[00:00:00] Speaker 02:
This is case number 21, 1764.

[00:00:02] Speaker 02:
We're going to draw inventions against Google LLC.

[00:00:09] Speaker 01:
Okay, I am there.

[00:00:13] Speaker 02:
Okay, press save when ready.

[00:00:17] Speaker 01:
Okay, so in this case, we are talking about a set of patents that have to do with when you add time scale modification to a system that doesn't have time scale modification.

[00:00:29] Speaker 01:
And the goal of this, when you add this, is that data time and presentation time need to be kept track of.

[00:00:38] Speaker 01:
Because in prior systems, you only had one time value.

[00:00:43] Speaker 01:
And in order to sync up the audio and video, you needed two different time values.

[00:00:50] Speaker 01:
You needed to know where you were in the stream once you speeded up the audio.

[00:00:55] Speaker 01:
So consensually, what happens in the context of all of the patents in this family is that the audio is sped up.

[00:01:03] Speaker 01:
We believe we need time scale modifications certainly in the context of the O5O patents, which call the time scale modification, whereas the others do not.

[00:01:13] Speaker 01:
And so it's another reason why time scale modification is something different than just speeding up and slowing down.

[00:01:19] Speaker 01:
But all of the patents deal with

[00:01:22] Speaker 01:
you sped up the audio, you do that processing first, and then you need both time values when you are displaying the information.

[00:01:33] Speaker 01:
So I will not, you know, the first argument, of course, is the median time scale modification that impacts claims one through four, the 20, 25, 31, 32, 34, 36, 40 through 43 and 45.

[00:01:49] Speaker 01:
And it's our position that

[00:01:52] Speaker 01:
time scale modifications should be interpreted as we have proposed, and that none of those claims are invalid in view of that construction.

[00:02:01] Speaker 01:
But there's a completely separate argument with regard to all of the patents in the 108 family, which is the subject of 1764, and that is that the board essentially read a limitation out of the claims,

[00:02:22] Speaker 01:
So if we look at the claims themselves and any of the patents in this family, the board basically rewrote the claims by importing into the claims a construction of reading out during rendering, essentially.

[00:02:43] Speaker 01:
So they made the presentation time from the beginning of the presentation

[00:02:49] Speaker 01:
And they made the data time for just the portion, which it's always going to be.

[00:02:56] Speaker 01:
And so in the context of the Nelson reference, as we set forth in our papers, it is never the case that Nelson presents a unit when the data time and the presentation time are not equal.

[00:03:17] Speaker 01:
In order for Nelson to render, to present anything, the data time and the presentation time must be equal in every embodiment.

[00:03:26] Speaker 01:
And that includes when you speed it up, because in the context of the Nelson reference, as Google's expert testified, you just make the clock go twice as fast.

[00:03:37] Speaker 01:
But in all cases, then you're going to do multiplication, and the data time is always going to equal the presentation time when a unit is presented.

[00:03:47] Speaker 01:
And so in each of the claims, for example, the 050 patent, in each of them they have a requirement that the data time parameter is not equal to the value of the presentation time.

[00:04:04] Speaker 01:
And so essentially what the board here has said, well, it doesn't matter if I'm sitting for 20 seconds.

[00:04:10] Speaker 01:
So let's take the example that at the beginning of the specification of the patents in this family, it says,

[00:04:17] Speaker 01:
If you present something that should be 60 seconds long, twice as fast, the presentation time is going to be 30 seconds, but the data time is going to be 60 seconds.

[00:04:30] Speaker 01:
And in the board's interpretation of Nelson, they say, well, presentation time is from the beginning of whenever the content starts being presented.

[00:04:41] Speaker 01:
So that means that if there is a 20 second gap,

[00:04:45] Speaker 01:
then the presentation time will be 80 seconds, even though nothing is being presented during that time.

[00:04:53] Speaker 01:
And so then when Nelson decides to present that unit, then they say, well, it's not equal, because now it's, so the presentation time has been 80 seconds, but the data time is a different value.

