`
`Filed on behalf of: Velos Media, LLC
`
`
`By: Barry J. Bumgardner
`Registration No. 38,397
` NELSON BUMGARDNER ALBRITTON P.C.
`
`3131 W. 7th Street, Suite 300
`
`Fort Worth, Texas 76107
`Telephone: (817) 377-3494
`Email: barry@nbafirm.com
`
`
`
`
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`______________________
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`______________________
`UNIFIED PATENTS, LLC
`Petitioner
`v.
`VELOS MEDIA, LLC
`Patent Owner
`______________________
`Case IPR2019-00757
`Patent 9,930,365
`______________________
`
`PATENT OWNER’S SUR-REPLY
`
`
`
`
`
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`TABLE OF CONTENTS
`
`SUMMARY OF THE RESPONSE .................................................................. 1
`
`
`I.
`
`II. KALKER .......................................................................................................... 2
`
`
`A. Scanning Circuit 41 does not “Segment” ......................................................... 6
`
`
`
`
`
`
`
`
`
`B. Kalker’s Decoder Doesn’t Partition Either ..................................................... 13
`
`C. Petitioner has not Presented Evidence that Novotny Applies to
`Sub-Blocks ...................................................................................................... 16
`
`D. Response to Some of Petitioner’s Gross Misstatements of Fact .................... 17
`
`E. Patent Owner’s Further Views on Footnote 6 ................................................ 18
`
`A. Unified’s Arguments Misunderstand the Law ................................................ 21
`
`
`
`
`
`
`III. PETITIONER HAS NOT CORRECTLY LISTED ALL REAL PARTIES-
`IN-INTEREST ................................................................................................ 21
`
`B. Unified Filed These IPRs at its Members’ Behest and for Their Benefit ...... 22
`
`
`IV. CONCLUSION ............................................................................................... 25
`
`
`
`
`
`Patent Owner’s Sur-Reply
`
`
`
`
`ii
`
`
`
`TABLE OF AUTHORITIES
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
` Page(s):
`
`
`Cases:
`
`Applications in Internet Time v. RPX Corp.,
`897 F.3d 1336 (Fed. Cir. 2018) ..................................................................... 22, 25
`
`
`Power Integrations, Inc. v. Semiconductor Components Industries, LLC,
`926 F.3d 1306 (Fed. Cir. 2019) ...................................................................... 22-23
`
`
`Other Authorities:
`
`Trial Practice Guide,
`77 Fed. Reg. 48,756 (Aug. 14, 2012) .................................................................. 22
`
`
`
`
`
`
`
`Patent Owner’s Sur-Reply
`
`
`
`
`iii
`
`
`
`
`
`Ex
`
`. 2001
`
`Ex
`
`. 2002
`
`Ex
`
`. 2003
`
`Ex
`
`. 2004
`
`Ex
`
`. 2005
`
`Ex
`
`. 2006
`
`Ex
`
`. 2007
`
`Ex
`
`. 2008
`
`Ex
`
`. 2009
`
`Ex
`
`. 2010
`
`Ex
`
`. 2011
`
`Ex
`
`. 2012
`
`Ex
`
`. 2013
`
`Ex
`
`. 2014
`
`Ex
`
`. 2015
`
`Ex
`
`. 2016
`
`Ex
`
`. 2017
`
`Ex
`
`. 2018
`
`Ex
`
`. 2019
`
`Ex
`
`. 2020
`
`Ex
`
`. 2021
`
`Ex
`
`. 2022
`
`Case IPR2019-00757
`Patent No. 9,930,365
`
`TABLE OF EXHIBITS
`
`Description and Short Name
`Statements made by Petitioner’s Co-Founder and COO
`concerning Patent Owner on LinkedIn
`ITU-T TELECOMMUNICATION
`STANDARDIZATION SECTOROFITU, SERIES H:
`AUDIOVISUAL AND MULTIMEDIA SYSTEMS-
`Infrastructure of audiovisual services — Coding of moving
`video, Advanced video coding for generic audiovisual
`services, Recommendation ITU-T H.264 (March 2009
`Definition of “separate” from Google Search (June 11,
`2019
`Unified Patents Press Release, April 9, 2019
`Unified Patents Press Release, January
`9, 2019
`Statements made by “VP Marketing @ Beamr”on
`Patent Owner’s [Proposed] Interrogatories to Petitioner
`Patent Owner’s [Proposed] Requests for Production to
`Petitioner
`Patent Owner’s [Proposed] Rule 30(b)(6) Notice of Oral
`Deposition of Petitioner Unified Patents Inc.
