`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`
`
`
`
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`
`
`
`
`
`SLING TV L.L.C., SLING MEDIA, L.L.C.,
`DISH NETWORK L.L.C., DISH TECHNOLOGIES L.L.C.
`Petitioners
`
`v.
`
`REALTIME ADAPTIVE STREAMING LLC,
`Patent Owner.
`____________________
`
`Case IPR2018-01342
`Patent 8,934,535 B2
`____________________
`
`
`
`
`PATENT OWNER’S SUR-REPLY
`
`
`
`
`
`
`

`

`
`
`
`
`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`
`TABLE OF CONTENTS
`
`
`I.
`
`Claim Construction ........................................................................................ 2
`A.
`“Access Profile” .................................................................................. 2
`1. Realtime’s proposal comes directly from the ’535 patent and is consistent
`with the claims and specification. ................................................................... 2
`2. Sling’s proposal is inconsistent with the specification. ................................... 3
`“Asymmetric Compressors” / “Asymmetric Data Compression” ...... 5
`B.
`New Arguments Not Permitted in Reply ....................................................... 5
`II.
`III. Ground 1: Dvir Does Not Anticipate Claims 1, 2, 9, 10, and 14 .................. 7
`The Petition fails to make a prima facie showing of
`A.
`anticipation. ......................................................................................... 7
`Sling fails to show that Dvir’s “sample” is a “data block.” ................ 9
`Sling fails to show that Dvir discloses “determining a parameter
`or attribute of at least a Portion of a data block” ............................... 12
`Sling fails to show that Dvir discloses an “access profile” under
`either Realtime’s proposed construction or the Board’s
`construction in IPR2018-01169. ....................................................... 14
`Sling has waived any argument that Dvir discloses
`“asymmetric” compression under the correct construction. ............. 15
`IV. Ground 2: Dvir Does Not Render Obvious Claims 1, 2, 9, 10, 14 .............. 17
`Sling adds a plethora of new obviousness theories found
`A.
`nowhere in the Petition that must be disregarded. ............................ 17
`Sling’s also waived any argument that “asymmetric”
`compression would be obvious under the correct construction. ....... 19
`V. Ground 3: Dvir and Ishii Do Not Render Obvious Claims 3–6, 8, 11,
`12. ................................................................................................................ 20
`VI. The Petition Is Time-Barred Under § 315(b). ............................................. 20
`VII. Conclusion ................................................................................................... 21
`
`B.
`C.
`
`D.
`
`E.
`
`B.
`
`
`
`i
`
`

`

`
`
`
`
`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`
`PATENT OWNER’S EXHIBIT LIST
`
`2002
`2003
`2004
`2005
`
`2006
`2007
`2008
`
`2009
`2010
`2011
`2012
`
`Exhibit No.1 Description
`2001
`Amended Complaint of June 6, 2017, in E.D. Tex. Case No. 17-
`cv-84
`Stipulated Motion in D. Colorado Case No. 17-cv-2097
`Proof of Service of Amended Complaint re: DISH Network LLC
`Proof of Service of Amended Complaint re: Sling TV LLC
`Proof of Service of Amended Complaint re: EchoStar
`Technologies LLC
`Proof of Service of Amended Complaint re: Sling Media LLC
`Defendants' Supplemental Rule 7.1 Disclosure Statement
`Defendants’ Unopposed Application for Extension of Time to
`Answer Complaint
`Defendants’ Invalidity Contentions
`Expert Declaration of Kenneth A. Zeger
`Transcript of Deposition of Scott Acton on May 10, 2019
`Realtime Adaptive Streaming v. Sling, et al., Civil Action No.
`1:17-CV-02097-RBJ, Dkt. 135-1 (D. Colo. Nov. 7, 2018), Expert
`Declaration of Alan Bovik
`Realtime Adaptive Streaming v. Sling, et al., Civil Action No.
`1:17-CV-02097-RBJ, Dkt. 151 (D. Colo. Jan. 11, 2019),
`Markman Order
`U.S. Patent App. Pub. US 2002/0144271 A1 for Appl. No.
`09/197,441 (“Behagen”)
`Transcript of Deposition of Scott Acton on May 10, 2019 in
`IPR2018-01331 on U.S. Patent No. 8,867,610.
`RFC 2435, RTP Payload Format for JPEG-compressed Video,
`October 1998
`
`2013
`
`2014
`
`2015
`
`2016
`
`
`
`
`
`
`
`1 Exhibits 2001–2009 were submitted with Patent Owner’s Preliminary Response.
`Exhibits 2010–2016 are submitted with Patent Owner’s Response.
`
`
`
`i
`
`

