`571-272-7822
`
`
`
`
`
`
`
`Paper No. 8
`Entered: October 25, 2017
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`____________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________
`
`APPLE INC.,
`Petitioner,
`
`v.
`
`SAINT LAWRENCE COMMUNICATIONS, LLC,
`Patent Owner.
`____________
`
`Case IPR2017-01244
`Patent 6,807,524 B1
`____________
`
`
`
`Before DANIEL N. FISHMAN, ROBERT J. WEINSCHENK, and
`MICHELLE N. ANKENBRAND, Administrative Patent Judges.
`
`FISHMAN, Administrative Patent Judge.
`
`
`
`
`DECISION
`Denying Institution of Inter Partes Review
`37 C.F.R. § 42.108
`
`
`
`
`
`
`IPR2017‐01244
`
`Patent 6,807,524 B1
`
`
`
`
`
`INTRODUCTION
`I.
`Apple, Inc. (“Petitioner”) filed a Petition (Paper 1, “Pet.”) requesting
`inter partes review of claims 1–21 and 29–42 (hereinafter the “challenged
`claims”) of U.S. Patent No. 6,807,524 B1 (Ex. 1001, “the ’524 patent”)
`pursuant to 35 U.S.C. §§ 311–319. Saint Lawrence Communications, LLC.
`(“Patent Owner”) filed a Patent Owner Preliminary Response (Paper 7,
`“Prelim. Resp.”). We have authority to determine whether to institute a trial
`under 35 U.S.C. § 314 and 37 C.F.R. § 42.4(a). An inter partes review may
`be instituted only if “the information presented in the petition . . . and any
`response . . . shows that there is a reasonable likelihood that the petitioner
`would prevail with respect to at least 1 of the claims challenged in the
`petition.” 35 U.S.C. § 314(a).
`We are not persuaded there is a reasonable likelihood that Petitioner
`would prevail in showing that the challenged claims are unpatentable.
`Pursuant to 35 U.S.C. § 314, we deny institution of an inter partes review as
`to all challenged claims of the ’524 patent.
`A.
`Real Parties in Interest and Related Matters
`Petitioner identifies Apple Inc. as the real party in interest. Pet. 1.
`Patent Owner identifies itself (Saint Lawrence Communications LLC) as the
`owner of the entire interest in the ’524 patent. Paper 3, 1.
`Both Petitioner and Patent Owner identify litigation matters relating to
`the ’524 patent in the U.S. District Court for the Eastern District of Texas
`captioned as: Saint Lawrence Communications LLV v. ZTE Corp. et al.,
`Case No. 2:15-cv-349-JRG; Saint Lawrence Communications LLC v.
`Motorola Mobility LLC, Case No. 2:15-cv-351-JRG; and Saint Lawrence
`Communications LLC v. Apple Inc, et al., Case No. 2:16-cv-082-JRG. Pet.
`
`2
`
`
`
`IPR2017‐01244
`
`
`
`Patent 6,807,524 B1
`
`
`5–6; Paper 3, 2. Petitioner also identifies six other litigations related to the
`’524 patent, all of which have been terminated. Pet. 5.
`B.
`The ’524 Patent
`According to the ’524 patent, digital encoding of speech/audio is
`widely applicable to numerous applications including audio/video
`teleconferencing, multimedia, and wireless applications. Ex. 1001, 1:19–23.
`Speech encoding (or any audio encoding) converts an audio signal (e.g.,
`speech) into a digital bitstream that can be transmitted to a receiver with a
`decoder, or stored for later retrieval by a device with a decoder, to reproduce
`the encoded audio signal. Id. at 1:33–40. For speech applications, early
`techniques utilized a narrow band of speech signals encoding only audio
`signals ranging between 200–3400 Hz (so-called “narrowband” encoding).
`Id. at 1:24–26. Some techniques utilized wideband encoding to provide
`better quality of speech reproduction—encoding signals ranging from about
`50 through about 7000 Hz. Id. at 1:26–30. In digital encoding, the speech
`signal is periodically sampled to generate a digitized value and the encoder
`is applied to the sequence of digitized values to reduce the number of bits
`required to represent each digitized sample value while maintaining good
`quality in the encoded sounds. Id. at 1:32–38.
