Apple Inc v. Saint Lawrence Communications LLC

1. Case Identification

• Case #: IPR2017-01075
• Patent #: 7,151,802
• Filed: March 13, 2017
• Petitioner(s): Apple Inc.
• Patent Owner(s): Saint Lawrence Communications LLC
• Challenged Claims: 1-3, 8-11, 16, 25-27, 32-35, 40, 49-50, and 52-53

2. Patent Overview

• Title: High-Frequency Content Recovery Method and Device for Wideband Speech Signal
• Brief Description: The ’802 patent discloses a method and device for recovering high-frequency content in a wideband speech signal that was lost during encoding. The technology uses Code Excited Linear Prediction (CELP) techniques and a high-frequency regeneration process that involves generating, shaping, and injecting noise into a synthesized speech signal.

3. Grounds for Unpatentability

Ground 1: Obviousness over Schnitzler in view of Pyke - Claims 1-3, 8-11, 16, 25-27, 32-35, 40, 49-50, and 52-53 are obvious over Schnitzler in view of Pyke.

• Prior Art Relied Upon: Schnitzler (“A 13.0 kbit/s Wideband Speech Codec Based on SB-ACELP,” 1998) and Pyke (“Extrapolation of Wideband Speech From the Telephone Band,” 1997 Master’s Thesis).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Schnitzler disclosed a wideband speech decoder that uses ACELP (a form of CELP) and high-frequency resynthesis, teaching the core CELP-related limitations of the challenged claims. This included a demultiplexer, pitch and innovative codebooks, a combiner, a synthesis filter, and an oversampler. Schnitzler’s high-frequency recovery module added spectrally shaped, gain-adjusted noise to improve the signal. However, Petitioner contended that Schnitzler injected this noise into an un-synthesized signal version of the lower frequency band.
  • Petitioner asserted that Pyke cured this deficiency. Pyke taught a system for extrapolating telephone-band speech to create a full-spectrum, pseudo-wideband signal. Critically, Pyke’s method involved generating a random white noise sequence, shaping it based on linear prediction coefficients, and then splicing (injecting) the resulting signal into an oversampled, synthesized telephone-band signal. Petitioner argued that combining Schnitzler's established CELP decoder architecture with Pyke's specific high-frequency regeneration technique of injecting shaped noise into a synthesized signal rendered the claims of the ’802 patent obvious.
  • Motivation to Combine: Petitioner asserted that a person of ordinary skill in the art (POSITA) would combine Schnitzler and Pyke to improve the sound quality of the synthesized speech. Both references addressed the same problem of achieving wideband quality at low bit rates using CELP-based high-frequency regeneration. Petitioner argued that the high-frequency regeneration (HFR) modules in both systems were functionally modular and could be interchanged. A POSITA would have been motivated to substitute Schnitzler’s HFR module with Pyke’s superior technique—which injected noise into the synthesized domain—as a simple substitution of one known element for another to obtain the predictable result of improved audio quality.
  • Expectation of Success: A POSITA would have had a high expectation of success because the systems were highly compatible. Both disclosed a CELP decoding portion followed in series by an HFR module, with a common input (a CELP-synthesized signal) and a common output (a high-frequency band added to that signal). This compatibility would have made the integration a matter of routine engineering.

4. Key Claim Construction Positions

• Petitioner dedicated significant argument to several means-plus-function claim terms, largely adopting constructions from a related district court case but arguing for the inclusion of additional structures disclosed in the ’802 patent’s specification.
• “spectral shaping unit for shaping the spectrum of the noise sequence...”: Petitioner adopted the function identified by the district court but argued that the corresponding structure should include not only the processor and filter identified by the court but also algorithms described in the specification where a gain adjusting module and LP filter provide inputs to the spectral shaper.
• “gain adjustment module...for producing a scaled white noise sequence”: Similarly, Petitioner adopted the court's identified function but contended the structure should also include algorithms where an energy computing module and spectral tilt calculator provide input signals to the gain adjustment module, which the district court’s construction allegedly ignored.
• “signal fragmenting device for receiving an encoded version of a wideband signal...”: Petitioner adopted the court's function but argued the corresponding structure should also encompass the demultiplexer (element 217 in Fig. 2 of the patent) which receives binary information and extracts synthesis model parameters as described in the specification.

5. Arguments Regarding Discretionary Denial

• Petitioner argued that discretionary denial under 35 U.S.C. § 325(d) would be inappropriate. While acknowledging two prior IPR petitions filed by other parties against the ’802 patent (IPR2016-00704 and IPR2015-01874), Petitioner asserted that the present petition relied on different prior art combinations and arguments.
• Specifically, Petitioner highlighted that its challenge to the independent claims relied on the Pyke reference, which was not used in the prior IPRs. Therefore, Petitioner contended that because it presented new prior art combinations and arguments, the petition fell outside the scope of § 325(d).

6. Relief Requested

• Petitioner requested the institution of an inter partes review and cancellation of claims 1-3, 8-11, 16, 25-27, 32-35, 40, 49-50, and 52-53 of the ’802 patent as unpatentable.