`

`
`
`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`The Petition fails to make a prima facie case of unpatentability. After seeing
`
`
`
`the Patent Owner Response, Petitioner Sling shifts gears and offers entirely new
`
`theories and arguments that are found nowhere in the Petition. Sling also offers a
`
`new, 61-page and 108-paragraph supplemental declaration from Dr. Acton as
`
`alleged support for its new positions. Both the declaration and Sling’s new theories
`
`and arguments should be disregarded.
`
`For example, the POR shows that Dvir does not teach or suggest the
`
`limitations “access profile” and “asymmetric” compression. Now, in reply, Sling
`
`offers new theories for how those limitations are satisfied under various
`
`constructions it could have but did not propose. Also, there no reason those theories
`
`could not have been included, at least in the alternative, in the Petition.
`
`Likewise, the POR shows that Dvir does not disclose various limitations, such
`
`as “data block,” “determining a parameter or attribute,” and “access profile.” Now,
`
`Sling does a wholesale rewrite of its anticipation theory. In reply for the first time,
`
`it argues in the alternative that each of these limitations is “obvious.” But none of
`
`these obviousness theories in the Petition. Thus, Sling essentially gives a new
`
`obviousness ground missing from the Petition
`
`Because the Petition does not present a prima facie case of obviousness,
`
`because new theories in reply are improper, and because all of Sling’s arguments
`
`fail regardless, the Board should confirm the patentability of the challenged claims.
`
`
`
`1
`
`

`

`
`
`I.
`
`Claim Construction
`
`A.
`
` “Access Profile”
`
`
`
`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`
`1.
`
`Realtime’s proposal comes directly from the ’535 patent
`and is consistent with the claims and specification.
`Realtime’s proposed construction “information that enables the controller to
`
`select a suitable compression algorithm that provides a desired balance between
`
`execution speed (rate of compression) and efficiency (compression ratio)” is correct.
`
`Realtime’s proposal comes verbatim from the ’535 patent, is consistent with the
`
`specification, and is supported by the claims. See POR at 13–16.
`
`Sling’s criticisms lack merit. First, Realtime’s proposal is not “aspirational”
`
`(Reply at 1)—it is simply the patents’ description what access profiles are. Even if
`
`the construction had a minimal degree of subjectivity (according to Sling), that does
`
`not mean it is indefinite or incorrect. See Nautilus, Inc. v. Biosig Instruments, Inc.,
`
`572 U.S. 898, 901 (2014) (“some modicum of uncertainty” in claims ensures “the
`
`appropriate incentives for innovation”). Sling’s assertion that Realtime’s proposal
`
`“is not informative as to what an ‘access profile is’” (Reply at 1) is question-begging.
`
`Both sides agree that “access profile” does not have a plain and ordinary meaning.
`
`Thus, it is appropriate to look to the intrinsic evidence to inform that meaning.
`
`Second, there is no inconsistency between Realtime’s construction and the
`
`claims or specification—and Sling identifies none. See Reply at 1–2. Sling appears
`
`
`
`2
`
`

