Speaker 02:
The only question your honors need to answer on this appeal is whether substantial evidence supports the board's factual finding that the beam former in figure 13.8 of Kellerman does not have parameters that remain fixed throughout a conference.

[00:15:12] Speaker 02:
And there is copious evidence to support that conclusion.

[00:15:15] Speaker 03:
That's the construction that the district court ultimately gave.

[00:15:19] Speaker 03:
But where is the support for the board's conclusion that they have to be fixed before the conference, or the parameters for the beam have to be fixed before the conference?

[00:15:34] Speaker 02:
Well, that is the board's construction, number one.

[00:15:38] Speaker 02:
The support for that comes from the specification of the 553 patent.

[00:15:42] Speaker 02:
Reading that, and both the district court and the board read it to understand correctly, that the reason why the beams are fixed is to prevent the AECs from having to readapt to changing beams.

[00:15:56] Speaker 02:
And the time when that matters is during a conference.

[00:15:59] Speaker 02:
And it doesn't matter outside of a conference.

[00:16:02] Speaker 02:
And so that's the basis for the construction.

[00:16:05] Speaker 02:
I also want to address this.

[00:16:06] Speaker 03:
But your view is that the board's construction is really saying they're set before the conference and then remain fixed during the conference.

[00:16:15] Speaker 02:
Exactly, Your Honor.

[00:16:16] Speaker 02:
It's just different ways of saying the same thing.

[00:16:19] Speaker 02:
And in fact, that is the way the board articulated its construction in terms of let's fix it before a conference.

[00:16:26] Speaker 02:
That's what the board's construction is.

[00:16:28] Speaker 02:
But as applied, the board applied the construction as, OK, if it's fixed before the conference, it's unchanging during the conference.

[00:16:37] Speaker 02:
It's really the same thing expressed two different ways.

[00:16:40] Speaker 03:
So what's your response to Ms.

[00:16:43] Speaker 03:
Addy's contention that aspects of the construction were actually given up during prosecution history?

[00:16:57] Speaker 02:
Well, during the prosecution of 553 patent, the Kajala reference was certainly distinguished.

[00:17:03] Speaker 02:
Kajala teaches continuous steering of beams, adaptive beam forming.

[00:17:09] Speaker 02:
And the applicants simply noted that that is one example of beam forming that is not fixed.

[00:17:18] Speaker 02:
That was certainly distinguished.

[00:17:21] Speaker 02:
That's all.

[00:17:22] Speaker 02:
That's all you can read from the prosecution history.

[00:17:28] Speaker 02:
So there are two ways to look at Kellerman.

[00:17:30] Speaker 02:
The first is open the reference, see the words.

[00:17:33] Speaker 01:
Kellerman structurally is pretty close to the claimed invention, right?

[00:17:36] Speaker 01:
It's doing the beamformer block first, then it's doing the acoustic echo cancellation second.

[00:17:44] Speaker 01:
And in the beamformer block, it's taking all the signals from all the different microphones

[00:17:51] Speaker 01:
and then making a certain number of beams that's less than the number of microphones.

[00:17:56] Speaker 01:
It's not just making a single beam, for example.

[00:17:59] Speaker 01:
Right?

[00:18:00] Speaker 01:
Correct, Your Honor.

[00:18:00] Speaker 01:
So all of these elements are just like your claimed invention.

[00:18:05] Speaker 02:
True.

[00:18:05] Speaker 02:
But the one distinction, the patentable distinction, is that Kellerman does not use fixed beams as the 553 patent.

[00:18:14] Speaker 01:
It says fixed beams.

[00:18:15] Speaker 02:
It says fixed beams, but what does it mean?

[00:18:19] Speaker 01:
what it means and there's a statement in Kellerman that says you know this system can be used for initial trial phase set up or the system can be used continuously and I assume continuously is in reference to over the course of the operation of the entire during a teleconference sure like Kajala so then it comes down to whether Kellerman is talking about

[00:18:45] Speaker 01:
a second embodiment where it's only operating and creating these beams during the initial trial phase, which sounds like something that you do during the setup for a conference, i.e.

[00:19:01] Speaker 01:
before a conference.

