Trials@uspto.gov
`571-272-7822
`
`
`
`
`
`Paper 36
`Date: March 8, 2024
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`____________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________
`
`VERANCE CORP.,
`Petitioner,
`
`v.
`
`MZ AUDIO SCIENCES, LLC,
`Patent Owner.
`____________
`
`IPR2022-01544
`Patent 7,289,961 B2
`____________
`
`Record of Oral Hearing
`Held Virtually: January 25, 2024
`____________
`
`Before KARL D. EASTHOM, DAVID C. MCKONE, and
`IFTIKHAR AHMED, Administrative Patent Judges.
`
`
`
`
`
`
`
`
`
`
`571-272-7822
`
`
`
`
`
`Paper 36
`Date: March 8, 2024
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`____________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________
`
`VERANCE CORP.,
`Petitioner,
`
`v.
`
`MZ AUDIO SCIENCES, LLC,
`Patent Owner.
`____________
`
`IPR2022-01544
`Patent 7,289,961 B2
`____________
`
`Record of Oral Hearing
`Held Virtually: January 25, 2024
`____________
`
`Before KARL D. EASTHOM, DAVID C. MCKONE, and
`IFTIKHAR AHMED, Administrative Patent Judges.
`
`
`
`
`
`
`
`
`
`
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`
`APPEARANCES:
`
`ON BEHALF OF THE PETITIONER:
`
`
`Babak Tehranchi
`PERKINS COIE, LLP
`11452 El Camino Real,
`Sand Diego, CA 92130-2080
`
`
`ON BEHALF OF THE PATENT OWNER:
`
`
`Sarah E. Spires
`SKIERMONT DERBY, LLP
`1601 Elm Street
`Suite 4400
`Dallas, Texas 75201
`
`
`
`
`
`The above-entitled matter came on for hearing on Thursday, January
`25, 2024, commencing at 10:00 a.m., via video teleconference.
`
`
`
`
`
`2
`
`
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`
`
`
`P R O C E E D I N G S
`- - - - -
`JUDGE EASTHOM: Good morning. This is Judge Easthom. Judge
`Ahmed and Judge McKone are on the video hearing with us. This is
`Verance Corp. versus MZ Audio Sciences, LLC, IPR 2022-01544, U.S.
`Patent 7,289,961 B2, and we’ll refer to that as the ’961 patent. We
`appreciate everyone’s willingness to be here. You have the right to be
`heard, so if at any time you have a problem hearing anything or the lines
`drop, please contact the Board personnel that helped you set this up. When
`you’re not speaking, please mute yourself and introduce yourself when you
`begin speaking, and pause before speaking after we ask a question so the
`court reporter creates a good record. Please also briefly pause after you turn
`to another slide or location in the record so that everyone can catch up. We
`have both parties’ demonstratives, and thank you for filing those. They’re
`helpful for us to resolve this issue, although they’re not evidence, as we
`know.
`Petitioner, you have the burden of persuasion. You have 45 minutes.
`Patent Owner also has 45 minutes. You can also each reserve rebuttal time.
`So, with that Petitioner, why don’t you please introduce yourself for the
`record and let me know what time you want to reserve for rebuttal, please?
`MR. TEHRANCHI: Thank you, Judge Eastman. My name is Babak
`Tehranchi. I’m representing Petitioner Verance Corporation. I’d like to
`reserve 30 minutes, and 15 minutes for the rebuttal. So, start with 30
`minutes and then reserve 15.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`3
`
`
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`
`JUDGE EASTHOM: Thank you, Mr. Tehranchi. Okay, who do we
`have for Patent Owner?
`MS. SPIRES: This is Sarah Spires from Skiermont Derby. On behalf
`of MZ Audio Sciences, LLC.
`JUDGE EASTHOM: Welcome, Ms. Spires.
`MS. SPIRES: Thank you.
`JUDGE EASTHOM: You may begin, Mr. Tehranchi.
`MR. TEHRANCHI: Thank you, Your Honor, may I share my
`demonstratives? I know you have it, but it may be easier to just go on,
`follow it.
`JUDGE EASTHOM: It’s totally up to you. We have them, and if you
`want to share them, that’s fine.
`MR. TEHRANCHI: Okay, so let me know if you could see them.
`JUDGE EASTHOM: Yes. Great.