`

`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`to argue that any construction is wrong unless it includes several terms repeated
`
`
`
`
`
`elsewhere in the claim. See id. This is meritless. Further, Sling’s argument that “the
`
`claims elsewhere explicitly recite criteria for selection” is wrong. Id. at 2. The quoted
`
`claim language of “based upon the determined parameter or attribute” is about
`
`selection of the access profile, not selection of the compressor.
`
`Third, Realtime’s construction is consistent with every example and
`
`embodiment in the ’535 patent. For instance, in column 12, each of the three access
`
`profiles is information for selecting a suitable compression algorithm. See POR 14–
`
`15. Those suitable algorithms are: (1) asymmetrical with slow compress and fast
`
`decompress; (2) asymmetrical with fast compress and slow decompress; and (3)
`
`symmetrical. Sling points to no embodiment that is excluded. The citation to the
`
`’535 patent at 4:28–44 is inapposite. It is merely a statement about lossless versus
`
`lossy compression and not an embodiment.
`
`2.
`Sling’s proposal is inconsistent with the specification.
`Sling’s proposed construction is “information that enables a controller to
`
`determine a compression routine that is associated with a data type of the data to be
`
`compressed.” Sling’s proposal must be rejected because is not the plain and ordinary
`
`meaning of “access profile” and is inconsistent with the specification. Sling argues
`
`that clear and unmistakable disclaimer is not required because “access profile” has
`
`no ordinary meaning. See Reply at 3. But that is an argument in favor of Realtime’s
`
`
`
`3
`
`

`

`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`proposal, not Sling’s. Realtime’s proposal comes directly from the ’535 patent (at
`
`
`
`
`
`8:8–12), whereas the specification never defines access profile using data type.
`
`Sling’s proposal is also inconsistent with the claims and specification. Of
`
`thirty claims of the ’535 patent, none recite “data type.” Further, Sling’s reliance on
`
`the preferred embodiment undermines its argument. In that embodiment, the
`
`specification describes “data profiles”—not “access profiles”—as enabling the
`
`controller to select a compression algorithm based on data type:
`
`In a preferred embodiment, the data profiles 15 comprise information
`
`regarding predetermined access profiles of different data sets, which
`
`enables the controller 11 to select a suitable compression algorithm
`
`based on the data type. For instance, the data profiles may comprise a
`
`map that associates different data types (based on, e.g., a file extension)
`
`with preferred one(s) of the compression algorithms 13.
`
`’535 patent at (emphasis added) at 11:31–38.
`
`
`
`In this embodiment, data profiles are associated with data types, whereas
`
`access profiles are associated with “data sets.” Indeed, the specification describes
`
`data profiles as a “map that associates different data types (based on e.g., a file
`
`extension)” to compression algorithms. This language is similar to Sling’s proposal.
`
`
`
`That access profiles do not require data types is further confirmed by the tables
`
`in columns 11 and 12. The column 11 table shows three exemplary access profiles
`
`
`
`4
`
`

`

`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`without mentioning data type. But the column 12 table shows three exemplary data
`
`
`
`
`
`profiles and adds a column about data types. This suggests that data profiles, not
`
`access profiles, are more closely related to data type.
`
`B.
`
`“Asymmetric Compressors” / “Asymmetric Data Compression”
`
`Realtime’s proposed construction is “compressors / data compression” “in
`
`which the execution time for the compression and decompression routines differ
`
`significantly.” This is the correct—both parties’ experts have agreed that this is the
`
`express definition provided by the patent. See POR at 11–12. Sling offers no real
`
`response. That Realtime’s construction includes the word “significantly” does not
`
`make it indefinite or incorrect. The Federal Circuit has repeatedly found many
`
`relative terms and words of degree not indefinite. See, e.g., Apple, Inc. v. Samsung
`
`Elecs. Co., 786 F.3d 983, 1002 (Fed. Cir. 2015) (“substantially centered” not
`
`indefinite); One-E-Way, Inc. v. ITC, 859 F.3d 1059, 1063 (Fed. Cir. 2017) (“virtually
`
`free of interference” not indefinite).
`
`II. New Arguments Not Permitted in Reply
`Sling’s reply makes a plethora of new arguments and theories that are entirely
`
`missing from the Petition. This is improper under PTAB and Federal Circuit
`
`precedent. The scope of this IPR trial limited to the grounds in the Petition, in which
`
`Sling was required to identify “with particularity . . . the grounds on which the
`
`challenge to each claim is based, and the evidence that supports the grounds for the
`
`
`
`5
`
`