[00:05:07] Speaker 01:
But the problem with that is the claims, because the claims all require a portion

[00:05:12] Speaker 01:
So it has to be, you have to be looking at data time and presentation time with regard to the same portion that isn't rendered.

[00:05:22] Speaker 01:
And so in the context of the 050 patent, for example, we look at the claim one, which is in column 23, starting at line 35.

[00:05:35] Speaker 01:
And it says the element A is maintaining a value of a presentation time parameter

[00:05:41] Speaker 01:
tangibly stored in a second computer medium and representing an amount of time elapsed during rendering of a portion of a temporal sequence of presentation data.

[00:05:54] Speaker 01:
So the time has to be the time during rendering, not time sitting idle, not time just deleting segments as Nelson does.

[00:06:03] Speaker 01:
And then, in addition to that, you have to maintain

[00:06:06] Speaker 01:
value of a data time parameter, tangibly stored in a third computer radio medium and representing an amount of time representing rendering system to render the portion, the same portion.

[00:06:19] Speaker 01:
So the way that the board has read Nelson, it's not the same portion.

[00:06:24] Speaker 01:
It's either not the same portion or it's always equal.

[00:06:28] Speaker 01:
There is no case in Nelson where it actually can either have

[00:06:35] Speaker 01:
you can't satisfy both not equal and the same portions.

[00:06:41] Speaker 01:
And so it just does not practice any of the cleanings of the 903 family.

[00:06:50] Speaker 01:
And I don't think there's actually a great report to this.

[00:06:57] Speaker 01:
The PTAB

[00:07:02] Speaker 01:
trying to explain it by saying, well, you can include things that happened after the information was presented.

[00:07:13] Speaker 01:
So you can look at, oh, those are still part of the sequence, but they were deleted.

[00:07:18] Speaker 01:
But that is not consistent with the claims.

[00:07:22] Speaker 01:
When the unit is deleted, it's no longer part of the portion.

[00:07:26] Speaker 01:
And so there is no embodiment in Elson where

[00:07:30] Speaker 01:
Data time and presentation time are not equal for a portion that is actually rendered.

[00:07:39] Speaker 01:
And the way in which the p-tab read during rendering out of the claim was improper and inconsistent with the claim language.

[00:07:51] Speaker 01:
Now, I understand Google is going to say there's a definition of presentation time at the beginning of the specification.

[00:07:58] Speaker 01:
But we still have the claims control.

[00:08:00] Speaker 01:
And so that definition of presentation time at the beginning of the specification is it's relating to, it doesn't preclude relating to a portion.

[00:08:13] Speaker 01:
So sure, it's the beginning of the content presentation, but in the context of the claims, the beginning of the content presentation has to be for a portion.

[00:08:23] Speaker 01:
And it has to be the same portion and the data time and the

[00:08:28] Speaker 01:
Presentation time must not be equal.

[00:08:30] Speaker 01:
And that is not satisfied by nationality.

[00:08:33] Speaker 01:
And so I will reserve the rest of my time, unless anyone has any questions.

[00:08:37] Speaker 02:
Yes.

[00:08:39] Speaker 02:
All right.

[00:08:39] Speaker 02:
Thank you.

[00:08:41] Speaker 02:
Mr. Zalberger, apologies for not getting it straight.

[00:08:46] Speaker 00:
No problem, Your Honor.

[00:08:48] Speaker 00:
May it please the court.

[00:08:50] Speaker 00:
The challenge patents specifically define the term presentation time.

[00:08:55] Speaker 00:
they provide at appendix 221, column 2, lines 1 to 4, that presentation time is a time elapsed since the beginning of the media content presentation.

[00:09:07] Speaker 00:
All the during rendering clause provides is that it keys the presentation time and the data time parameters to the same media work.

[00:09:17] Speaker 00:
You have a single media presentation at a given time.

[00:09:20] Speaker 00:
You're going to start the presentation.

[00:09:21] Speaker 00:
It's going to be proceeding.

[00:09:23] Speaker 00:
And the during rendering clause ensures that the presentation time and the data time are measuring that same presentation.

