PTAB

IPR2025-00460

Samsung Electronics Co Ltd v. Cerence Operating Co

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Techniques for Generating Synthetic Speech with Contrastive Stress
  • Brief Description: The ’291 patent discloses techniques for improving text-to-speech (TTS) systems by analyzing text input to identify tokens (e.g., words) that should contrast with each other and then generating synthetic speech that applies contrastive stress to the differing portions.

3. Grounds for Unpatentability

Ground 1A: Obviousness over Walker, Matsumoto, and Malsheen - Claims 1-4, 6-11, 13-17, and 19-20 are obvious over Walker in view of Matsumoto and Malsheen.

  • Prior Art Relied Upon: Walker (Application # 2001/0049602), Matsumoto (Application # 2007/0233492), and Malsheen (Patent 5,634,084).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Walker discloses a foundational text-to-speech (TTS) system that converts text, such as a weather report, into audible speech for a user. Matsumoto improves upon such systems by teaching a method to make synthesized speech more natural and easier to understand. It achieves this by comparing sequential, similar sentences (e.g., "Today's weather in Tohoku is fair" vs. "Today's weather in Kanto is fair"), identifying corresponding words that differ (e.g., "Tohoku" vs. "Kanto"), and modifying their prosody (pitch, volume, duration) to emphasize the difference. Malsheen teaches a text normalization pre-processing step, specifically expanding abbreviations like state names into full words ("NC" to "North Carolina") to ensure proper pronunciation by a TTS engine. Petitioner asserted that the combined teachings disclose identifying a first portion of a token that differs from a corresponding token ("Kanto" vs. "Tohoku") and a second portion that does not, and then assigning contrastive stress to the differing portion.
    • Motivation to Combine: A person of ordinary skill in the art (POSITA) seeking to improve the naturalness of Walker's TTS system for weather reports would be motivated to implement Matsumoto's method for emphasizing differing information, a common feature in such reports. Because weather reports often contain abbreviations for locations, a POSITA would also incorporate the well-known text normalization technique taught by Malsheen to ensure correct pronunciation before applying Matsumoto’s prosodic adjustments.
    • Expectation of Success: Petitioner contended that a POSITA would have had a reasonable expectation of success because combining these known elements—a base TTS system, a prosodic enhancement module, and a text normalization pre-processor—involved applying known techniques to a known system to achieve the predictable result of more natural and accurate synthesized speech.

Ground 1B: Obviousness over Walker, Matsumoto, Malsheen, and Bellegarda - Claims 1-4, 6-11, 13-17, and 19-20 are obvious over Walker, Matsumoto, and Malsheen in view of Bellegarda.

  • Prior Art Relied Upon: Walker (Application # 2001/0049602), Matsumoto (Application # 2007/0233492), Malsheen (Patent 5,634,084), and Bellegarda (Patent 7,313,523).
  • Core Argument for this Ground:
    • Prior Art Mapping: This ground was presented as an alternative to Ground 1A, addressing a specific claim limitation requiring not assigning contrastive stress to the non-differing portion of a token. Petitioner noted that Matsumoto teaches deemphasizing words that are the same in consecutive sentences. To the extent deemphasizing could be considered a form of stress, Petitioner introduced Bellegarda. Bellegarda teaches a TTS system that tracks previously spoken words and applies no accent or change to newly synthesized words that are the same as those recently spoken. This explicitly teaches the negative limitation of not applying stress to unchanged portions.
    • Motivation to Combine: A POSITA implementing the Walker-Matsumoto-Malsheen combination might wish to avoid deemphasizing repeated words to improve clarity or reduce computational load. Bellegarda provides an obvious solution: modify the system to leave the prosody of identical, corresponding words unchanged. This would be a simple substitution of one known technique (deemphasis in Matsumoto) for another (no change in Bellegarda) to achieve the predictable result of emphasizing only the differing words.
    • Expectation of Success: The modification would be straightforward for a POSITA, as Matsumoto's system already includes user-settable minimum and maximum variation coefficients. A POSITA could, as taught by Bellegarda, simply set the minimum variation coefficient to 1.0, which would result in no prosodic change for identical words.

4. Key Claim Construction Positions

  • "Contrastive Stress": Petitioner argued that, based on the ’291 patent’s specification and the prosecution history of a related patent, "contrastive stress" is a term of art referring to a specific type of emphasis meant to contrast words. It is characterized by increasing prosody (e.g., increased pitch, amplitude, or duration). Petitioner contended that deemphasizing a word by decreasing its prosody, as taught in Matsumoto for non-differing words, does not constitute applying "contrastive stress" as understood in the patent.

6. Arguments Regarding Discretionary Denial

  • Petitioner argued that discretionary denial under Fintiv would be inappropriate. The petition was filed approximately 12 months before the scheduled trial date in the parallel district court litigation, a timeline Petitioner argued is speculative and weighs only minimally, if at all, in favor of denial. Petitioner also asserted that investment in the parallel litigation was low at the time of filing, with key events like claim construction and expert reports not yet due. Finally, Petitioner noted an incomplete overlap in issues, as this petition challenges claims that are not asserted in the district court case, meaning the IPR would promote efficiency by resolving issues the court will not address.

7. Relief Requested

  • Petitioner requests institution of an inter partes review and cancellation of claims 1-4, 6-11, 13-17, and 19-20 of the ’291 patent as unpatentable.