`

`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`challenge to each claim.” § 312(a)(3). Though “introduction of new evidence in the
`
`
`
`
`
`course of the trial is to be expected,” the “shifting of arguments is not.” Pfizer Inc.
`
`v. Chugai Pharm., IPR2017-01357, Paper 56, at 19 (Nov. 28, 2018) (declining to
`
`consider petitioner’s “subsequent arguments” on Reply which “amount[ed] to an
`
`entirely new theory of anticipation absent from the petition”); see also Wasica v.
`
`Con’t Auto Sys., 853 F.3d 1272, 1286 (Fed. Cir. 2017).
`
`These requirements are important. “[U]nlike district court litigation—where
`
`parties have greater freedom to revise and develop their arguments over time and in
`
`response to newly discovered material—the expedited nature of IPRs bring with it
`
`an obligation for petitioners to make their case in their petition to institute.”
`
`Intelligent Bio-Sys. v. Illumina Cambridge, 821 F.3d 1359, 1369 (Fed. Cir. 2016).
`
`Petitioners are thus forbidden from providing “a new mapping of the claims to the
`
`prior art in light of the gaps that Patent Owner pointed out in its Response.” Arista
`
`Networks, Inc. v. Cisco Sys. Inc., IPR2016-00308, Paper 42, at 14 (May 24, 2017).
`
`Thus, petitioners may not remedy fatal deficiencies in their Petition with new
`
`arguments in reply. See Wasica, 853 F.3d at 1286 (“Rather than explaining how its
`
`original petition was correct, Continental’s subsequent arguments amount to an
`
`entirely new theory of prima facie obviousness absent from the petition. Shifting
`
`arguments in this fashion is foreclosed by statute, our precedent, and Board
`
`guidelines.”). Likewise, it is “improper to present in reply new evidence (including
`
`
`
`6
`
`

`

`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`new expert testimony) that could have been presented in a prior filing.” PTAB
`
`
`
`
`
`Revised Tr. Practice Guide at 15 (August 2018 Update). A reply “that raises a new
`
`issue or belatedly presents evidence may not be considered.” Id. (emphasis added).2
`
`III. Ground 1: Dvir Does Not Anticipate Claims 1, 2, 9, 10, and 14
`A. The Petition fails to make a prima facie showing of anticipation.
`
`Sling concedes that Dvir is directed to multiple embodiments, and that neither
`
`Petitioner nor its expert, Dr. Acton, provided any discussion, analysis, or mention of
`
`embodiments with the original Petition. Thus, the Petition fails to make a prima
`
`showing of anticipation and Ground 1 should be rejected.
`
`For the first time in reply, Dr. Acton introduces nine pages of new opinions
`
`concerning Dvir’s embodiments. See Acton Supp. Decl. (Ex. 1031) at ¶¶ 7–18. This
`
`is improper. All of this could and should have been included with the Petition
`
`because they were required to show anticipation. These paragraphs and Sling’s
`
`arguments relying on them (Reply at 8–11) must be disregarded.
`
`Law of Anticipation
`
`Realtime’s statement of anticipation is correct. To anticipate, a reference must
`
`disclose “all of the limitations arranged or combined in the same way as recited in
`
`
`2 See also Acceleration Bay, LLC v. Activision Blizzard Inc., 908 F.3d 765, 775 (Fed.
`
`Cir. 2018); Argentum Pharm. LLC v. Research Corp. Techs., Inc., IPR2016-00204,
`
`Pap. 85, 5-6 (Mar. 22, 2017); 77 Fed. Reg. 48,756, 48,767.
`
`
`
`7
`
`