`Patent Owner’s [Proposed] Notice of Deposition of
`Shawn Ambwani
`Unified Patents’ Collaborative Deterrence Approach
`from Internet Archive Wayback Machine
`Unified Patents Press Release, March 1, 2019
`Unified Patents Press Release, October 9, 2019
`Unified Patents Press Release, November 8, 2018
`Unified Patents Press Release, December6, 2018
`Excerpt from Redacted Transcript of Deposition of Kevin
`Jakel filed in IPR2019-00194 (Ex. 1029
`Declaration of Iain Richardson (“Richardson Decl.”
`Curriculum Vitae of Iain Richardson
`Transcript of Deposition of Immanuel Freedman, Ph.D,
`Taken on October 22, 2019 (“Freedman Depo.”
`“Graph” (Exhibit 3 from Freedman Depo.
`“Graph” (Exhibit 4 from Freedman Depo.
`“Graph” (Exhibit 6 from Freedman Depo.
`
`
`
`Patent Owner’s Sur-Reply
`
`IV
`
`
`
` Case IPR2019-00757
` Patent No. 9,930,365
`9/4/17 Email from Kevin Jakel with Attachments
`9/14/17 Email from Shawn Ambwani with Attachments
`11/24/17 Email from Shawn Ambwani with Attachments
`12/3/17 Email from Kevin Jakel with Attachments
`1/18/18 Email from Shawn Ambwani with Attachments
`1/19/18 Email from Shawn Ambwani with Attachments
`2/3/18 Email from Shawn Ambwani with Attachments
`2/9/18 Email from Shawn Ambwani with Attachments
`1/2/18 Email from Shawn Ambwani with Attachments
`12/6/19 Kevin Jakel Deposition Transcript
`Final Written Decision from Unified Patents, Inc. v.
`Uniloc USA, Inc., IPR2018-00199, Paper 33 (PTAB May
`31, 2019)
`Petitioner’s Supplemental Second Voluntary
`Interrogatory Responses
`Unified Patents Member Agreement
`Unified Patents Press Release, November 8, 2018
`Unified Patents Press Release, April 9, 2019
`Unified Patents Press Release, January 9, 2019
`LinkedIn Post from Shawn Ambwani
`Email from Fred Telecky of Velos
`Mass Email Titled “Unified Files IPR Against US
`9,338,449 Owned by Velos Media LLC”
`Unified Patents Newsletter from November 2018
`Part Two of Written Interview given by Kevin Jakel to
`abovethelaw.com
`Transcript of Deposition of Kevin Jakel
`Unified Success Page
`
`Ex. 2102
`Ex. 2103
`Ex. 2109
`Ex. 2111
`Ex. 2113
`Ex. 2114
`Ex. 2122
`Ex. 2127
`Ex. 2132
`Ex. 2138
`Ex. 2150
`
`Ex. 2151
`
`Ex. 2152
`Ex. 2153
`Ex. 2154
`Ex. 2155
`Ex. 2156
`Ex. 2157
`Ex. 2158
`
`Ex. 2159
`Ex. 2160
`
`Ex. 2161
`Ex. 2162
`
`Patent Owner’s Sur-Reply
`
`
`
`
`v
`
`
`
`
`I.
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`SUMMARY OF THE RESPONSE
`Petitioner’s entire case hangs on two instances of the word “segmented” in
`
`Column 5 of Kalker. Based on these two words, Petitioner has constructed a tale
`
`about how Kalker renders obvious the Challenged Claims. But, this particular tale
`
`comes undone when the implications of Petitioner’s claim are examined. In
`
`addition, Patent Owner has brought to light positions independent of Petitioner’s
`
`“segmenting” claims that show with equal force how little Kalker has to do with the
`
`Challenged Claims. In short, Petitioner, not Patent Owner, ignores the express
`
`teachings of the reference in a futile attempt to force Kalker into the mold of the
`
`Challenged Claims.
