DCT

2:26-cv-00878

Cerence Operating Co v. Sony Group Corp

Log In
Key Events
Complaint
complaint Intelligence

I. Executive Summary and Procedural Information

  • Parties & Counsel:
  • Case Identification: 2:26-cv-00878, E.D. Tex., 10/01/2026
  • Venue Allegations: Venue is alleged to be proper based on Defendant Sony Group Corporation being a foreign entity not resident in the United States. For the other Sony defendants, venue is based on their being registered to do business in Texas, transacting business in the district, and maintaining regular and established places of business in Plano and Frisco, Texas.
  • Core Dispute: Plaintiff alleges that Defendant’s consumer electronics with voice processing capabilities, including the PlayStation 5 and noise-cancelling headphones, infringe five U.S. patents related to speech recognition, noise suppression, and voice command processing.
  • Technical Context: The technology at issue involves conversational artificial intelligence (AI) and voice control systems, which are integral to the functionality of modern automotive, mobile, and home entertainment products.
  • Key Procedural History: The complaint alleges a complex history between the parties. Plaintiff Cerence was spun out of Nuance Communications, Inc., which reportedly entered a ten-year strategic partnership with Defendant Sony's subsidiary, Gracenote, in 2012 to develop voice-controlled interfaces. Plaintiff also notes prior litigation against Defendants filed in August 2025 and a licensing presentation shared with Defendants in November 2025 that identified some of the patents now asserted in this case. This history may be relevant to the allegations of pre-suit knowledge and willful infringement.

Case Timeline

Date Event
2004-09-23 ’872 Patent Priority Date
2006-12-14 ’815 Patent Priority Date
2012-01-01 Nuance and Gracenote (Sony subsidiary) partnership begins (approx. date)
2012-06-05 ’872 Patent Issue Date
2012-11-06 ’815 Patent Issue Date
2013-03-12 ’750 Patent Priority Date
2016-02-05 ’899 Patent Priority Date
2019-03-28 ’073 Patent Priority Date
2019-10-01 Cerence becomes a separate public company (approx. date)
2020-09-22 ’899 Patent Issue Date
2021-08-10 ’750 Patent Issue Date
2024-03-12 ’073 Patent Issue Date
2025-08-01 Cerence files "Prior Litigation" against Sony (approx. date)
2025-11-18 Cerence provides "Licensing Presentation" to Sony
2026-10-01 Complaint Filing Date

II. Technology and Patent(s)-in-Suit Analysis

U.S. Patent No. 11,929,073 - "Hybrid Arbitration System"

  • Patent Identification: U.S. Patent No. 11,929,073, “Hybrid Arbitration System,” issued March 12, 2024.

The Invention Explained

  • Problem Addressed: In hybrid voice interface systems, processing occurs both on-device ("embedded") and on remote servers ("cloud-based") Compl. ¶31 ’073 Patent, col. 1:19-33 These two paths can produce different speech recognition results, and the cloud-based result typically has higher latency. The technical problem is how to efficiently select the best result to respond to a user, balancing the speed of the on-device system with the potential for higher accuracy from the cloud system '073 Patent, background
  • The Patented Solution: The patent describes an arbitration method for selecting between the local and cloud speech recognition results '073 Patent, abstract Upon receiving a result from the local on-device processor, an "arbitrator" determines whether to select that result immediately or to wait for the result from the cloud service '073 Patent, col. 2:5-12 This decision is based on analyzing features of the local result '073 Patent, abstract If the system waits, it then compares the two results to select the final output, for example by using a classifier to determine which is better '073 Patent, Fig. 2
  • Technical Importance: This arbitration logic aims to improve user experience by reducing perceived latency in voice assistant interactions, responding immediately when local processing is deemed sufficient while retaining the option to use more powerful cloud resources for more complex queries.

Key Claims at a Glance

  • The complaint asserts at least one exemplary independent claim Compl. ¶34 The patent contains independent method claim 1 and independent system claim 14.
  • Key elements of independent claim 1 include:
    • Receiving a first speech recognition result determined by a local computing device.
    • Receiving first features, determined using the first speech recognition result.
    • Determining whether to select the first speech recognition result or to wait for a second speech recognition result from a cloud service, based on the first result and first features.
  • The complaint reserves the right to assert additional claims Compl. ¶33

U.S. Patent No. 10,783,899 - "Babble Noise Suppression"

  • Patent Identification: U.S. Patent No. 10,783,899, “Babble Noise Suppression,” issued September 22, 2020.