`

`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`the claim.” Therasense, Inc. v. Becton, Dickinson and Co., 593 F.3d 1325, 1332
`
`
`
`
`
`(Fed. Cir. 2010). Blue Calypso does not diminish this black letter rule. It merely
`
`noted that a reference may anticipate if a POSITA “would at once envisage the
`
`claimed arrangement or combination.” Reply at 7. But knowing that the claim
`
`elements “could have been arranged” is not enough. Therasense, 593 F.3d at 1332.
`
`The claimed arrangement must be immediately apparent to a POSITA.
`
`Here, Sling falls far short of that prima facie showing. Realtime shows how
`
`Sling and Dr. Acton rely on at least two different Dvir embodiments but provides
`
`zero discussion of those embodiments. See POR 27–28. Thus, a POSITA would not
`
`understand, based on the evidence and arguments presented that Dvir anticipates any
`
`claims. POR at 28 ; Zeger Decl. (Ex. 2010) ¶¶ 100–101.
`
`Sling’s argument that Realtime failed to consider whether a POSITA would
`
`“at once envisage the claimed arrangement” (Reply at 10, n. 4) fails for two reasons.
`
`First, Realtime did consider that. It explained that a POSITA reading Dvir would
`
`readily understand that not all the teachings are meant for the same system and that
`
`there is no one system for which all of Dvir’s teachings apply. POR at 28; Zeger
`
`Decl ¶¶ 100–101. Second, Sling’s argument again attempts to turn the burdens on
`
`its head. It was Sling’s burden offer support for its proposition. It never did. Now it
`
`tries to fault Realtime based on an argument it never made.
`
`
`
`8
`
`

`

`
`
`
`
`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`
`B.
`
`Sling fails to show that Dvir’s “sample” is a “data block.”
`
`Sling reply confirms that it is identifying a “sample” of “groups of rasters” in
`
`Dvir as the claimed “data block.” See Reply at 11. But Sling fails to show that Dvir’s
`
`“sample” meets the agreed construction of “data block”—i.e., “a single unit of data,
`
`which may range in size from individual bits through complete files or collection of
`
`multiple files.” Instead, Sling makes two arguments in reply but neither adequately
`
`address the disputed portion of the construction: “single unit of data.”
`
`First, Sling argues that Realtime “rewrites” the construction to “further limit
`
`‘data block to proximately-located data having ‘common characteristics or
`
`functionality.” Reply at 11–12. Realtime proposes no such “rewrite.” Realtime’s
`
`position is simply that Sling has not shown that Dvir’s “sample” meets the claim
`
`construction—i.e., Sling has not shown that Dvir’s “sample” is “a single unit of
`
`data.” Realtime gave many reasons and explanations for why this. See POR at 29–
`
`31; Zeger Decl. ¶¶ 102–09.
`
`Sling’s Reply does nothing to show otherwise. Instead, the Reply argues that
`
`“data block” can be “unbounded in size,” and need not be “contiguous or proximate
`
`to another.” Reply at 12. Even if this is true, there must be some evidence or
`
`explanation for why Dvir’s “sample” constitutes “a single unit of data.” Sling’s
`
`assertion that “video data files” or “collection of multiple files” may not be
`
`
`
`9
`
`

`

`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`“contiguous” (Reply at 12) does not support its arguments because Dvir’s “sample”
`
`
`
`
`
`is not a “video data file” or a “collection of multiple files.”
`
`But as Realtime and its expert have shown, Dvir’s “sample” comprises
`
`“groups of rasters” which may be “formed from various parts of the input data, such
`
`as the beginning, middle, and end.” Zeger Decl. ¶ 107. Indeed, “‘sampling’ means
`
`that the entire input data is not analyzed.” Id. “Rather, only a ‘sample’ or (likely
`
`relatively small) portion of the input data are analyzed.” Id. For these reasons, Sling
`
`has not shown that Dvir’s sample constitutes a single unit of data. Sling does not
`
`meaningfully dispute any of these points in its Reply.
`
`Second, Sling argues that Dvir’s “sample” is a “data block” purportedly
`
`because other things (e.g., “a DVD movie” or a “frame” or “adjacent frames”) may
`
`be a “data block.” Reply at 12–13. But whether or not those other things are “data
`
`block” does not mean that Dvir’s “sample” is one. For example, while “a DVD
`
`movie” may be a “single unit of data”—e.g., because it is typically packaged as a
`
`single unit that can be selected and viewed as a single unit—that does not show that
`
`Dvir’s “sample” is also a “single unit of data.” And while “a frame” or “adjacent
`
`frames” may be a “single unit of data,” that does not show that Dvir’s “sample” is
`
`also a “single unit of data.” For example, while an input data (e.g., a DVD movie,
`
`frame, or adjacent frame) may be a “single unit of data,” Dvir’s “sample” may be
`
`
`
`10
`
`