`According to the ’524 patent, one widely accepted encoding technique
`for providing a good balance between the bit rate and the resulting quality is
`so-called Code Excited Linear Predictor (“CELP”) encoding. Id. at 1:41–43.
`The ’524 patent summarizes CELP encoding as follows:
`[T]he sampled speech signal is processed in successive blocks of
`L samples usually called frames where L is some predetermined
`number (corresponding to 10-30 ms of speech). In CELP, a
`linear prediction (LP) synthesis filter
`is computed and
`transmitted every frame. The L-sample frame is then divided
`
`3
`
`
`
`
`
`IPR2017‐01244
`
`Patent 6,807,524 B1
`
`
`
`into smaller blocks called subframes of size N samples, where
`L=kN and k is the number of subframes in a frame (N usually
`corresponds to 4-10 ms of speech). An excitation signal is
`determined in each subframe, which usually consists of two
`components: one from the past excitation (also called pitch
`contribution or adaptive codebook) and the other from an
`innovative codebook (also called fixed codebook). This
`excitation signal is transmitted and used at the decoder as the
`input of the LP synthesis filter in order to obtain the synthesized
`speech.
`An innovative codebook in the CELP context, is an
`indexed set of N-sample-long sequences which will be referred
`to as N-dimensional codevectors. Each codebook sequence is
`indexed by an integer k ranging from 1 to M where M represents
`the size of the codebook often expressed as a number of bits b,
`where M=2b.
`To synthesize speech according to the CELP technique,
`each block of N samples is synthesized by filtering an
`appropriate codevector from a codebook through time varying
`filters modelling the spectral characteristics of the speech signal.
`At the encoder end, the synthesis output is computed for all, or a
`subset, of the codevectors from the codebook (codebook search).
`The retained codevector is the one producing the synthesis output
`closest to the original speech signal according to a perceptually
`weighted distortion measure. This perceptual weighting is
`performed using a so-called perceptual weighting filter, which is
`usually derived from the LP synthesis filter.
`Id. at 1:44–2:8.
`According to the ’524 patent, CELP encoding has been widely
`adopted for encoding telephone band (narrowband) sound signals (i.e.,
`ranging between 200 and 3400 Hz). Id. at 2:9–14. In such applications, the
`speech signal is typically sampled at a bit rate of about 8000
`samples/second. Id. at 2:13–14. By contrast, wideband speech encoding
`applications typically sample the speech signal at a higher bit rate of about
`
`4
`
`
`
`IPR2017‐01244
`
`
`
`Patent 6,807,524 B1
`
`
`16000 samples/second to further enhance quality of the encoded speech. Id.
`at 2:14–16. However, according to the ’524 patent, problems arise when
`applying CELP techniques for wideband signal encoding. Id. at 2:17–20. In
`particular, the frequency range of signals to be encoded typically has higher
`energy levels in the lower range of frequencies as compared to the higher
`range of frequencies (a property often referred to as “spectral tilt”) that is
`exacerbated by wider dynamic range of wideband signals to be encoded. Id.
`at 2:24–27. The ’524 patent discloses that a perceptual weighting filter of
`the CELP encoder is modified to adapt to wideband signals and preemphasis
`filters may be utilized to boost the energy of the higher range of frequencies.
`Id. at 2:27–34. However, the ’524 patent also discloses that such
`modifications to the perceptual weighting filter are inefficient for encoding
`wideband signals. Id. at 2:49–57.
`The ’524 patent purports to resolve these problems with a particular
`arrangement of filters in a perceptual weighting device (i.e., an encoder) for
`digitizing wideband audio signals (e.g., speech). Id. at 2:66–3:21. Figure 1,
`reproduced below, is a block diagram of an exemplary CELP-type wideband
`encoding device according to the ’524 patent. Id. at 6:48–50.
`
`
`
`5
`
`
`
`IPR2017‐01244
`
`
`
`Patent 6,807,524 B1
`
`
`Figure 1 of the ’524 patent shows sampled input speech signal 114
`comprising successive frames wherein each frame comprises a number of
`sub-frames. Id. at 6:51–59. Frames and sub-frames are represented as
`parameter values encoding the sampled speech signal. Id. at 6:60–7:37.