`MR. TEHRANCHI: Again, thank you for the opportunity to present
`this case. In our case, we have three grounds. First ground is based on the
`combination of Srinivasan, Cabot, and Kudumakis. The second ground,
`we’re adding Hobson for some dependent claims. And the third ground is
`based on Kudumakis, Tilki, and Cabot. This is on page two of the
`demonstratives.
`On page three, I have reproduced the claim language, claim 1 and
`claim 4. What I like to point out is that I think both parties agree there is no
`controversy that the actual text that is not in the red box is taught by
`Srinivasan, in the Srinivasan combination. And it basically entails getting
`the audio signal, dividing it into time frames and frequency components,
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`
`
`4
`
`
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`then taking at least two of those components and altering the face of at least
`one of those components.
`The added limitations that were added during prosecution was about
`selecting the fundamental tone and the overtone. And that is where the
`combination with Cabot and Kudumakis comes into picture. I’d also like to
`note that there are also watermark extraction claims, which more or less are
`similar to the encoding. They’re decoding the watermark from the
`composite signal.
`On page four of the demonstrative, this is just to show that
`Kudumakis basically has these steps. You divide it up into time frames, into
`frequency components, and then it selects two components. It calls them F1
`and F0. And it changes the phase of one component, and it talks about
`changing the component with a lower amplitude; so, it changes the phase of
`that with respect to a reference frequency component.
`On page five of the demonstratives, as you can see, Srinivasan talks
`about selecting a frequency range of 4.8 kHz to 6 kHz. And it says you need
`to do so because that provides better audibility results. But however,
`Srinivasan also recognizes that embedding watermarks in terms of audibility
`is not perfect; and in fact, acknowledges in that column 11, that we’ve
`highlighted that says, yes, if you embed watermarks and you are to change
`the phase of the component to be zero or 180 degrees, it could result in some
`audible artifacts. And it goes on to explain that there is an alternative
`method where it actually uses certain neighborhoods as opposed to changing
`the phase to be zero or 180. So, there is an acknowledgment that there might
`be some audibility issues, and it proposes some methodology to fix that.
`And also, we want to point out that the language in Srinivasan about
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`
`
`5
`
`
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`selecting the 4.8 to 6 kHz, we don’t believe it’s an absolute language. In
`other words, going back to figure 2, which is on page four of the
`demonstrative, there is no frequency band requirement in that flowchart, and
`neither is in the claims of Srinivasan.
`Why is that important? That’s because in our proposed combination,
`we may have to move away from that 4.8 kHz to 6 kHz. And in our
`combination, the proposal is based on Cabot, which is basically, they tested
`a setup where they changed the phase of third harmonic with respect to the
`fundamental. They were able to show that these phase changes, if you make
`it to the fundamental -- to the harmonics, to the harmonic series, they do not
`produce audible artifacts for certain phase changes. So, in this experiment,
`they showed that if you change the phase under 22 degrees or up to 15
`degrees, it was absolutely impossible to hear that phase change.
`However, if you go a little bit higher, there might be artifacts. But in
`the conclusion, Cabot says, these are really subtle. And it’s also notable that
`the way Cabot conducted the experiment, it was designed to be able to detect
`the audible artifacts. And what I mean by that is they had pristine laboratory
`conditions where they had listeners that at the beginning, they refused to
`believe that there were any changes to these signals until they actually
`showed them on the oscilloscope and showed the traces. They went through
`this training process where they were able to switch between, basically the
`original content that was not changed and then switch back to the watermark
`content. Not watermark, but the content with a base change. They went
`back and forth as many times as they wanted. And at the end, when they got
`to 22 and a half degrees, after all these experiments, about 60 percent of the
`time they were able to tell that there was an artifact.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`
`
`6
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`
`JUDGE MCKONE: This is Judge McKone. Can I ask you a question
`
`here?
`
`MR. TEHRANCHI: Absolutely.
`JUDGE MCKONE: So, if we accept your argument here that Cabot
`shows a phase shift for the third harmonic versus the fundamental is either
`subtle or inaudible. Are you arguing that Cabot suggests using the
`fundamental and the third harmonic to encode data? Or are you simply
`arguing that Cabot shows that that wouldn’t cause a problem in Srinivasan?