`

`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`“formed from various parts of the input data, such as the beginning, middle, and
`
`
`
`
`
`end.” Zeger Decl. ¶ 107.
`
`Sling’s argument is therefore a non-sequitur. That other things, meaningfully
`
`different from Dvir’s samples, may or may not be the claimed “data block” does not
`
`show that Dvir’s samples is one. Importantly, none of Sling’s other examples, such
`
`as a frame or file, are at all similar Dvir’s sample. And they are all different in ways
`
`that show Dvir’s sample is not “a single unit of data.”
`
`Sling also attempts to introduce a new obviousness theory for “data block.”
`
`As discussed below, this amounts to a new theory—and indeed new ground—
`
`entirely missing from the Petition. This must be disregarded. But that Sling would
`
`try to argue obviousness and fail to show it further confirms that Dvir’s sample is
`
`not the claimed “data block.”
`
`Sling argues: “It would at least be obvious” that Dvir’s “sample” would
`
`comprise “related” data, purportedly because the sampling would come “from a
`
`single frame or related frames.” Reply at 13–14. This fails for the same reason
`
`discussed above. Just because “a single frame or related frames” might be a “data
`
`block” does not show that Dvir’s “sample” from that purported “data block” is also
`
`a “data block” / “a single unit.” Indeed, Sling provides no explanation how Dvir
`
`performs “sampling”—and it is unrebutted that sampling can be “formed from
`
`various parts of the input data, such as the beginning, middle, and end.” Zeger Decl.
`
`
`
`11
`
`

`

`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`¶ 107. It was Sling’s burden to show why Dvir’s “sample” is a single unit of data.
`
`
`
`
`
`Sling’s argument that some other thing (e.g., “a single frame or related frames”) is a
`
`single unit does meet its burden. Sling’s obviousness contention does not even make
`
`sense. Dvir’s “sample” is not a single unit of data. And Sling offers no explanation
`
`about why or how a POSITA modify it to be a single unit of data.
`
`C.
`
`Sling fails to show that Dvir discloses “determining a parameter
`or attribute of at least a Portion of a data block”
`
`Sling’s argument in reply regarding “parameter or attribute” confirms that it
`
`has abandoned any potential argument that a single 8x8 block in Dvir meets the “data
`
`block” element. See Reply at 14. Indeed, in the Patent Owner’s Response, Realtime
`
`explained why a single 8x8 block in Dvir cannot meet the “data block” element
`
`because the “parameter or attribute” Sling identified is not that of a single 8x8 block.
`
`Sling has no response and has now abandoned any assertion that Dvir’s 8x8 block
`
`can be the claimed “data block.”
`
`
`
`Instead, Petitioner identifies entirely new argument—that “data block” could
`
`be “a frame, adjacent frames, or entire video.” Reply at 14–15. Neither Petitioner
`
`nor its expert has previously disclosed the argument that “a frame, adjacent frames,
`
`or entire video” meets the “data block” element. See, e.g., Acton Dep. (Ex. 2011) at
`
`27–33. This attempt to change theories for the first time in reply, with zero support
`
`in the Petition, must be rejected. Sling’s argument must be disregarded.
`
`
`
`12
`
`