`Sampled input speech signal 114 is down-sampled by module 101 to
`increase encoding efficiency by down-sampling the input signal by a 4:5
`ratio (i.e., reducing the number of sampled frames). Id. at 7:45–58. The
`down-sampled speech signal is applied to high-pass filter 102 to eliminate
`unwanted low frequency noise (e.g., below 50 Hz). Id. at 7:59–63. The
`output of high-pass filter 102 is applied to preemphasis filter 103 to enhance
`the higher frequency range of the speech signal. Id. at 7:66–8:18. The
`output of preemphasis filter 103 is applied to LP analysis quantization and
`interpolation calculator 104 and also applied to perceptual weighting filter
`105. Id. at 8:19–9:31. Linear prediction analysis of the preemphasized
`speech signal is performed in module 104. Id. at 8:19–58. Perceptual
`weighting filter 105 applies weighting to the preemphasized speech signal to
`enable further processing to determine the optimal encoding for each frame
`to minimize any error between an encoded value to be later synthesized as a
`speech signal and the input speech signal. Id. 8:59–9:31. In an exemplary
`embodiment, the ’524 patent discloses its perceptual weighting filter is new
`in that it uses a “fixed denominator.” Id. at 9:37–41. The ultimate
`parameters representing the optimal encoding of the input sampled speech
`signal are applied to multiplexor 112 for transmission to a receiver for
`decoding or for storage and later retrieval to decode the encoded speech.
`
`6
`
`
`
`
`
`IPR2017‐01244
`
`Patent 6,807,524 B1
`
`
`
`Illustrative Claim
`C.
`Claims 1, 8, 15, 29, and 36 are the challenged independent claims of
`the ’524 patent.1 Independent claim 1, reproduced below, is exemplary of
`the challenged claims:
`1. A perceptual weighting device for producing a
`perceptually weighted signal in response to a wideband
`speech signal in order to reduce a difference between the
`wideband speech signal and a subsequently synthesized
`wideband speech signal, said perceptual weighting device
`comprising:
`a) a signal preemphasis filter responsive to the
`wideband speech signal for enhancing a high frequency
`content of the wideband speech signal to thereby produce
`a preemphasised signal;
`b) a synthesis filter calculator responsive to said
`preemphasised signal for producing synthesis filter
`coefficients; and
`c) a perceptual weighting filter, responsive to said
`preemphasised signal and said synthesis filter coefficients,
`for filtering said preemphasised signal in relation to said
`synthesis filter coefficients to thereby produce said
`perceptually weighted signal, said perceptual weighting
`filter having a transfer function with fixed denominator
`
`
`1 Patent Owner refers to claims 1, 8, 15, 29, and 36 as “independent claims.”
`See generally PO Resp. Claim 1, unquestionably an independent claim,
`recites a perceptual weighting device comprising a combination of elements.
`Claim 15 recites an encoder that comprises the perceptual weighting device
`of claim 1. Claims 29 and 36 recite a cellular transceiver and a cellular
`network, respectively, that each comprise an encoder of claim 15. We take
`no position as to whether claims 15, 29, and 36 are properly phrased as
`independent claims or whether they are simply further dependent claims
`dependent from claim 1. The distinction is irrelevant to our analysis and our
`Decision below.
`
`7
`
`
`
`IPR2017‐01244
`
`Patent 6,807,524 B1
`
`
`
`
`
`whereby weighting of said wideband speech signal in a
`formant region is substantially decoupled from a spectral
`tilt of said wideband speech signal.
`Alleged Grounds of Unpatentability
`D.
`The Petition sets forth the following asserted grounds of
`unpatentability:
`References
`Salami2 and Kroon3
`Salami, Kroon, and
`Makamura4
`Salami, Kroon, Lim5, and
`Admitted Prior Art (“APA”)6
`Salami, Kroon, Lim, APA,
`and Makamura
`
`The Petition also relies on the Declaration of Dr. Jordan Cohen
`(Ex. 1003) as support for the various contentions. Patent Owner relies on
`the Declaration of Dr. Oded Gottesman (Ex. 2004) in support of its
`contentions.