[00:19:03] Speaker 01:
So why does the board get to read that in the opposite?

[00:19:09] Speaker 02:
Because it was Schur's burden below to show that the initial training phase

[00:19:14] Speaker 02:
would happen before a conference begins.

[00:19:16] Speaker 02:
And they failed to do that.

[00:19:18] Speaker 02:
In fact, Kellerman is silent as to when this initial training phase occurs.

[00:19:24] Speaker 02:
And there is an admission from Schur's counsel below during the oral argument to that effect.

[00:19:30] Speaker 02:
And you can see that on appendix page 8996.

[00:19:33] Speaker 02:
It's on lines 13 and 14 of that page.

[00:19:38] Speaker 02:
And that was critical to the board's reasoning.

[00:19:43] Speaker 02:
And there's confirmation of that elsewhere in the record.

[00:19:46] Speaker 02:
I want to direct your honor to appendix page 6021.

[00:19:52] Speaker 02:
This is testimony from Dr. Kellerman during one of his depositions when he was explaining exactly what we're talking about.

[00:19:59] Speaker 02:
And he said, quote, we would initially run adaptive beam

[00:20:04] Speaker 02:
an adaptive beam former which picks our two voices mainly, right, and learns the background noise statistics.

[00:20:10] Speaker 02:
And then we would use it as a fixed beam former for the rest of the conference.

[00:20:15] Speaker 02:
What he's clearly saying here is that this initial training phase that determines what the beam should be happens after the conference has begun.

[00:20:24] Speaker 01:
Like maybe the first couple minutes of a conference?

[00:20:27] Speaker 02:
It's probably more like the first few seconds.

[00:20:30] Speaker 02:
I mean, depending on the technology and how well the algorithms perform, it could be fractions of a second.

[00:20:39] Speaker 02:
But the point is, it is definitely happening after the conference has begun because it has to.

[00:20:46] Speaker 02:
It has to use the talker's voices to do that initial adaptation.

[00:20:52] Speaker 02:
One important point here is that when we're talking about adaptive systems,

[00:20:56] Speaker 02:
They adapt their coefficients their parameters based upon use operation There's not a separate design and operation of an adaptive filter It is if a person comes in into the room in the middle of a conference call and sits down and speaks then the adaptive process starts all over

[00:21:18] Speaker 02:
It could yes now in the 553 patent its system would handle that differently because it uses fixed beams and A different fixed beam which probably had not been utilized before would pick up that new talker in the position where he or she is seated and Sometimes when that person is talking that fixed beam would be selected How do you understand Kellerman to handle that exact same situation?

[00:21:45] Speaker ?:
I?

[00:21:45] Speaker 02:
In Kellerman's situation, because the beams are fundamentally adaptive, but just held fixed for brief periods of time to allow the AECs to converge to their proper coefficients, what would happen is one of the beams, we don't know which one, would adapt to find the new speaker.

[00:22:06] Speaker 02:
And during that time, Kellerman would

[00:22:12] Speaker 02:
There would be leak through in the AECs, but then Kellerman would hold the beams fixed while the AECs adapt to that new situation.

[00:22:25] Speaker 01:
Would it select a new subset of beams?

[00:22:29] Speaker 02:
It could, very possibly.

[00:22:31] Speaker 02:
Yes.

[00:22:33] Speaker 02:
And if the selection of the subset is during the conference,

[00:22:38] Speaker 02:
It's a change in the beat.

[00:22:39] Speaker 01:
When you say it could do this or it may do that, it makes me a little nervous as to what actually the reference teaches and discloses to one of scale and the art.

[00:22:50] Speaker 02:
Sure, Your Honor.

[00:22:51] Speaker 02:
So the board never found affirmative disclosure

[00:22:56] Speaker 02:
that Kellerman keeps the beams fixed for the entire duration.

[00:23:00] Speaker 02:
And yes, granted, there are statements in Kellerman that discuss, using permissive language, the time varying nature of its beams.

[00:23:10] Speaker 02:
But the key thing is,

[00:23:13] Speaker 02:
Kellerman never affirmatively says, keep the beams fixed for the entire conference.