`
`There are many examples of Petitioner twisting Kalker’s disclosure to suit its
`
`ends. For instance, Kalker describes a segmentation circuit that segments and a
`
`scanning circuit that scans. Kalker, 2:60-3:24. But, according to Petitioner, the
`
`scanning circuit also segments, even though this functionality is nowhere described.
`
`On a related topic, Kalker’s picture transform circuit is shown using the same
`
`segmentation map as the map encoding circuit. According to Petitioner, however,
`
`they use different maps. Kalker states that its picture blocks are not sub-dividable
`
`and teaches away from the use of sub-blocks. Id., 1:51-53. But, contrary to this
`
`disclosure, Petitioner somehow finds sub-blocks in Kalker. As a final example,
`
`Petitioner repeatedly states that decoders perform the inverse functions of an
`
`Patent Owner’s Sur-Reply
`
`
`
`1
`
`
`
`
`encoder, but then insists that both Kalker’s encoder and decoder perform segmenting
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`functions. These are just some of the logical inconsistencies that flow from
`
`Petitioner’s positions.
`
`The key to resolving this IPR is to simply understand Kalker. With a proper
`
`understanding, the deficiencies of Kalker as applied to the Challenged Claim are
`
`readily apparent.
`
`II. KALKER
`Patent Owner extensively discussed the technical aspects of Kalker in its
`
`Patent Owner Response (“POR”, Paper 18). It will briefly review some of the more
`
`salient aspect of Kalker as an introduction to its comments below.
`
`The bulk of Kalker’s disclosure is related to encoding a segmentation map.
`
`See Kalker, Summary of the Invention. The segmentation map is produced by
`
`segmentation circuit 3 which analyzes a “video picture or part thereof.” Id., 2:45-
`
`46. No details are provided about the actual method employed by the segmentation
`
`circuit to determine the appropriate block sizes. The output of segmentation circuit
`
`3 is a segmentation map. Id., 2:64-3:2. The map consists of a plurality of variable
`
`sized blocks. Id. Examples of segmentation maps are shown in Figs. 3, 7, and 9 of
`
`Kalker. Segmentation circuit 3 is mentioned once in the Petition and is not relied
`
`upon to show any aspect of the ’365 Patent and/or the Challenged Claims.
`
`Patent Owner’s Sur-Reply
`
`
`
`2
`
`
`
`Map-encoder circuit 4 takes as input the segmentation map output by
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`
`
`segmentation circuit 3 and encodes the contents of the segmentation map for
`
`transmission. Id., Fig. 1 (reproduced below), 3:1-7, and 19-25. Map-encoder circuit
`
`4 is comprised of scanning circuit 41, a linear combination circuit 42, a run-length
`
`coder 43, and a Huffman coder 44. Id., Fig. 2 (also reproduced below). Scanning
`
`circuit 41 “scans the segmentation map on the basis of a grid corresponding to the
`
`smallest block size” in one embodiment, and scans “on the basis of the largest block
`
`size” in another embodiment. Id., 3:23-24 and 5:31-35. The algorithm carried out
`
`by scanning circuit 41 is shown in Fig 4. Scanning circuit 41 simply reads the two-
`
`dimensional segmentation map and outputs a one-dimensional series of block size
`
`codes corresponding to the segmentation map.
`
`Patent Owner’s Sur-Reply
`
`
`
`
`
`3
`
`
`
`
`
`
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`
`
`Importantly, scanning circuit 41 doesn’t alter the segmentation map provided
`
`to it by segmentation circuit 3. Its job is to scan it and turn it in to a string of numbers.
`
`See, e.g., id. at 5:58. This functionality is confirmed by Fig. 4, which “shows a flow
`
`chart illustrating the operation of the segmentation map-encoding circuit . . .” and
`
`the description of Fig. 4. Id., 2:22-25.
`
`Indeed, the segmentation map cannot be altered by scanning circuit 41. As
`
`shown in Fig. 1, the same segmentation map that is provided to map encoder 4
`
`(which includes scanning circuit 41) is also provided to transform circuit 1.