`
`Basis
`103(a)
`103(a)
`
`103(a)
`
`Challenged Claims
`1, 8, 15, 29, and 36
`2, 3, 9, 10, 16, 17, 30, 31, 37,
`and 38
`6, 13, 20, 34, and 41
`
`103(a)
`
`4, 5, 7, 11, 12, 14, 18, 19, 21,
`32, 33, 35, 39, 40, and 42
`
`
`
`
`2 R. Salami et al., Real-Time Implementation Of A 9.6 Kbit/S ACELP
`Wideband Speech Coder, Globecom’92 – IEEE Global Telecommunications
`Conference (1992). Ex. 1008 (“Salami”).
`3 P. Kroon, Regular-Pulse Excitation—A Novel Approach To Effective And
`Efficient Multipulse Coding Of Speech, IEEE Transactions on Acoustics,
`Speech, & Signal Processing (1986). Ex. 1005 (“Kroon”).
`4 U.S. Patent No. 5,295,224. Ex. 1021 (“Makamura”).
`5 J. S. Lim et al., Enhancement & Bandwidth Compression Of Noisy Speech,
`Proceedings of the IEEE, Vol. 67, No. 12 (1979). Ex. 1014 (“Lim”).
`6 Petitioner identifies certain claimed features as admitted prior art based on
`citations within the ’524 patent. Pet. 57.
`
`8
`
`
`
`IPR2017‐01244
`
`Patent 6,807,524 B1
`
`
`
`
`
`II. ANALYSIS
`A. General Principles of Law
`A patent claim is unpatentable under 35 U.S.C. § 103(a) if the
`differences between the claimed subject matter and the prior art are “such
`that the subject matter[,] as a whole[,] would have been obvious at the time
`the invention was made to a person having ordinary skill in the art to which
`said subject matter pertains.” KSR Int’l Co. v. Teleflex Inc., 550 U.S. 398,
`406 (2007). The question of obviousness is resolved on the basis of
`underlying factual determinations, including: (1) the scope and content of
`the prior art; (2) any differences between the claimed subject matter and the
`prior art; (3) the level of skill in the art; and (4) objective evidence of
`nonobviousness, i.e., secondary considerations. Graham v. John Deere Co.,
`383 U.S. 1, 17–18 (1966).
`Claim Construction
`B.
`As a step in our analysis for determining whether to institute a review,
`we determine the meaning of the claims for purposes of this Decision. In an
`inter partes review, a claim in an unexpired patent shall be given its broadest
`reasonable construction in light of the specification of the patent in which it
`appears. 37 C.F.R. § 42.100(b); see also Cuozzo Speed Techs., LLC v. Lee,
`136 S. Ct. 2131, 2142–46 (2016) (upholding the use of the broadest
`reasonable interpretation standard).
`Under the broadest reasonable construction standard, claim terms are
`generally given their ordinary and customary meaning, as would be
`understood by one of ordinary skill in the art in the context of the entire
`disclosure. In re Translogic Tech., Inc., 504 F.3d 1249, 1257 (Fed. Cir.
`2007). “[A] claim construction analysis must begin and remain centered on
`
`9
`
`
`
`IPR2017‐01244
`
`
`
`Patent 6,807,524 B1
`
`
`the claim language itself.” Innova/Pure Water, Inc. v. Safari Water
`Filtration Sys., Inc., 381 F.3d 1111, 1116 (Fed. Cir. 2004). “Though
`understanding the claim language may be aided by the explanations
`contained in the written description, it is important not to import into a claim
`limitations that are not a part of the claim.” SuperGuide Corp. v. DirecTV
`Enters., Inc., 358 F.3d 870, 875 (Fed. Cir. 2004). Only terms that are in
`controversy need to be construed and only to the extent necessary to resolve
`the controversy. See Wellman, Inc. v. Eastman Chem. Co., 642 F.3d 1355,
`1361 (Fed. Cir. 2011); Vivid Techs., Inc. v. Am. Sci. & Eng’g, Inc., 200 F.3d
`795, 803 (Fed. Cir. 1999).