[00:09:30] Speaker 00:
And I think Dr. Bonsallet made this absolutely clear at his deposition.

[00:09:34] Speaker 00:
He asked him what the during rendering clause means.

[00:09:38] Speaker 00:
And that Appendix 1508, this is page 70, lines 13 to 18 of his deposition transcript, we asked him what during rendering means in the claim.

[00:09:48] Speaker 00:
And he said it was, quote, during the presentation.

[00:09:51] Speaker 00:
So during would be the duration from beginning to end.

[00:09:55] Speaker 00:
And rendering would be synonymous with presentation.

[00:09:58] Speaker 00:
That matches precisely with the definition from the specification of the patents that I just read.

[00:10:05] Speaker 00:
The specification further makes clear that things like presentation delays that might happen in the middle of a presentation are, in fact, accounted for by the presentation time parameter.

[00:10:17] Speaker 00:
For one, the patents talk about this concept of a cumulative rendition period.

[00:10:22] Speaker 00:
They say that's one way to calculate a presentation time.

[00:10:26] Speaker 00:
And what's interesting is that the patents provide that the cumulative rendition period might not just be the sum of its parts.

[00:10:32] Speaker 00:
It might not equal just the time it takes to present one presentation unit, another one, another one, another one.

[00:10:39] Speaker 00:
They say it could take into account some other factors.

[00:10:41] Speaker 00:
And that makes sense in light of Dr. Schonfeld's testimony.

[00:10:44] Speaker 00:
Dr. Schonfeld explained

[00:10:46] Speaker 00:
at appendix 1974 to 1975, that it's actually more often than not the case that when you're presenting media like this in a streaming scenario where things are coming from a server to a client, that you'll have delays.

[00:10:58] Speaker 00:
Things will usually take a little bit longer than you expect them to take.

[00:11:02] Speaker 00:
And then when we confronted Dr. Bancelot with this portion of the specification, he couldn't exclude that delays were taken into account by these other factors that are referenced in the specification.

[00:11:12] Speaker 00:
And he simply threw up his hands and said he couldn't really understand what the inventors were talking about.

[00:11:16] Speaker 00:
That's an appendix 1470 to 1472.

[00:11:21] Speaker 00:
Further to that,

[00:11:22] Speaker 00:
Varentum's interpretation of the claims would mean that data time and presentation time are always the same if timescale modification has not happened, regardless of the definition of timescale modification.

[00:11:35] Speaker 00:
The specification, again, rejects that understanding.

[00:11:38] Speaker 00:
At appendix 2 to 8, column 16, lines 44 to 50, the specification provides that in a traditional player, these two times, that's presentation time and data time, are the same.

[00:11:51] Speaker 00:
But in a player capable of variable speed playback, or simply capable of differentiating between data time and presentation time, those two values are, in general, different.

[00:12:05] Speaker 00:
Again, when we asked Dr. Boncellet to reconcile Verentim's construction with that portion of the specification, he said, it's not obvious what the inventors were thinking of at the time when they wrote this.

[00:12:16] Speaker 00:
That's an appendix 1496 to 1497.

[00:12:22] Speaker 00:
In any event, I think there's a very big gradient between what Varentum is arguing, saying scenarios where, say, there's a 20-second pause in a presentation and what Nelson actually teaches.

[00:12:37] Speaker 00:
When Nelson talks about presentation units that come late, that they might get deleted, that's in the context of an ongoing presentation.

[00:12:44] Speaker 00:
And Nelson makes absolutely clear that presentation units

[00:12:49] Speaker 00:
when they arrive late, they might arrive too late to present.

[00:12:53] Speaker 00:
But that data is still processed, because the data in those presentation units still might affect the presentation of subsequent units.

[00:13:01] Speaker 00:
Because the way modern systems work, you're not going to just have the data for a given presentation at a single time.

[00:13:07] Speaker 00:
The data interacts with each other.

[00:13:09] Speaker 00:
So data from a previous presentation unit affects subsequent presentations.

[00:13:13] Speaker 00:
And Dr. Bonsoleil and Dr. Schoenfeld agreed on that point.