`MR. TEHRANCHI: We’re suggesting both, and specifically the fact
`that a person of ordinary skill in the arts, reading Cabot, would understand
`that if you affected the phase change that Srinivasan is talking about, if you,
`instead of using random frequency components that Srinivasan describes, if
`you were to use the harmonic series and specifically change the third
`harmonics phase, then that would not be audible. It would not produce any
`audible artifacts. And in fact --
`JUDGE MCKONE: But Cabot doesn’t discuss encoding data. Right?
`It’s simply an experiment to show how audible a phase shift would be,
`correct?
`MR. TEHRANCHI: Correct. That’s correct.
`JUDGE MCKONE: Okay. So, do you have any affirmative evidence
`that a skilled artisan would have chosen the fundamental and the third
`harmonic as the candidates for encoding data?
`MR. TEHRANCHI: So, I answer that in two steps. So, one is -- and I
`move to my slide, oh geez, to the Kudumakis slide, which is on page 10. So,
`Kudumakis was about watermarking. And Kudumakis talked about
`selecting the locations of fundamental and overtones for embedding
`
`
`
`7
`
`
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`watermarks. So, selecting those locations for watermarking purposes was
`known. Now --
`JUDGE MCKONE: I guess, as an initial matter, Kudumakis is the
`reference that you’re citing to show that you would encode in the
`fundamental and the third harmonic?
`MR. TEHRANCHI: It would use those positions to embed the
`watermarks. And it also stated that it’s also advantageous to use those
`positions because they vary across the content in sort of an unknown
`fashion, randomly. So, it would be harder for somebody who’s trying to
`attack and remove the watermarks to be able to locate them.
`JUDGE MCKONE: Okay. And I just want to confirm that it’s
`Kudumakis and not Cabot that you’re citing to and showing the affirmative
`evidence that the fundamental and the third harmonic are your choices for
`encoding?
`MR. TEHRANCHI: That’s correct. In addition, and again, it’s with
`combination with Srinivasan and Cabot, because Srinivasan says, pick two
`frequencies, any two frequencies, and it shows two different ways of picking
`those. And it’s more or less a random way of picking those. And Cabot
`says, well, it’s really, if you change the phase of harmonic series and the
`third overtone, it wouldn’t be audible or detectable, or it would be very
`subtle. And --
`JUDGE MCKONE: That would rebut an argument that the
`fundamental and the third harmonic would cause audible problem that
`Srinivasan wants to avoid?
`MR. TEHRANCHI: That’s correct. That’s correct.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`
`
`8
`
`
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`
`JUDGE AHMED: Mr. Tehranchi, speaking of Kudumakis, you
`talked about randomness. Doesn’t Kudumakis teach not just using those two
`frequencies, but also adding a random offset to those frequencies? Why then
`would a person of skill ignore that teaching and use exactly those
`frequencies?
`MR. TEHRANCHI: Randomness is an added benefit. And in our
`view, it’s not necessary to use that, because then you also need to then
`generate that random offset; you need to communicate that random offset to
`the detector. So, it does improve the randomness, perhaps, but it also adds
`some overhead to the embedding process. So, it is part of it. It could be part
`of it, but we’re not using that for this combination.
`JUDGE AHMED: Okay.
`JUDGE EASTHOM: And just to make clear, this is Judge Easthom,
`none of these references, or your modification doesn’t involve changing the
`phase of the fundamental. Right? You’re not doing anything to the
`fundamental, is that right?
`MR. TEHRANCHI: That’s absolutely correct.
`JUDGE EASTHOM: And then in Srinivasan, Srinivasan selects the
`two overtones based on the fundamental. Right?
`MR. TEHRANCHI: So, Srinivasan just says, pick two. And I’ll go
`back to its figure 2. It basically says two. Now, there are different ways that
`it --
`
`JUDGE EASTHOM: Well, in Srinivasan, I think in column 8, it says
`they’re going to select the highest frequency, the strongest frequency, Imax.
`Right?
`MR. TEHRANCHI: Right. And then they’re going to find some --
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`
`
`9
`
`
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`
`JUDGE EASTHOM: Right. They shift from that to select two
`overtones, right? They --
`MR. TEHRANCHI: Correct. So, they use that as one reference, and
`then they use the two --
`JUDGE EASTHOM: Generally, I think the patent says if you pick
`the strongest frequency component, that’s the fundamental, typically.
`Right?
`MR. TEHRANCHI: As the reference, correct.