`Petitioner proposes interpretations of a number of terms in the ’524
`patent. Pet. 15–19. Petitioner points to a Markman Order in a litigation
`related to the ’524 patent in the U.S. District Court for the Eastern District of
`Texas in support of many of Petitioner’s proposed interpretations. Id. (citing
`Ex. 1019).
`Patent Owner “opposes the Petitioner’s proposed claim
`constructions.” PO Resp. 13. However, with the exception of the term
`“wideband speech signal,” Patent Owner does not identify a problem with
`any of Petitioner’s proposed interpretations. Although the parties express
`some slight difference in their respective interpretations of “wideband
`speech signal,” we determine that it is unnecessary to construe any claim
`terms for this Decision.
`B.
`Alleged Obviousness Over Salami and Kroon
`The Petition asserts that claims 1, 8, 15, 29, and 36 (all independent
`claims of the challenged claims) would have been obvious over the
`combined teachings of Salami and Kroon. Pet. 14–46.
`
`10
`
`
`
`IPR2017‐01244
`
`Patent 6,807,524 B1
`
`
`
`Salami (Ex. 1008)
`1.
`According to Salami, known CELP algorithms suffered from
`computational complexity that precluded real-time implementation for
`transmission of high quality digitized voice signals. Ex. 1008, 22.7 Salami
`purports to resolve this problem by disclosing an improved CELP coding
`structure (Salami refers to its structure as “Algebraic” CELP or simply
`“ACELP”). Id.
`Figure 1 of Salami is reproduced below with our annotation in red.
`
`
`
`
`Figure 1 of Salami shows a synthesis model (a decoder) for decoding a block
`of speech samples. Ex. 1008, 23. The speech samples are represented as
`encoded parameters enclosed by our added red box annotation. Each
`encoded sample is applied to algebraic codebook and shaping filter within
`dynamic algebraic codebook. Id. The decoded value is scaled by a gain
`factor (“G”) determined by the encoded parameters and the result thereof is
`
`
`7 Although the Petition cites portions of Exhibit 1008 and other exhibits
`using the page numbers of the underlying publication, we cite to the page
`numbering added to the Exhibits in compliance with our rules.
`
`11
`
`
`
`IPR2017‐01244
`
`Patent 6,807,524 B1
`
`
`applied to two time varying filters—first, long-term predictor (“LTP”)
`followed by short-term predictor (“STP”).
`The LTP filter is expressed in Salami as:
`
`
`
`
`where β is the pitch gain parameter and α is the pitch
`delay parameter. Id. The STP filter (also known as the linear prediction
`(“LP”) filter) is represented in Salami as:
`
`
`where p is the predictor order and ai are the predictor coefficients. Id.
`Speech samples may be encoded as shown in Salami’s Table 2
`reproduced below.
`
`
`Salami’s Table 2 describes the allocation of bits in encoded speech samples.
`Id. at 25.
`
`12
`
`
`
`IPR2017‐01244
`
`Patent 6,807,524 B1
`
`
`
`Salami teaches that the encoded parameters (pitch delay, pitch gain,
`codebook address, and codebook gain) are determined “using an analysis-
`by-synthesis technique” in which an encoded signal is selected from all
`candidate innovation sequences that minimizes a difference from the
`“original signal according to a perceptually weighted distortion measure.”
`Id. at 23. The error weighting filter is expressed in Salami as:
`
`
`
`
`
`where 0 < γ < 1. Id.
`
`To improve the LP analysis in encoding, Salami discloses, inter alia, that a
`preemphasis filter is applied to the input speech signal to emphasize higher
`frequencies. Id.
`
`Kroon (Ex. 1005)
`2.
`Kroon discloses a speech encoding/decoding technique asserted top
`lower computational complexity. Ex. 1005, 5. Kroon discloses, in pertinent
`part, that time-varying weighting filters contribute to computation
`complexity and, thus, discloses a time-invariant error-weighting filter to
`reduce computational complexity. Id. at 8.
`3.