[00:23:18] Speaker 02:
It was Schur's burden to prove that that is disclosure in Kellerman, and they couldn't do it.

[00:23:23] Speaker 02:
The board reasonably found that that disclosure is not there.

[00:23:27] Speaker 03:
What do we do with the fact that once apparently they decided that Dr. Kellerman was going off the reservation, they got a new expert

[00:23:39] Speaker 03:
And the new expert said, well, let's really look at what one of skill in the art would understand from this disclosure rather than having someone try to teach us what they might have been thinking at the time.

[00:23:51] Speaker 03:
And that new expert expressly said that one would understand that there are periods of use of fixed beams.

[00:24:02] Speaker 02:
but not periods that are coextensive with the entire duration of the conference.

[00:24:08] Speaker 02:
And you hit on an important point, Judge O'Malley.

[00:24:11] Speaker 02:
They threw Dr. Kellerman under the bus for two reasons.

[00:24:14] Speaker 02:
He said two things highly damaging to Schur's case.

[00:24:17] Speaker 02:
First, he testified in his deposition that

[00:24:22] Speaker 02:
When when the word fixed is used it is always in reference to some observation interval and the observation interval that is relevant to the Kellerman reference is on the order of seconds and he said that

[00:24:44] Speaker 02:
For example, he said at page 5993, quote, time invariant and fixed always is associated to a time interval that you consider.

[00:24:54] Speaker 02:
So nothing is forever.

[00:24:55] Speaker 02:
So time invariant or fixed beamforming may be fixed for only a certain period of time, say 10 seconds comma 100 milliseconds.

[00:25:05] Speaker 03:
Lower.

[00:25:05] Speaker 03:
And that could just be a logical explanation that whenever you have

[00:25:09] Speaker 03:
something that's time invariant.

[00:25:11] Speaker 03:
It could be time invariant for two seconds.

[00:25:13] Speaker 03:
It could be time invariant for, you know, two days.

[00:25:18] Speaker 03:
But the real question is, in the use of these beams, did it actually teach time invariance throughout the conference?

[00:25:29] Speaker 02:
No, the Kellerman reference does not.

[00:25:32] Speaker 02:
That's why the board scoured the reference.

[00:25:36] Speaker 02:
to look for some disclosure of holding the beams fixed for the entire conference.

[00:25:41] Speaker 02:
And they couldn't find that.

[00:25:42] Speaker 02:
And there's reasonable support for that.

[00:25:45] Speaker 02:
It's in Kellerman's testimony, Mr. Kellerman the man, Dr. Kellerman.

[00:25:50] Speaker 02:
It comes from the document itself.

[00:25:52] Speaker 02:
And it also comes from the testimony of Dr. Loy, Clear One's expert, who provided testimony about how one of skill and the art would read the Kellerman reference.

[00:26:02] Speaker 02:
And his reasoning tracks almost exactly the board's reasoning.

[00:26:08] Speaker 03:
So does Kellerman steer the beams, much like the prior art that Clear One distinguished in its prosecution history?

[00:26:20] Speaker 02:
Certainly.

[00:26:21] Speaker 02:
Either beams are selected and the selection happens or changes during the conference, or the beams are steered and adapted during the conference.

[00:26:35] Speaker 03:
So the beams respond to the

[00:26:38] Speaker 02:
background noises and the voices Yes, they do it's adaptive beamforming fundamentally Kellerman is about adaptive beamforming and how to integrate AEC into a time varying beamforming and That that's the default for Kellerman We know that because the relevant section with figure 13.8 is section 13.5 beginning on page 709

[00:27:05] Speaker 02:
And it refers to integration of time-varying beamforming and AEC.

[00:27:11] Speaker 02:
Time-varying beamforming is described in the immediately preceding paragraph on that same page, 709.

[00:27:20] Speaker 02:
And there, Dr. Kellerman as the author is talking about the fact that you have signal-dependent time-varying beamformers

[00:27:32] Speaker 02:
That's the context.

[00:27:33] Speaker 02:
He even says that, look, to make this work well for the AECs, you have to hold the echo path model constant for a sufficiently long time.