`

`
`
`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`Further, that Sling feels it necessary to shift positions only reveals the
`
`
`
`
`
`weakness of its invalidity arguments. The Petition only identifies “sample” as
`
`meeting the claimed “data block” and Sling doubles down on it in the reply while
`
`discussing that claim element. See Reply at 11. But also in Reply—and only while
`
`discussing the “parameter or attribute” claim element—it asserts that “a frame,
`
`adjacent frames, or entire video” is a data block. This only confirms that Sling has
`
`not shown that “data block” and “parameter or attribute” limitations are met.
`
`
`
`Indeed, Sling’s careless assertions and unpredictable shifting only confirms it
`
`has not conducted a proper invalidity analysis. The different limitations of claim 1
`
`of the ’535 patent have common terms that take antecedent basis from each other.
`
`For example, claim 1[a] requires “determining a parameter or attribute of at least a
`
`portion of a data block . . ..” Claim 1[b] requires “selecting an access profile from
`
`among a plurality of access profiles based upon the determined parameter or
`
`attribute.” And claim 1[c] requires “compressing the at least the portion of the data
`
`block. . . ” The terms “parameter or attribute” and “data block” take antecedent basis
`
`from earlier limitations and repeat throughout the claim.
`
`Sling never clearly identifies what these terms are, much less show how they
`
`are satisfied by the same thing for each limitation. Sling’s anticipation theory
`
`therefore fails for all claim limitations of the ’535 patent.
`
`
`
`13
`
`

`

`
`
`
`
`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`
`D.
`
`Sling fails to show that Dvir discloses an “access profile” under
`either Realtime’s proposed construction or the Board’s
`construction in IPR2018-01169.
`
`Sling offered no argument in its Petition that Dvir discloses an “access profile”
`
`under Realtime’s construction and should not be allowed to make a new argument
`
`in reply. Dr. Acton’s new opinions under this construction are also improper. See
`
`Acton Supp. Decl. ¶¶ 38–49. Both Dr. Acton’s new opinions, and Sling’s arguments
`
`relying on them, should be disregarded. See Reply at 15–17.
`
`Regardless, Dvir does not disclose access profiles under Realtime’s
`
`construction. As Dvir (and Sling’s arguments about Dvir) make clear, its disclosures
`
`are all about data type. Realtime’s proposed construction does not include “data
`
`type” and is about selecting a compression algorithm based on balancing
`
`compression speed and efficiency. Dvir’s description of “compression profiles” is
`
`silent as to these concepts. See POR at 35–36; Zeger Decl. ¶¶ 117–120.
`
`Sling’s argument about “motion estimation” (Reply at 15–16), to the extent
`
`even understandable, is unsupported and incorrect. Sling appears to argue that the
`
`selection of any algorithm would meet Realtime’s construction because any
`
`algorithm inherently balances compression speed and efficiency. Not true. That an
`
`algorithm may be used for a particular data type, such as MPEG, does not mean that
`
`the algorithm, or the particular settings of the algorithm, are selected to balance
`
`compression speed and efficiency. It only means it is compatible.
`
`
`
`14
`
`

`

`
`
`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`Nor does Dvir disclose selecting any settings of a compression algorithm,
`
`
`
`such as increasing the motion estimation settings of MPEG compression, based on
`
`the data. Sling relies solely on Dvir’s selection based on data type. This alone, with
`
`no other selection consideration or criteria, does not satisfy Realtime’s construction.
`
`Sling also concedes that Dvir does not disclose “access profile” under the
`
`Board’s construction in IPR2018-01169. See Reply at 17. Thus, if this construction
`
`is adopted, Sling’s entire anticipation argument (Ground 1) must be rejected.
`
`E.
`
`Sling has waived any argument that Dvir discloses “asymmetric”
`compression under the correct construction.
`
`Sling waived any argument Dvir discloses “asymmetric” compression under
`
`Realtime’s construction. Sling deliberately decided to omit this construction from
`
`the Petition. Sling’s only opinions and arguments were premised on its incorrect
`
`construction that omits “significantly” or “execution time.” See POR at 37–39. It
`
`should not be rewarded for that strategic and unreasonable decision. Sling is not
`
`permitted to add new opinions and arguments under a different construction for the
`
`first time in reply. Thus, Dr. Acton’s new opinions and Sling’s new arguments
`
`should be disregarded. See Acton Decl. ¶¶ 58–63; Reply at 18–19.
`
`Sling’s only excuse that it “never had an opportunity to consider” Realtime’s
`
`construction (Reply at 19) is demonstrably false. That construction comes from the
`
`patent itself, which gives an express definition of “asymmetric compression.” See
`
`
`
`15
`
`