`Analysis: Independent Claims 1, 8, 15, 29, and 36
`Petitioner identifies the recited features of independent claim 1 in the
`combined disclosures of Salami and Kroon. Pet. 27–37. Specifically, the
`preamble of claim 1 recites, “A perceptual weighting device for producing a
`perceptually weighted signal in response to a wideband speech signal in
`order to reduce a difference between the wideband speech signal and a
`subsequently synthesized wideband speech signal.” Petitioner argues Salami
`
`13
`
`
`
`IPR2017‐01244
`
`
`
`Patent 6,807,524 B1
`
`
`discloses such a device in that Figure 1, as reproduced supra, discloses a
`decoder that decodes perceptually weighted signals received from an
`encoder, the signals representing a wideband speech signal, and discloses the
`encoded signals reduce the difference between the synthesized speech signal
`and the original speech signal. Pet. 27–30 (citing Ex. 1008, 22–23, Fig. 1;
`Ex. 1003 ¶¶ 114 et seq.8).
`Claim 1 further recites the elements that comprise the claimed
`perceptual weighting device as including: a preemphasis filter; a synthesis
`filter calculator responsive to the preemphasis filter output; and a perceptual
`weighting filter responsive to the preemphasis filter output and the synthesis
`filter calculator output. The Petition identifies the recited preemphasis filter
`as Salami’s disclosure of applying a preemphasis filter to the input speech
`signal to enhance a high frequency range of the signal. Pet. 30–31 (citing
`Ex. 1008, 23). The Petition identifies the recited synthesis filter calculator
`as a linear prediction (LP) filter disclosed in Salami as the short-term
`predictor (STP) and asserts this filter is responsive to the preemphasized
`signal generated by the preemphasis filter. Id. at 31–32 (citing Ex. 1008,
`23). Lastly, the Petition identifies the recited perceptual weighting filter as
`Salami’s disclosure of an error-weighting filter and asserts the error-
`weighting filter is responsive to the preemphasized signal of the preemphasis
`filter. Id. at 32–35 (citing Ex. 1008, 23–24; Ex. 1003 ¶¶ 94–98, 114 et seq.).
`
`
`8 Petitioner cites paragraphs “114 et seq.” of Dr. Cohen’s Declaration
`(Ex. 1003) in support of several of its contentions. Pet. 30, 33, 35, 39, 40,
`42–46. Petitioner’s non-specific references to all paragraphs numbered 114
`and beyond are improper. Petitioner leaves it to this panel to determine
`which of paragraphs 114 through 161 of the Declaration are offered in
`support of its contentions.
`
`14
`
`
`
`IPR2017‐01244
`
`
`
`Patent 6,807,524 B1
`
`
`Petitioner acknowledges that Salami fails to disclose that the transfer
`function of the recited perceptual weighting filter has a fixed denominator
`(id. at 23), but argues Kroon, in combination with Salami, discloses a fixed
`(time-invariant) denominator for such a filter transfer function (id. at 35–36
`(citing Ex. 1005, 8–9; Ex. 1003 ¶¶ 90–92, 104–113)).
`Patent Owner argues, inter alia, the proposed combination of
`references fails to teach the recited preemphasis filter and the recited
`perceptual weighting filter. PO Resp. 32–37. More specifically, Patent
`Owner argues all challenged independent claims include an encoder, but
`Salami discloses a decoder. Id. at 33. Patent Owner contends claims 15, 29,
`and 36 explicitly recite an encoder while claims 1 and 8 are directed to an
`encoder because they each recite features that generate synthesized
`wideband speech with minimal differences from a wideband speech signal—
`i.e., a wideband speech signal is input to generate (encode) a synthesized
`wideband speech signal therefrom. Id. (citing Ex. 2004 ¶ 86). Patent Owner
`contends the ordinarily skilled artisan would have understood that a
`perceptual weighting filter would only exist in an encoder and that the
`decoder of Salami does not have a preemphasis filter. Id. Patent Owner
`asserts Petitioner’s argument is mere speculation regarding the components
`of an encoder based on Salami’s disclosure of a decoder. Id. at 34.