`
`Transform circuit 1 uses the picture block sizes specified in the segmentation map
`
`to perform transforms on the picture blocks. Id. 2:49-52 (“transform circuit 1 []
`
`subjects picture blocks having a variable block size S to a picture transform. In the
`
`present example, the block size is 4*4, 8*8 or 16*16 pixels.”). On the decoder side,
`
`the segmentation map that flows out of map encoder 4 is used to perform an inverse
`
`transform on the blocks transformed by transform circuit 1. Id., Fig. 1. If there is a
`
`Patent Owner’s Sur-Reply
`
`
`
`4
`
`
`
`
`mismatch between the block sizes used by the transform and inverse transform
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`circuits, the output of the inverse transform circuit will be meaningless.
`
`The functionality of scanning circuit 41 was further confirmed by Petitioner’s
`
`expert. When questioned about Fig. 9, which is the subject of Kalker’s Column 5,
`
`Petitioner’s expert was asked to confirm that the figure below (Ex. 2019, Exhibit 4
`
`to Petitioner’s expert’s deposition) was the full segmentation map represented by
`
`Figure 9 of Kalker:
`
`
`
`Deposition of Petitioner’s Expert Dr. Freedman (Ex. 2019), 107:20-108:1. As
`
`compared to Figure 9 of Kalker, this figure makes clear that the segmentation map
`
`is a grid/matrix composed of numbers between 1 and 3, which represent block size
`
`codes in Kalker. Note that this segmentation map does not contain any data
`
`describing the picture itself (e.g., data representing pixels in the picture). This
`
`Patent Owner’s Sur-Reply
`
`
`
`5
`
`
`
`
`segmentation map simply describes how segmentation circuit 3 decided to divide a
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`picture into blocks, and in that sense is “related” to the picture.
`
`A.
`Scanning Circuit 41 does not “Segment”
`In the Petition and in Petitioner’s Reply (Paper 26), Petitioner repeatedly
`
`parrots two sentences in Column 5 of Kalker that contain the word “segmented.”
`
`The first sentence states, “[t]his block is segmented into smaller blocks and will now
`
`completely be scanned before proceeding to the next 16*16 block.” In its Reply,
`
`Petitioner continues to falsely assert that this sentence is evidence that scanning
`
`circuit 41, which is part of map-encoder circuit 4, is performing some sort of
`
`segmenting action. If Petitioner is to be believed, Kalker’s scanning circuit 41
`
`receives a segmentation map from segmentation circuit 3 and further
`
`segments/divides it. This is simply contrary to the express disclosures of Kalker.
`
`Patent Owner addressed this issue extensively in POR. POR at 34-39. In
`
`addition to this analysis, Patent Owner makes the following observations. The
`
`phrase “this block” in the sentence from Kalker quoted above refers to scanning a
`
`region of the segmentation map “on the basis of the largest block size”, which, for a
`
`block size code of 3, is 16x16 (in comparison, Fig. 3 shows how scanning is carried
`
`out on the basis of the smallest block size). See id. at 15-16 (discussing Kalker’s
`
`two uses of the term “block”). The phrase “this block” does not refer to any picture
`
`block having a block size code. Instead, it is simply repeating what was discussed
`
`Patent Owner’s Sur-Reply
`
`
`
`6
`
`
`
`
`in the prior paragraph of Kalker (Kalker, 5:31-35) – that the segmentation map in
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`this embodiment is scanned in 16x16 regions (the largest block size) instead of 4x4
`
`regions (the smallest block size) as shown in Fig 3.
`
`This same paragraph gives context to the portion of Kalker that Petitioner
`
`heavily relies upon. This paragraph states:
`
`In a further embodiment of the scanning circuit 41 (see FIG. 2), the
`segmentation map is scanned on the basis of the largest block size. If a
`block comprises smaller blocks, it is scanned on the basis of the next
`smaller block size. This is an iterative process.
`
`A key sentence from above states that if a “block” contains smaller blocks, it is
`
`scanned on the basis of the next smaller block size. Again, this use of “block” does
`
`not refer to a block having a size code, rather, it refers to the region of the
`
`segmentation map scanned by scanning circuit 41. Conclusive proof of this
`
`statement is found in the next paragraph, where Kalker discusses scanning four
`
`16x16 blocks. The top left 16x16 “block” (referred to as a region of the
`
`segmentation map in POR) contains a single picture block represented by block size
`
`code 3. See Kalker, Fig. 9 and Ex. 4 to Petitioner’s expert’s depo. Kalker discloses
`
`the following about the processing of this region of the segmentation map:
`
`First, the top left 16*16 block is analyzed. As this block is not further
`divided into smaller blocks, the block size code S=3 is generated.