`JUDGE EASTHOM: As a reference. And Srinivasan does that, and
`then they shift it up to a higher frequency from 4.8. That’s the preferred
`range. So, you’re not trying to change the fundamental. And Srinivasan is
`basing overtones off the fundamental, and so is your combination, the way I
`understand it. Am I understanding that correctly?
`MR. TEHRANCHI: That’s correct. So, we’re not changing the
`fundamental or the strongest frequency in the combination. It’s the
`overtone. And we’re very specific that based on Cabot, at least, we’re
`looking at the third harmonic, which has been proven that it doesn’t produce
`audible artifacts.
`JUDGE EASTHOM: And so, the fundamental is just, I mean, the
`third harmonic is just a special case of an overtone, is that right?
`MR. TEHRANCHI: That’s correct. So, what Cabot started, the
`premise was that it talks about in its background that there were studies done
`and the prior studies have shown that if you change the phase of harmonic
`series, and it’s not specific to the third harmonic, you won’t be able to hear
`that phase change. And then it devised a very specific test involving a third
`harmonic in that example, just to show that that is the case.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`
`
`10
`
`
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`
`JUDGE EASTHOM: And part of that is because a third harmonic is
`symmetrical with respect to the fundamental, according to what your expert
`relies on, and I think it’s Lipshitz or maybe Cox, I’m not sure.
`MR. TEHRANCHI: That’s correct, because, you know, and I think
`that’s in the slide 9 that Lipshitz actually says that because these are
`symmetrical signals, the harmonics are. So, their effects, their audible
`effects in changing the phase is much less pronounced. So, that is the
`premise. Correct.
`JUDGE EASTHOM: But as my colleague was putting, that doesn’t
`have anything to do with embedding phases -- in Lipshitz, does that deal
`with encoding and watermarking?
`MR. TEHRANCHI: It’s not about watermarking. And we’re using
`all of these, Risset and Lipshitz and Cabot to show that changing the phase
`of harmonics would not be audible. So, you put a person of ordinary skill in
`the art, when they’re presented with a scenario that they have to pick two
`frequency components and change the phase of the frequency components,
`as Srinivasan describes in its own technique, they would understand, reading
`these other references, that it’s been known for a long time, that if you
`change the phase of harmonic series, that would have less audible effects
`overall in the watermarking. And that’s obviously something that is a
`benefit. It’s something that you want to do in a watermarking system.
`JUDGE MCKONE: I want to do a quick follow up here. You’re not
`arguing that Risset and Lipshitz affirmatively suggest using the fundamental
`and the third harmonic to encode data. Rather, you’re arguing that if you
`choose the fundamental and the third harmonic, Risset and Lipshitz show
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`
`
`11
`
`
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`that it wouldn’t cause audibility problems that Srinivasan might be
`concerned about. Is that correct?
`MR. TEHRANCHI: That’s correct.
`JUDGE MCKONE: Thank you.
`MR. TEHRANCHI: That’s correct.
`JUDGE AHMED: Mr. Tehranchi, going back to Srinivasan, you
`talked about the flexibility. Srinivasan says 4.8 to 6 kHz range. You said
`that Srinivasan teaches some flexibility there. If we disagree with you that
`there is that flexibility in Srinivasan, does your combination still work?
`MR. TEHRANCHI: The combination works because Srinivasan
`relied, again, the technique that it used, it was based on more or less, I
`characterized it as being selecting the two frequencies randomly. So, when
`you select the frequencies randomly, then there might be the case that, yes,
`you should choose that 4.8 kHz to 6 kHz, because that would be a better
`position in terms of audibility. But in the combination, we’re not proposing
`to use random frequency components for effectuating the phase embedding.
`But we’re saying that a person of ordinary skill in the art would have
`understood that they would switch -- they would select the harmonic series
`and they would change the phase of one of the overtones. And in that case,
`what Cabot and Risset and all these other evidence, and our experts testify
`that the phase would be practically inaudible, even if you lowered that 4.8.
`And in addition to that, and that’s another part of this, is that changing that
`frequency range to a lower frequency range, perhaps, to enable using the
`harmonic series, would provide other benefits. So, shifting it to a lower
`frequency range would allow you to embed for telephony applications,
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`
`
`12
`
`
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`which we’ve said, it cuts it off the bandwidth for that is about 4 kHz. So,
`Srinivasan’s system won’t work there.