`Regarding the recited preemphasis filter, Patent Owner acknowledges that
`Salami discusses a preemphasis process, but contends Salami discloses
`preemphasis to improve LP analysis—not for the recited purpose of claim 1
`to enhance a high-frequency range of the wideband voice signal before
`further filter processing within the encoder. Id. 35–36. Therefore, Patent
`Owner argues, “[a]ccordingly, Salami would not have taught ‘a perceptual
`
`15
`
`
`
`
`
`IPR2017‐01244
`
`Patent 6,807,524 B1
`
`
`weighting filter, responsive to said preemphasized signal and said synthesis
`filter coefficients, for filtering said preemphasized signal’” as required by
`the claims. Id. at 36–37.
`After considering the parties’ arguments and cited evidence, we
`determine that Petitioner has not shown sufficiently that Salami teaches the
`identified limitations of claim 1. Figure 1 of Salami is a decoder of encoded
`speech signals—i.e., a “synthesis model” capable of decoding speech
`encoded according to the ACELP techniques disclosed in Salami. Ex. 1008,
`23 (“Figure 1 shows the synthesis model in ACELP coders.”). Although
`Salami’s Figure 1 shows encoded speech signals received by the decoding
`elements, the figure does not disclose the structure of an encoder that
`generates those encoded signals.
`Although Salami does not provide a figure depicting the structure of
`an encoder, we find Salami discloses use of a preemphasis filter in speech
`encoding. Ex. 1008, 23 (“The first is to preemphasize the input speech
`signal. . . . it emphasizes the higher frequencies in the speech signal”). We
`also find that Salami discloses some form of perceptual weighting and an
`element that generates linear predictor (LP) coefficients:
`The LP coefficients are determined using the method of linear
`prediction analysis by minimizing the mean-squared prediction
`error. The pitch parameters (delay and gain) and the codeboook
`[sic] parameters (address and gain) are determined at the encoder
`using an analysis-by-synthesis technique. In this technique, the
`synthetic speech is computed for all candidate innovation
`sequences in the codebook retaining the particular codeword that
`produces the output closer to the original signal according to a
`perceptually weighted distortion measure.
`
`Id.
`
`16
`
`
`
`IPR2017‐01244
`
`Patent 6,807,524 B1
`
`
`
`
`
`Thus, we determine Salami discloses the individual elements of the
`encoding device of, for example, claim 1. However, we agree with Patent
`Owner that Salami does not disclose that the preemphasized input speech
`signal is used as an input to a perceptual weighting filter as recited in claim
`1. See PO Resp. 35–37. Specifically, claim 1 specifies both the synthesis
`filter calculator and the perceptual weighting filter are responsive to the
`preemphasized signal generated by the preemphasis filter, which is, in turn,
`responsive to the input wideband speech signal, and further specifies that the
`perceptual weighting filter is also responsive to the coefficients generated by
`the synthesis filter calculator. In other words, the claim requires an
`arrangement of the various elements of an encoder in which the output of the
`preemphasis filter (the preemphasized signal) is applied to the synthesis
`filter calculator and applied to the perceptual weighting filter, and the
`coefficients generated by the synthesis filter calculator are applied to the
`perceptual weighting filter. An excerpt of Figure 1 of the ’524 patent,
`reproduced below, shows a block diagram of relevant portions of such an
`encoder arrangement.
`
`
`
`17
`
`
`
`
`
`IPR2017‐01244
`
`Patent 6,807,524 B1
`
`
`The above excerpt of Figure 1 of the ’524 patent depicts preemphasis filter
`103, coupled to receive wideband speech signals (“Sp”) and configured to
`generate preemphasized speech signal (“S”). Ex. 1001, 7:64–8:18. The
`output of preemphasis filter 103 (“S”) is applied to LP analysis quantization
`and interpolation calculator 104 and perceptual weighting filter 105. Id. at
`8:19–9:45. Petitioner directs us to no disclosure in Salami of such an
`arrangement of filters in an encoding device.
`Specifically regarding the perceptual weighting filter, Petitioner
`argues:
`Second, Salami discloses that the perceptual weighting filter is
`responsive to the pre-emphasized signal. For instance, Salami
`discloses that the pre-emphasis filter operates on the original
`input wideband speech signal (see [1.1]), and the preemphasized
`signal is sent to the LP filter for LP analysis (see [1.2]); then, the
`preemphasized signal is input to the perceptual weighting filter:
`“The LP coefficients are determined using the method of
`linear prediction analysis”; then, Salami discloses that
`“the synthetic speech is computed . . . according to a
`perceptually weighted distortion measure.”