`
`Patent Owner’s Sur-Reply
`
`
`
`7
`
`
`
`
`Kalker, 5:38-40. How does Kalker’s scanning circuit 41 know that this block is “not
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`further divided into smaller blocks”? It knows this because the picture block in this
`
`scanning area is represented by block size codes of 3. Since Kalker’s picture blocks
`
`are not divisible into smaller blocks, once Kalker’s encounters a picture block of size
`
`3 (i.e., a 16x16 picture block represented by S=3) in a segmentation map, it can skip
`
`that picture block and move to the next 16x16 region of the segmentation map for
`
`processing. The same concept is repeated again when Kalker encounters a picture
`
`block size of 2 – “More particularly, the top left 8*8 block is now analyzed. As it is
`
`not further divided, the block size code S=2 is generated.” Id., 5:37-40. As
`
`discussed above, once scanning circuit 41 encounters a picture block having a block
`
`size code 2 (i.e., an 8x8 picture block represented by S=2), it can skip that picture
`
`block and move to the next picture block, because it knows that picture blocks are
`
`not divisible.
`
`The second use of the word “segmented” in Column 5 of Kalker mirrors the
`
`first. It simply states that the 8x8 region being analyzed at the moment is divided
`
`into picture blocks that are smaller than 8x8 (i.e., they are picture blocks having a
`
`block size code of 1). Thus, this area is “segmented” into four, 4x4 blocks, instead
`
`of one 8x8 block.
`
`To be clear, this portion of Column 5 is describing the operation of Kalker’s
`
`scanning circuit 41. So, Petitioner’s reading of “segmented” as a verb requires
`
`Patent Owner’s Sur-Reply
`
`
`
`8
`
`
`
`
`scanning circuit 41 to be carrying out this “segmenting.” If Petitioner were to be
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`believed, this 8x8 area would have been “something else” prior to being divided by
`
`scanning circuit 41. But, Kalker states that Fig. 9 is a “segmentation map” which is
`
`output by segmentation circuit 3. This map is the initial input to scanning circuit 41.
`
`Petitioner’s argument requires that Fig. 9 not be the segmentation map provided to
`
`scanning circuit 41 from segmentation circuit 3, as it shows certain blocks in their
`
`alleged “post-segmented” state (i.e., the state they are in after “the block” was
`
`segmented). According to Petitioner’s logic, this segmentation map is some altered
`
`version of what was initially received by scanning circuit 41 from segmentation
`
`circuit 3. But, this sort of alteration is never described or mentioned in Kalker. This
`
`is but one more example of the inconsistencies that flow from Petitioner’s view of
`
`Kalker’s scanning circuit 41 performing some function other than scanning as
`
`described in Fig. 4.
`
`In fact, such behavior is required for Kalker’s encoder and decoder to
`
`function. Kalker’s transform circuit 1 uses the “locally optimal block size codes”
`
`generated by the segmentation circuit to apply the correspondingly-sized forward
`
`transforms when transforming the actual picture data. See id., Fig. 1 (showing signal
`
`Xin (e.g., a video picture) and the output of segmentation circuit 3 being used as input
`
`by transform circuit 2) and 2:45-46. Thus, using the segmentation map shown in
`
`Fig. 9 for example, transform circuit 1 may perform a Discrete Cosine Transform on
`
`Patent Owner’s Sur-Reply
`
`
`
`9
`
`
`
`
`ten separate blocks of video data (i.e., the three 16x16, three 8x8, and four 4x4
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`picture blocks shown in Fig. 9, applying 16x16, 8x8, or 4x4 transform sizes,
`
`respectively), resulting in ten separate DCT blocks.
`
`It is axiomatic that any inverse transform circuit operating on data
`
`representing such a series of DCT blocks would have to perform inverse transforms
`
`on the basis of the same number of blocks and block sizes used to transform the data.