`Using the harmonic series, as Kudumakis explains, would basically
`allow different areas of the content to be embedded in sort of a random
`fashion, and it moves dynamically across the content, so it makes the attack
`more difficult. And also in malicious attacks, one of the basic attacks on
`audio watermarks is low-pass filtering, and most of the energy of the audio
`content and speech content is at the lower frequencies. So, by lowering that
`frequency band, changing it, you would gain a lot of robustness. You would
`prevent a watermark to be removed based on normal processing, based on
`intentional attacks on it. So, there are a bunch of benefits that are gained by
`shifting that frequency range.
`JUDGE AHMED: And going back to Kudumakis. It’s a different
`encoding technique, right? It’s the notch encoding. Why would a person of
`skill in the art look to that? And it does not do it at the exact frequencies. It
`does it off the frequencies. Why would a person look to that when
`implementing Srinivasan?
`MR. TEHRANCHI: SO, again, yes, you’re correct, Your Honor.
`Kudumakis talks about a specific type of embedding algorithm. So, it says
`you use a notch to remove the frequencies and because you -- and then you
`insert your own tones in those locations. And because of that specific way
`of embedding, then it says you don’t change the stronger frequency
`components, but you go to the neighborhood and change one of those. So, it
`is that specific embedding process that’s done that then Kudumakis uses to
`say that you embed at the edges of those frequencies. Now, in our
`combination, we’re not relying on the embedding process of Kudumakis
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`
`
`13
`
`
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`because Srinivasan already has this phase-embedding technique,
`quantization phase embedding, that says you pick two frequencies, you
`change the phase of those frequencies. So --
`JUDGE MCKONE: I do have a question on Kudumakis still. So,
`Kudumakis discusses choosing its placement of the encoding to be near, but
`not on the fundamental or a harmonic. And it seems to have a description in
`here that you’re choosing those locations because the stronger fundamentals
`that are harmonics, if you do not alter them, they will mask the changes that
`you make to the frequencies nearby. Is that correct?
`MR. TEHRANCHI: So, if you have a strong fundamental frequency
`and then that, because it is -- any strong signal in general, would mask other
`smaller signals that is around them, right. So, if you are changing the
`fundamental itself, that’s certainly the case. And in that case, they’re
`actually removing that fundamental or removing the area of that
`fundamental and then adding their own signals in there. In our combination,
`at the end of proposal is, you change the third harmonic, which in a
`harmonic series, it is the lower amplitude. So, there might be some masking
`there, but we’re not removing anything. And for that reason, you don’t need
`to go to the edges or to the neighborhoods of that third harmonic.
`JUDGE MCKONE: But doesn’t Kudumakis stand for the principle
`that if you are making modifications to the stronger signals, in his case a
`notch rather than a phase change, but making changes to the stronger
`signals, such as the fundamental and the overtones of the fundamental, those
`could cause undesirable audio or audibility problem. Is that a fair reading of
`Kudumakis?
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`
`
`14
`
`
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`
`MR. TEHRANCHI: In case, when you remove those components and
`you try to replace those components with something completely different,
`right. In our combination, we’re not removing those. We’re changing the
`phase of the third harmonic. And phase change is practically inaudible in
`most cases. And in case of harmonic series, it’s even better in terms of
`audibility.
`JUDGE MCKONE: Well, my concern here, though, is we have three
`references in the combination. Srinivasan discusses choosing its encoding
`range to be in a range where it believes, or Srinivasan believes it’s inaudible
`or barely audible. And Kudumakis chooses its encoding to be away from
`the fundamental and the overtones so that it doesn’t introduce audibility
`problems. So, we have two of the references discussing potential audibility
`problems with the encoding in the range that your combination would
`suggest. Why should we not be concerned about two of the references
`perhaps pointing in a direction other than your combination?
`MR. TEHRANCHI: So, in terms of Srinivasan, again, it uses, it
`changes --the way it selects the frequencies is more or less random. So, it’s
`not using the harmonic series to affect the phase change. So, for that reason,
`the audibility issue that it ports that you would need to put it in the higher
`end in that 4.8 to 6 kHz, that’s not there when you select the harmonic
`series. In Kudumakis, the reason that you have audibility issues is that
`because you’re removing, you’re doing cancer surgery as an analogy, you’re
`removing the signals, and then you’re adding something completely foreign
`in its place. So, the notch embedding, which was also acknowledged in,
`actually, Srinivasan, it does produce audible artifacts if you’re actually
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`
`
`15
`
`
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`removing the signal. That’s why Srinivasan, at part of it says, don’t do that
`cancer surgery, go just change the phase of those signals.