`Ex-1008, [23]. Therefore, Salami states here that the pre-
`emphasized wideband speech signal is first used for LP analysis,
`and next, the pre-emphasized signal is perceptually weighted and
`used to compute the excitation codeword. Ex-1003, ¶¶ 114 et
`seq.
`Pet. 33 (emphases in original). We are not persuaded by Petitioner’s
`argument. Petitioner highlights two quotations from Salami and excludes
`significant text between the two quotes. Including and highlighting the text
`excluded by Petitioner, Salami discloses:
`The LP coefficients are determined using the method of linear
`prediction analysis by minimizing the mean-squared prediction
`error. The pitch parameters (delay and gain) and the
`
`18
`
`
`
`IPR2017‐01244
`
`Patent 6,807,524 B1
`
`
`
`
`
`codeboook [sic] parameters (address and gain) are determined
`at the encoder using an analysis-by synthesis technique. In this
`technique, the synthetic speech is computed for all candidate
`innovation sequences in the codebook retaining the particular
`codeword that produces the output closer to the original signal
`according to a perceptually weighted distortion measure.
`Ex. 1008, 23. We discern no teaching or suggestion in this text of Salami
`that the preemphasized signal is used on a perceptual weighting filter. We
`find no support in the cited portions of Salami for Petitioner’s assertion that
`the preemphasized signal is first applied to LP analysis and then the same
`preemphasized signal is perceptually weighted (i.e., applied to a perceptually
`weighted filter). Petitioner’s argument in the Petition does not explain
`specifically how the cited text teaches the identified claim limitations and,
`thus, amounts to little more than an unsupported conclusory remark. Dr.
`Cohen merely repeats the same argument without providing any further
`explanation. Ex. 1003, 65 (the portion of the claim chart of paragraph 114
`for this element of claim 1).
`Patent Owner argues, “[i]n contrast to the ’524 patent in which the
`preemphasised signal is used for both the LPC analysis and the perceptual
`weighting filter (Ex. 1001, Figure 1 (reproduced above)), a POSITA would
`have understood Salami’s encoder to filter the input speech as shown
`below.” PO Resp. 36; see also Ex. 2004 ¶ 92. Patent Owner’s diagram of a
`possible alternate arrangement of elements in an encoder of Salami is
`reproduced below.
`
`
`
`19
`
`
`
`IPR2017‐01244
`
`
`
`Patent 6,807,524 B1
`
`
`Patent Owner’s proposed alternate arrangement shows an input speech
`signal (“S(n)”) applied to preemphasis filter, the output of which is applied
`to LPC analysis to generate linear prediction coefficients (“A(z)”) applied to
`“Salami Encoder” along with the original input speech signals (“S(n)”).
`We have no basis for presuming Patent Owner’s suggested
`arrangement of elements for an encoder in Salami is a correct interpretation.
`Nonetheless, we agree that, in the absence of disclosure within Salami
`regarding the specific structure of its encoder, Petitioner has not shown
`sufficiently that the ordinarily skilled artisan would have perceived Salami
`to disclose the particular arrangement recited in, for example, claim 1. Thus,
`Petitioner has not shown sufficiently that the ordinarily skilled artisan would
`have understood Salami to disclose or suggest application of a
`preemphasized signal to a perceptual weighted filter as required by claim 1.
`Independent claim 8 includes a similar recitation of encoder features
`for which Petitioner relies on the same argument as for claim 1. Pet. 37–38.
`Claim 15 recites “a perceptual weighting device as recited in claim 1” and,
`thus, incorporates the same encoder limitations for which Petitioner relies on
`the same arguments as for claim 1. Id. at 38. Claims 29 and 36 each recite
`“an encoder for encoding a wideband speech signal as recited in claim 15”
`and, thus, each incorporates the same encoder features as claim 1. For these
`encoder features of claims 29 and 36, Petitioner relies on the same
`arguments as for claim 1. Id. at 44, 46.
`Accordingly, we are no