`
`In the example above, if a DCT circuit transformed blocks of video data into ten
`
`separate DCT blocks, the inverse circuit would have to know the number of DCT
`
`blocks and the size of the transform to be applied to each (16x16, 8x8, or 4x4 in the
`
`example above) in order to generate a proper output signal. See Pet. at 6 (“The
`
`[decoded] coefficients are then subject to inverse transformation operations to
`
`reverse the DCT process”) and 52 (“[Kalker’s] receiving station applies the relevant
`
`block size S to the inverse transform circuit for decoding encoded data”).
`
`Petitioner’s argument leads to Kalker’s receiving station applying the wrong
`
`block size S to the inverse transform circuit and therefore cannot be correct.
`
`Consider the following example with respect to Kalker’s Fig. 9. Assume that the
`
`segmentation map output by segmentation circuit 3 only consisted of four 16x16
`
`(i.e., S=3) picture blocks, but that scanning circuit 41 further “segmented” this
`
`segmentation map to what is shown in Fig. 9, as alleged by Petitioner (consisting of
`
`the three 16x16, three 8x8, and four 4x4 picture blocks, represented by block size
`
`Patent Owner’s Sur-Reply
`
`
`
`10
`
`
`
`
`codes 3, 2, 2, 2, 1, 1, 1, 1, 2, 3, 3 – Id., 5:50). This is exactly what Petitioner is
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`assuming when it cites to the word “segmented” as being used by as a verb in
`
`Column 5. See Freedman Depo., 108:19-110:6. This would result in a flow of
`
`information as shown below:
`
`Patent Owner’s Sur-Reply
`
`
`
`
`
`11
`
`
`
`
`Kalker, Fig 1 (annotated to show the illogical result of scanning circuit 41 (part of
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`the map encoder) “segmenting” according to Petitioner’s allegations). When the
`
`decoded segmentation map and transform coefficients reach inverse transform
`
`circuit 8, there is a “collision”, for lack of a better word. The inverse transform
`
`circuit will try to perform ten inverse transforms of varying sizes on coefficient data
`
`corresponding to four equal-size pixel blocks, resulting in non-sensical data being
`
`output as Xout.
`
`The sole purpose of encoding and transmitting the segmentation map is to
`
`inform inverse transform circuit 8 of the correct sequence of block sizes used to
`
`perform the original transforms, a concept to which Petitioner agrees wholeheartedly
`
`- “Fundamentally, the entire point of sending block size codes from the encoder to
`
`the decoder is to enable the decoder to segment an image in the exact same way it
`
`was segmented at the encoding side so that the blocks of image data can be
`
`reconstructed and decoded in exactly the inverse of the way they were encoded.”
`
`Reply at 7.
`
`In sum, if one credits Petitioner’s argument about the word “segmented”, one
`
`is immediately faced with a number of issues that are simply irreconcilable with
`
`Kalker’s disclosure. The disclosure of Kalker would have to be completely rewritten
`
`to fit this narrative. The Board should decline the invitation to do so.
`
`Patent Owner’s Sur-Reply
`
`
`
`12
`
`
`
`
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`B. Kalker’s Decoder Doesn’t Partition Either
`The discussion above is somewhat irrelevant to the Challenged Claims in that
`
`Kalker’s Figs. 3, 4, and 9, as well as Petitioner’s oft relied upon portion of Column
`
`5 all relate to Kalker’s encoder whereas the Challenged Claims are all directed to a
`
`decoder (or the functionality thereof). Thus, the time spent by Petitioner twisting
`
`Kalker’s encoder into something it is not is really immaterial because it is Kalker’s
`
`decoder that is material to the Challenged Claims. It is also interesting to note that
`
`nearly all of Petitioner’s citations in the sections of its Reply dealing with Kalker’s
`
`decoder are to Kalker’s encoder. See Reply at 9-11. There are no citations to
`
`Kalker’s Figure 6, for instance, that describe the operation of Kalker’s decoder or to
`
`the corresponding discussion thereof in Kalker’s specification.
`
`As an initial matter, Petitioner never really attempts to reconcile the obvious
`
`logical disconnect regarding how Kalker’s encoder and decoder can both partition
`
`data, given Petitioner’s view that video encoders generally perform inverse steps as
`
`video decoders. Petitioner spends one sentence addressing this issue in its Reply,
`
`but it never addresses the fundamental issue of how partitioning big blocks into small
`
`ones on the encoder side is also carried out on the decoder side. Is the result even
`
`smaller blocks? Petitioner has no answer.