[00:27:45] Speaker 02:
And if you look at the diagram on the next page, especially part C, you can see that adaptation is enabled in this example of five and a half seconds of a conference.

[00:27:59] Speaker 02:
It's enabled for half a second.

[00:28:01] Speaker 02:
and the beam formers are held fixed for the other five seconds.

[00:28:06] Speaker 02:
And it's during the time when the beam formers are held fixed that the AEC adaption is enabled.

[00:28:15] Speaker 01:
Is it your understanding that when Kellerman talks about how he can monitor the fixed beams and learning of optimum beam formers for deciding upon the G function, and that can be carried out during initial training phase only?

[00:28:32] Speaker 01:
that we're just talking about those first few milliseconds.

[00:28:35] Speaker 01:
And so everything else that you're talking about during the course of the teleconference, those things are for a separate independent teaching in Kellerman.

[00:28:49] Speaker 01:
It's not relevant at all to this particular disclosure embodiment about the initial training phase only.

[00:28:59] Speaker 02:
It is really all we have to really understand and focus and zero in on is the initial training phase only embodiment I think that's One of the stronger embodiments for sure, but it falls short because as we've discussed the initial training phase likely happens during a conference at the at least the initial start of a conference for

[00:29:28] Speaker 02:
milliseconds a few seconds who knows how long but it doesn't matter how long it's during a conference and then the other important finding that the board found is that Even assuming for the sake of argument that this initial training phase happens before the conference the board could find no evidence in Kellerman that the beams would remain fixed throughout the duration of the conference and

[00:29:55] Speaker 02:
And that's a finding that the board made, and there's substantial evidence in support of that.

[00:29:59] Speaker 03:
I think your time is up.

[00:30:02] Speaker 02:
Thank you, Your Honors.

[00:30:06] Speaker 03:
I'll give you two minutes for rebuttal.

[00:30:08] Speaker 02:
Thank you.

[00:30:09] Speaker 04:
I have three points.

[00:30:11] Speaker 04:
Hopefully I can make them.

[00:30:12] Speaker 04:
First, counsel focused on could and may repeatedly in his discussion.

[00:30:18] Speaker 04:
And that's because everything in the design section of Kellerman is optional.

[00:30:23] Speaker 04:
And he describes it as could or may or can.

[00:30:27] Speaker 04:
Secondly, counsel argues that Kellerman's system is a time-varying system.

[00:30:33] Speaker 04:
However, that's interesting, because so is the claimed 553 system.

[00:30:39] Speaker 04:
Here's where it goes.

[00:30:40] Speaker 04:
There's a time-invariant beamformer in both.

[00:30:44] Speaker 04:
There's also an AEC that's time-varying.

[00:30:47] Speaker 04:
There's also, in the claim, it talks about the selection process at the multiplexer.

[00:30:52] Speaker 04:
And in the Kellerman, it talks about voting.

[00:30:55] Speaker 04:
That's time varying.

[00:30:57] Speaker 04:
However, and all of those things are not disputed.

[00:31:00] Speaker 04:
They're present in Kellerman and in the 553 patent.

[00:31:03] Speaker 04:
We're only talking about the time invariant beam former.

[00:31:07] Speaker 04:
And Kellerman is specific that you can't track time varying beams with the AEC.

[00:31:15] Speaker 04:
So you're using time invariant.

[00:31:16] Speaker 04:
With respect to the claim construction, the claim itself has no timing restrictions.

[00:31:22] Speaker 04:
And the only construction that deals with the prosecution history and that deals with what Clear One was trying to distinguish is a construction that recognizes that the claim is not, that fixed beams are not adjustable.

[00:31:41] Speaker 04:
Your Honor, this appeal turns on one issue.

[00:31:44] Speaker 04:
And that issue is whether Kellerman discloses fixed beams.

[00:31:47] Speaker 04:
It does.

[00:31:49] Speaker 04:
It does at page 709.

[00:31:51] Speaker 04:
It does at figure 13.8.

[00:31:54] Speaker 04:
It does from 709 to 711.

[00:31:57] Speaker 04:
And that's the end of the story.

[00:32:00] Speaker 04:
Thank you, Your Honors.

[00:32:02] Speaker 04:
Thank you.