`

`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`’535 patent at 9:63–67. The patent states: “An asymmetrical data compression
`
`
`
`
`
`algorithm is referred to herein . . .” and therefore presents a clear case of
`
`lexicography. Sling knew this and should be held to its decision to (a) present a
`
`broader construction and (b) only offer a theory under its broader construction. At
`
`the minimum, it could and should have presented alternate opinions under the
`
`express definition provided by the patent.
`
`Sling was always plainly aware that Realtime’s construction was correct
`
`before it filed the Petition. The Petition cites three earlier-filed IPRs from other
`
`petitions on the ’535 patent: IPR2018-01169, IPR2018-01170, and IPR2018-00833.
`
`See Pet. at 4. In those IPRs, the petitioners proposed the same lexicographical
`
`definition as Realtime’s proposal.
`
`Further, Sling submitted a district court expert declaration from Dr. Bovik
`
`dated November 7, 2018 (Ex. 2012). This was months before Realtime proposed the
`
`construction in this IPR. And Dr. Bovik, Sling’s own expert agreed that Realtime’s
`
`proposal is the “express definition” provided by the ’535 patent:
`
` [T]he applicant of the [’535 patent] chose to expressly define the “symmetry”
`or “asymmetry” of an algorithm in terms of run-time. The patent defined an
`asymmetrical compression algorithm as “one in which the execution time for
`the compression and decompression routines differ significantly.” I agree with
`the positions recently taken by Dr. Zeger and Realtime in this litigation that
`the [’535] patent[] provide an express definition of this term.
`
`Ex. 2012 (Declaration of Dr. Bovik) ¶ 32.
`
`
`
`16
`
`

`

`
`
`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`These facts confirm that Sling could and should have proposed and offered
`
`
`
`opinions under Realtime’s proposal in the original Petition. Instead, Sling decided
`
`to roll the dice and gain an advantage by proposing and relying on a construction
`
`that was incorrect and unreasonable. Now, it must be precluded from offering any
`
`opinion or argument under Realtime’s construction.
`
`IV. Ground 2: Dvir Does Not Render Obvious Claims 1, 2, 9, 10, 14
`In arguing that these claims are obvious over Dvir, the Petition focuses solely
`
`on “asymmetric” compression. Pet. at 40–42. The Petition and Dr. Acton do not
`
`provide a complete obviousness analysis, and instead appear to rely on obviousness
`
`to plug in a hole for the “asymmetric” compression limitation. Id. Thus, Ground 2
`
`depends largely on Ground 1 and fails for the same or similar reasons.
`
`A.
`
`Sling adds a plethora of new obviousness theories found nowhere
`in the Petition that must be disregarded.
`
`Now for the first time in Reply, Sling expands its Ground 2 obviousness
`
`theory beyond all recognition. By grouping Grounds 1 and 2 together, it attempts to
`
`add a plethora of new opinions that were never included in the Petition. For example,
`
`Sling now argues that “data block” would be obvious (Reply at 13–14); that
`
`“determining a parameter or attribute” would be obvious (id. at 14–15); and that
`
`“access profiles” would be obvious (16–18). These new theories are reflected in the
`
`table of contents of Sling’s Reply:
`
`
`
`17
`
`

`

`
`
`
`
`IPR2018-01342 (’535 Patent)
`PATENT OWNER’S SUR-REPLY
`
`
`
`
`
`
`Of course, none of these obviousness theories were preserved in the Petition.
`
`Ground 2 in the Petition was about two pages and only presented an alternative
`
`theory that the “asymmetric” compression limitation would be obvious. See Pet. at
`
`40–42. These new obviousness theories must be disregarded in their entirety. See,
`
`e.g., Pfizer v. Chugai Pharmaceutical, IPR2017-01357, Paper 56 at 19 (PTAB Nov.
`
`28, 2018) (petitioner’s new theory was an “impermissible shift of its anticipation
`
`theory” because rather than “explaining how its original petition was correct,” the
`
`argument amounted to “a new theory . . . ab