`
`Likewise, Petitioner never addresses the unambiguous statement in Kalker
`
`that “only block-size codes are transmitted [from the encoder to the decoder] for
`
`Patent Owner’s Sur-Reply
`
`
`
`13
`
`
`
`
`blocks which are not divided into smaller blocks” other than to say it is only found
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`in the Summary of the Invention section of Kalker. This one sentence, which is
`
`characterizing “the invention” in Kalker, dooms Petitioner’s positions by itself. If
`
`the picture blocks represented by the block size codes transmitted to Kalker’s
`
`decoder are indivisible (i.e., not capable of being divided into smaller blocks), what
`
`blocks remain for Kalker’s decoder to partition (regardless of what is happening in
`
`Kalker’s encoder)? The answer is none – there are no blocks available to Kalker’s
`
`decoder which are capable of being divided.
`
`In further support of Patent Owner’s position, Petitioner inadvertently slips-
`
`up in its analysis of Kalker’s decoder in the Reply. Petitioner states:
`
`When the syntax element S=3 is received at the decoder, the inverse is
`true, i.e., the current block is identified as a 16x16 block and the
`decoder can immediately proceed to the next block without additional
`scanning because the decoder is aware that S=3 is the syntax element
`representing the largest block size of 16x16. The receipt of an S=3
`allows the decoder to determine (or recognize) that the current block is
`equal to the maximum size block.
`
`Contrary to PO’s suggestion, not all blocks in Kalker are the same size,
`i.e., the decoder must determine whether the size of the current block is
`the maximum so that it can skip forward, or if not the maximum size,
`the decoder must continue analyzing the sub-blocks.
`
`Patent Owner’s Sur-Reply
`
`
`
`14
`
`
`
`
`Reply at 9-10.1 Petitioner slips-up when it states that Kalker’s decoder determines
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`whether the size of the current block is the maximum so that it can skip forward, or
`
`if not the maximum size, the decoder continues analyzing the “sub-blocks.” Here,
`
`Petitioner is admitting that if Kalker’s decoder encounters a block size represented
`
`by a block size code 3 (i.e., the maximum block size), the decoder skips forward
`
`because it understands that there are no sub-blocks associated with this block. This
`
`is one of the rare areas where Petitioner and Patent Owner agree when it comes to
`
`Kalker. Just like Kalker’s encoder can advance to a certain extent when
`
`encountering picture blocks having a block size of 3 or 2, Kalker’s decoder can do
`
`the same. See POR at 44-49 (illustrating how Kalker’s decoder writes sixteen 3’s to
`
`a segmentation map when encountering a single block size code of 3 in the stream
`
`of block size codes it receives from Kalker’s encoder), and see also Kalker, steps 64
`
`and 65 in Kalker’s Fig. 6 (illustrating a flowchart of Kalker’s decoding process
`
`which skips over non-zero locations in a segmentation map (i.e., locations in a
`
`
`1 Patent Owner never stated that Kalker’s blocks are all the same size – it said that
`
`Kalker contemplates variable sized blocks, but that these blocks are independent
`
`from one another and not sub-blocks of another block (a distinction that seems lost
`
`on Petitioner). See, e.g., POR at 18.
`
`Patent Owner’s Sur-Reply
`
`
`
`15
`
`
`
`
`segmentation map that have been previously assigned a block-code value) before
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`extracting another block size code (step 61)).
`
`Compare this to the language of the Challenged Claims, which states that
`
`when a current block has a starting size equal to the maximum block size, it starts
`
`the partitioning process:
`
`determining that a current block of a plurality of blocks of the
`sequence of pictures has a starting size equal to the maximum size using
`the second syntax element;
`
`partitioning the current block to obtain a plurality of sub-blocks
`for the current block . . . .
`
`’365 Patent (Ex. 1001), Claim 1. Stated another way, when Kalker’s decoder
`
`encounters a maximum size block, it skips to the next block, whereas the Challenged
`
`Claim requires partitioning the current block when it is determined to be the
`
`maximum block size. In short, Kalker’s decoder’s treatment of maximum size
`
`picture blocks, as described by Petitioner, is opposite of that required by the
`
`Challenged Claims and therefore doesn’t disclose or otherwise suggest this aspect
`
`of the Challenged Claims.