[00:13:16] Speaker 00:
Dr. Bonsoleil's testimonies at appendix 1548, page 110, lines 19 to 25, where he explained or he agreed that these deleted presentations, it's not really deleted, it's almost a misnomer.

[00:13:30] Speaker 00:
They arrive late, so they're not themselves presented, but the data in them is still processed.

[00:13:34] Speaker 00:
And Dr. Schoenfeld likewise testified at appendix 1968.

[00:13:39] Speaker 00:
And you can see that also in Nelson itself.

[00:13:41] Speaker 00:
in Appendix D of Nelson, it's Appendix D of Nelson, it's Appendix 1342 in the joint appendix, where Nelson refers to either consume and present or consume and process.

[00:13:55] Speaker 00:
So deleted presentation units, they're consumed and processed rather than consumed and presented.

[00:14:00] Speaker 00:
But the data is still actually processed.

[00:14:03] Speaker 00:
So if there are some

[00:14:05] Speaker 00:
outlier scenarios where maybe you could have a system that data comes in 20 seconds after the fact.

[00:14:11] Speaker 00:
The user's long gone.

[00:14:12] Speaker 00:
It's after the presentation is done.

[00:14:14] Speaker 00:
Could that be a scenario that's not included in claims?

[00:14:16] Speaker 00:
I don't know.

[00:14:18] Speaker 00:
But Nelson is not that case.

[00:14:19] Speaker 00:
Nelson is a case where data is constantly coming in.

[00:14:22] Speaker 00:
There'd be fractions of seconds where these come slightly too late.

[00:14:26] Speaker 00:
But the data is still used.

[00:14:28] Speaker 00:
It's still an ongoing presentation.

[00:14:30] Speaker 00:
And the reference time base in Nelson still measures the presentation time.

[00:14:34] Speaker 00:
I would also note that there's really no dispute, or there is no dispute, that Nelson's calculated parameters and embedded parameters are, in fact, the data time.

[00:14:44] Speaker 00:
And in fact, it's eerily similar to what the specification

[00:14:48] Speaker 00:
In the challenge patents teach, where the specification talks about how data time can be explicit or derivable, at appendix 225, column 9, lines 33 to 39, that lines up precisely with Nelson's technique of having either a calculated or embedded parameter that lines up with the data time.

[00:15:07] Speaker 00:
If your honors have no further questions, I'm happy to return the remainder of my time.

[00:15:14] Speaker 02:
Any more questions?

[00:15:16] Speaker 02:
Thank you.

[00:15:16] Speaker 02:
Thank you.

[00:15:17] Speaker 01:
I'm sorry, I was muted.

[00:15:18] Speaker 01:
I apologize for that.

[00:15:22] Speaker 01:
So I'm going to go back to the claim itself because that's what we have to look at when we talk about this.

[00:15:29] Speaker 01:
So what I just heard from Google's counsel was presentation time is just a clock.

[00:15:36] Speaker 01:
It starts running when you start the presentation and it just keeps going.

[00:15:40] Speaker 01:
And there's no dispute that the only time that Nelson actually presents a unit

[00:15:47] Speaker 01:
is when the date and time, when the clock lines up and it says, I was expecting this unit at one minute.

[00:15:54] Speaker 01:
If it comes in, if it's there at one minute, then I will present it.

[00:15:58] Speaker 01:
If it is late, I do not present it.

[00:16:01] Speaker 01:
It does not get rendered.

[00:16:02] Speaker 01:
So I just heard a lot about there's some processing that happens and something happens to the data, but it does not get rendered.

[00:16:11] Speaker 01:
That is not disputed.

[00:16:13] Speaker 01:
So when we go to the claim itself,

[00:16:15] Speaker 01:
This is about, and also I didn't, there's nothing about maintaining that value.

[00:16:19] Speaker 01:
It's not stored anywhere.

[00:16:20] Speaker 01:
It's just, you look at the clock, look at the clock.

[00:16:25] Speaker 01:
Is it on time?