`JUDGE MCKONE: But it seems to me that what you’re arguing is,
`okay, we’ve got two references, Srinivasan and Kudumakis, that seem to
`suggest that your selected placement of the encoding could cause noise
`issues, but that Srinivasan and Kudumakis were incorrect about that; they
`misunderstood, and using Cabot to show that they misunderstood and that
`those placements would be possible without audibility problems. And what
`I’m looking for now is the affirmative reason why we would have chosen, or
`a skilled artisan would have chosen the placement that your combination
`requires.
`MR. TEHRANCHI: Okay, right. Again, so, in Srinivasan, the reason
`that Srinivasan says -- and we don’t agree that it’s absolutely, it has to be in
`that 4.8 to 6 reason that it picked those is because it did not contemplate
`using the harmonic series. So, it’s not that it says don’t put them in
`harmonic -- in those areas. Just it didn’t contemplate, it didn’t think about
`using the harmonic series. And Cabot says, well, if you change the phase of
`the harmonic series, your audibility issues go away. Basically, that’s our
`position in terms of changing the phase of the signal. And when we get to
`Kudumakis, Kudumakis did not talk about using phase quantization
`embedding in how --
`JUDGE EASTHOM: Counsel, can I interrupt you for a second? I’m
`sorry. You have about two and a half, two minutes left in your main case.
`But doesn’t Kudumakis encode at the lower frequencies that my friend was
`asking you about, my colleague?
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`
`
`16
`
`
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`
`MR. TEHRANCHI: I’m sorry, Kudumakis, could you repeat the
`question?
`JUDGE EASTHOM: Does Kudumakis’s third harmonic, that’s
`encoding something at a lower frequency, right?
`MR. TEHRANCHI: Kudumakis is not very specific about which
`frequency range that it uses.
`JUDGE EASTHOM: But the third harmonic, I thought, was, I think
`your friend and people are sort of saying the third harmonic generally falls,
`it could fall anywhere, really, but generally it might fall under 4000. I don’t
`know. If the fundamental’s 1000, the third harmonic will be 3000; if the
`fundamental’s 2000, third harmonic will be 6000. So, the third harmonic
`can vary. But Kudumakis generally says encode at the third harmonic or
`right around it. So, it’s encoding at what parties here seem to say are lower
`frequencies below what Srinivasan teaches. Right?
`MR. TEHRANCHI: Right. So, yes, if we embed it in the third
`harmonic, it could fall any -- pretty much in the upper band. So, we are
`changing the third harmonic, so, in situations that the third harmonic is in
`that 4 kHz or above, that may be the issue -- may be the case, but it could
`also fall below that as well.
`JUDGE MCKONE: All right. So, if we accept your arguments that
`encoding, that Cabot shows that encoding at the first -- or the fundamental
`and the third harmonic would be problem free, vis-a-vis Srinivasan and
`Kudumakis. Have you shown more than a skilled artisan could have made
`the combination, when what we’re looking for, the question we have to
`answer is whether a skilled artisan would have made the combination?
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`
`
`17
`
`
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`
`MR. TEHRANCHI: I think that both is the case in terms of the actual
`process of what to do it, Srinivasan already talks about that. And we’ve
`talked about that in our papers. Srinivasan has the frequency selection
`capabilities. Kudumakis talks about FFTs and spectrums, et cetera, to be
`able to locate those harmonics; and changing the phase of the components is
`not an issue. It’s within the realm of person of ordinary skill in the art. And
`Srinivasan talks about that as well.
`And in terms of would they be making that combination? And our
`answer is absolutely yes, because Srinivasan just has changed the phase of
`the quantization phase embedding. It’s just a matter of whether a person
`ordinarily skilled in the art would have thought about using the harmonic
`series to do that. And the answer is yes, because Cabot and Lipshitz and
`these other corroborating references showed that if you made that change, it
`would be inaudible, practically inaudible, in most situations.