`
`C.
`
`Petitioner has not Presented Evidence that Novotny Applies to Sub-
`Blocks
`Petitioner’s application of Novotny to the Challenged Claims in the Petition
`
`and its expert’s declaration never mention sub-blocks or 4x4 blocks. Thus,
`
`Patent Owner’s Sur-Reply
`
`
`
`16
`
`
`
`
`Petitioner’s citation to paragraph 65 of its expert’s declaration in the Reply to
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`support its (current) position that “Novotny’s Intra4x4 encoding mode instructs the
`
`decoder that a 4x4 region, which is a sub-block of a larger 16x16 macroblock, is
`
`encoded in intra-prediction mode” is improper for the simple reason that it doesn’t
`
`support Petitioner’s statement. See Reply at 11-12. As discussed in POR, Novotny’s
`
`MB (macroblock) type “specifies how a macroblock . . . is partitioned (or
`
`segmented) and/or encoded.” POR at 56 (quoting Novotny, ¶50). Paragraph 50 of
`
`Novotny doesn’t mention the applicability of its MB types to sub-blocks and neither
`
`do the Petition or Petitioner’s expert’s declaration. Whatever is left in the Reply is
`
`simply attorney argument that cannot take the place of evidence, especially when it
`
`is contradicted by the portion of Novotny relied upon by Petitioner. Given that the
`
`relevant portions of the Challenged Claims require the disclosure of a third syntax
`
`element applicable to the coding of sub-blocks (i.e., not macroblocks), the evidence
`
`supplied by Petitioner falls short and doesn’t disclose this third syntax element that
`
`specifies the coding of sub-blocks.
`
`D. Response to Some of Petitioner’s Gross Misstatements of Fact
`Petitioner makes several misstatements of fact in its Reply that Patent Owner
`
`would like to correct. Petitioner starts out by saying that “PO admits the term
`
`‘segmenting’ is synonymous with the term ‘partitioning,’ and that both terms refer
`
`to the division of larger blocks into smaller sub-blocks.” Reply at 1. The portion of
`
`Patent Owner’s Sur-Reply
`
`
`
`17
`
`
`
`
`this sentence referring to sub-blocks is false. Petitioner never equates “segmenting”
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`with the creation of sub-blocks. Petitioner’s next statement is likewise objectively
`
`false – “Kalker repeatedly refers to segmenting blocks into sub-blocks throughout
`
`its disclosure.” Id. As mentioned before, Kalker never uses the term “sub-blocks”
`
`once (or any other term that would be understood as referring to sub-blocks).
`
`Petitioner also tries to spin a deposition quotation from Patent Owner’s expert that
`
`he agrees with Kalker’s segmenting blocks into sub-blocks. Id. at 5-6. Simply
`
`reading the quote found below Petitioner’s false statement in its Reply shows that
`
`Patent Owner’s expert made no such statement. Id. at 6.
`
`E.
`Patent Owner’s Further Views on Footnote 6
`The topic of redefining the size of blocks associated with a particular block
`
`size code comes up again in the Reply. As an initial matter, it is helpful to review
`
`the evidence on both sides of this issue. When asked about the written support in
`
`Kalker that best describes the alleged capability in Kalker to transmit information
`
`from Kalker’s encoder to its decoder relating to the change of the size of a block
`
`associated with a block size code 3, for example, Petitioner’s expert cited to the last
`
`step of Claim 1 of Kalker as the “only” support in Kalker about such transmissions.
`
`Freedman Depo., 92:12-97:15. On the other hand, Patent Owner’s expert, Dr. Iain
`
`Richardson, whose seminal book, H.264 and MPEG-4 Video Compression (2003),
`
`is cited by Petitioner in this Petition (Ex. 1011) and most of the other Petitions it has
`
`Patent Owner’s Sur-Reply
`
`
`
`18
`
`
`
`
`filed against Patent Owner’s patents, explains in great detail why Petitioner’s claim
`
` Case IPR2019-00757
` Patent No. 9,930,365
`
`is wrong. S