[00:16:27] Speaker 01:
Present it.

[00:16:28] Speaker 01:
If it's not on time, delete it.

[00:16:30] Speaker 01:
It might get processed in some other way, but it's not rendered.

[00:16:33] Speaker 01:
And so when we look at the claim itself, and I'm now in appendix 283, which is 903 claim one, column 27, starting at lines 19.

[00:16:45] Speaker 01:
So it's a method performed by at least one machine for rendering temporal sequence presentation data in a machine-implemented rendering system.

[00:16:54] Speaker 01:
The temporal sequence presentation data is tangibly stored in a first computer medium, and then it is comprised of maintaining the value of the presentation time parameter, tangibly stored in a second computer readable medium, and representing an amount of time elapsed during rendering of a portion.

[00:17:15] Speaker 01:
not since the beginning of time during rendering of a portion.

[00:17:20] Speaker 01:
And then when you go down to element C, now you have to maintain a second value.

[00:17:26] Speaker 01:
The second value is a data time parameter tangibly stored in a third computer read of low memory, which we didn't hear anything about either, and rendering the amount of time required to render the portion, the same portion, the amount of time

[00:17:44] Speaker 01:
required to render the same portion that was actually rendered, the antecedent basis proportion, and storing that value in a machine-readable memory.

[00:17:57] Speaker 01:
And then the final element is wherein the value of the presentation time parameter is not equal to the value of the data time parameter.

[00:18:06] Speaker 01:
So that is completely contrary to Nelson.

[00:18:10] Speaker 01:
only renders units when the data time and the presentation time are equal.

[00:18:17] Speaker 01:
And it does not render units when they are not, and it does not store these values.

[00:18:22] Speaker 01:
In fact, what we just heard was that the value of the presentation time is just a clock that keeps on ticking since the beginning of the presentation.

[00:18:31] Speaker 01:
And so it is very different from the claims of the invention and

[00:18:39] Speaker 01:
And this applies to all of the claims in the 903 family and in appeal number 17, please.

[00:18:46] Speaker 01:
60 votes.

[00:18:48] Speaker 02:
I believe we have the argument.

[00:18:50] Speaker 02:
Do you need another minute or so for the last word?

[00:18:53] Speaker 01:
The last word is time scale modification.

[00:19:00] Speaker 01:
And what I just did there is,

[00:19:02] Speaker 01:
I think Google, even though they didn't advocate it as their claim construction, and they relied on statements from prior counsel as the primary basis for their claim construction of another patent in another case involving other parties, Google has been unequivocally clear in this litigation in its unprompted invalidity contentions as to its understanding of what timescale modification means to a person of ordinary skill in the art.

[00:19:30] Speaker 01:
And that's what we're here to do.

[00:19:31] Speaker 01:
It's a technical term.

[00:19:33] Speaker 01:
It's a term that has a meaning.

[00:19:36] Speaker 01:
It's like saying you didn't put a definition of semiconductor in a patent.

[00:19:41] Speaker 01:
You wouldn't go look and see what a semiconductor is.

[00:19:44] Speaker 01:
For audio signal processing, time scale modification has a defined meaning.

[00:19:48] Speaker 01:
There is nothing other than saying, yes, the specification says you speed up and slow down.

[00:19:54] Speaker 01:
Of course it speeds up and slows down.

[00:19:56] Speaker 01:
That's part of it.

[00:19:57] Speaker 01:
But when you say timescale modification as opposed to fast forward, except for in the 433 patent and the other references that are cited in consulate, you're talking about a technique that tries to preserve things as intelligible.

[00:20:12] Speaker 01:
And that is timescale modification.

[00:20:14] Speaker 01:
We will acknowledge that when it was unprompted in its own ability to conduct tests.

[00:20:24] Speaker 02:
Thank you.

[00:20:25] Speaker 02:
Thanks to all counsel, Mr. Mori, Mr. Zalberg, and Mr. Modi.

[00:20:29] Speaker 02:
The case is taken under submission.

[00:20:31] Speaker 02:
That concludes this panel's arguments for this morning.