`I think I may be out of time. Let me just, if I may, unless there are
`other questions, I just want to touch on the Tilki reference, which would be
`similar in this case. So, Tilki is also talking about phase embedding. So, it
`selects two frequency components. And it selects the phase of one
`frequency component with respect to the frequency component in the
`previous FFT bin. So, that the difference is that it says here is how you do
`the selection; this is how you would select it. It also talks about an example
`that you would embed above 2 kHz. Again, this will be, again, evidence or
`something that a person of ordinary skill in the art would see if you were
`comparing it to Srinivasan that said do it in 4 kHz. Here Tilki said you
`would do it above 2 kHz. So, the frequency range has shifted. And again,
`that’s not a big deal, because when you’re doing phase embedding to begin
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`
`
`18
`
`
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`with, it’s not very audible. And now, if you use the same combination of
`Cabot and Kudumakis to select the harmonic series to embed, those audible
`issues would completely go away, or practically go away. So, that is the
`basic combination.
`One of the issues that the Patent Owner has brought up is that the
`embedding bitrate may go down if you use the harmonic series instead of the
`preceding bin, or doing that phase modification in the Tilki combination.
`And the answer is, yes, that is correct. But there are other ways to gain that
`loss in data rate, namely use additional harmonics in that series. Just don’t
`use the third harmonic. And also, it’s known in the industry, and a person of
`ordinary skill would have known that there are tradeoffs in all of this. And
`you don’t really need a high data rate channel for watermarks. In fact, for
`copy control watermarks, you only need one bit that says copy or don’t
`copy. In other cases, you may need a few bits to show an ID. But in any
`case, there are other references, evidence that we’ve presented in our papers
`that says you don’t really need a high data rate watermark. And there’s a list
`of improvements that you would have obtained by implementing Tilki.
`Again, the robustness, because again, that selected frequency is not fixed. It
`moves within the content. Tilki’s methodology is very simple. And then it
`also provides security in terms of low-pass filtering and malicious attacks.
`So --
`JUDGE EASTHOM: Counsel, you have about 10 minutes left if you
`stop in 20 seconds, if you want to wrap up.
`MR. TEHRANCHI: I think this is where I’m going to stop. There are
`other slides about the actual implementation. How would the person of
`ordinary skill in the art would implement. These are all in our papers. And I
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`
`
`19
`
`
`
`IPR2022-01544
`Patent 7,289,961 B2
`
`will not discuss ground two which is based on Hobson analogous art unless I
`may use some of my rebuttal time if Patent Owner brings that up. But our
`papers fairly well cover that second ground.
`Thank you, Your Honors.
`JUDGE EASTHOM: Thank you, Counsel. I’m sorry, did I cut you
`
`off?
`
`MR. TEHRANCHI: No, I was going to ask you do I have about 10
`minutes for my rebuttal?
`JUDGE EASTHOM: Ten minutes, yes.
`MR. TEHRANCHI: Perfect. Thank you.
`JUDGE EASTHOM: And then, you’re welcome. And we’ll hear
`from your friend Ms. Spires, now Patent Owner.
`MS. SPIRES: Thank you, Your Honor.
`JUDGE EASTHOM: Do you want to reserve rebuttal Ms. Spires?
`MS. SPIRES: Yes, I’d like to reserve 10 minutes please.
`JU
Accessing this document will incur an additional charge of $.
After purchase, you can access this document again without charge.
Accept $ Charge
Still Working On It
This document is taking longer than usual to download. This can happen if we need to contact the court directly to obtain the document and their servers are running slowly.
Give it another minute or two to complete, and then try the refresh button.
A few More Minutes ... Still Working
It can take up to 5 minutes for us to download a document if the court servers are running slowly.
Thank you for your continued patience.
This document could not be displayed.
We could not find this document within its docket. Please go back to the docket page and check the link. If that does not work, go back to the docket and refresh it to pull the newest information.
Your account does not support viewing this document.
You need a Paid Account to view this document. Click here to change your account type.
Your account does not support viewing this document.
Set your membership
status to view this document.
With a Docket Alarm membership, you'll
get a whole lot more, including:
- Up-to-date information for this case.
- Email alerts whenever there is an update.
- Full text search for other cases.
- Get email alerts whenever a new case matches your search.
One Moment Please
The filing “” is large (MB) and is being downloaded.
Please refresh this page in a few minutes to see if the filing has been downloaded. The filing will also be emailed to you when the download completes.
Your document is on its way!
If you do not receive the document in five minutes, contact support at support@docketalarm.com.
Sealed Document
We are unable to display this document, it may be under a court ordered seal.
If you have proper credentials to access the file, you may proceed directly to the court's system using your government issued username and password.
Access Government Site