The Invention Explained

  • Problem Addressed: Standard speech detection and noise reduction methods struggle in environments with "babble noise"—that is, background speech from interfering speakers '899 Patent, col. 5:1-12 This interfering speech can be falsely identified as desired user speech, degrading the performance of voice-controlled applications '899 Patent, col. 5:13-20
  • The Patented Solution: The patent proposes a system with a "soft speech detector" that, instead of making a simple yes/no decision, calculates a "speech detection result" indicating the likelihood of the user's speech being present '899 Patent, abstract ’899 Patent, col. 6:12-20 This likelihood is determined using signal characteristics such as kurtosis, which helps distinguish a single speaker's voice from a mix of background voices '899 Patent, col. 7:50-53 This likelihood result is then used to dynamically control the aggressiveness of a noise suppressor, allowing for stronger attenuation of babble noise during speech pauses '899 Patent, col. 13:20-49
  • Technical Importance: This technology allows for more robust voice control and recognition in acoustically challenging, crowded public spaces where multiple people may be speaking simultaneously.

Key Claims at a Glance

  • The complaint asserts at least one exemplary independent claim Compl. ¶46 The patent contains independent method claim 1 and independent system claim 11.
  • Key elements of independent system claim 11 include:
    • A "soft speech detector" configured to dynamically determine a "speech detection result indicating a likelihood of a presence of the foreground speech components."
    • A "noise suppressor" coupled to the detector, configured to dynamically compute "spectral weighting coefficients based on the speech detection result."
    • The noise suppressor is further configured to apply the coefficients to the audio signal to suppress background noise in a dynamic manner.
  • The complaint reserves the right to assert additional claims Compl. ¶45

U.S. Patent No. 8,306,815 - "Speech Dialog Control Based on Signal Pre-Processing"

  • Patent Identification: U.S. Patent No. 8,306,815, “Speech Dialog Control Based on Signal Pre-Processing,” issued November 6, 2012 Compl. ¶55
  • Technology Synopsis: The patent describes a speech dialog system where a signal pre-processing unit generates not only an "enhanced speech signal" for recognition, but also a separate "analysis signal" '815 Patent, abstract This analysis signal conveys non-semantic information about the input audio, such as the noise level, speaker location, or the speaker's pitch '815 Patent, col. 2:33-45 A control unit uses this analysis signal to manage the user dialog, for example by adjusting output volume in a noisy environment or tailoring content based on the detected age or gender of the user '815 Patent, Figs. 4, 6
  • Asserted Claims: The complaint asserts at least one exemplary independent claim Compl. ¶58 Independent claims 1 and 20 are present in the patent.
  • Accused Features: The complaint accuses the Sony PlayStation 5 of infringement, focusing on its voice command and dialog functionalities Compl. ¶56 Compl. ¶66

U.S. Patent No. 11,087,750 - "Methods and Apparatus for Detecting a Voice Command"

  • Patent Identification: U.S. Patent No. 11,087,750, “Methods and Apparatus for Detecting a Voice Command,” issued August 10, 2021 Compl. ¶69
  • Technology Synopsis: The patent discloses a method for a mobile device to perform "always-on" monitoring for a voice command, including while in a low-power mode, without requiring an explicit trigger phrase '750 Patent, col. 2:7-13 The method uses a multi-stage processing hierarchy to conserve power, often starting with low-power processors for initial analysis '750 Patent, col. 13:8-16 The system may also use "contextual cues"—such as device motion, location, or time of day—to assist in accurately detecting a voice command while minimizing power consumption '750 Patent, abstract ’750 Patent, col. 5:59-63
  • Asserted Claims: The complaint asserts at least one exemplary independent claim Compl. ¶72 Independent claim 1 is the primary method claim in the patent.
  • Accused Features: The complaint accuses the Sony PlayStation 5, specifically its voice command functionality, of infringement Compl. ¶70 Compl. ¶78

U.S. Patent No. 8,194,872 - "Multi-Channel Adaptive Speech Signal Processing System with Noise Reduction"

  • Patent Identification: U.S. Patent No. 8,194,872, “Multi-Channel Adaptive Speech Signal Processing System with Noise Reduction,” issued June 5, 2012 Compl. ¶83
  • Technology Synopsis: The patent describes a noise reduction system for microphone arrays that uses an adaptive beamformer to focus on a desired speech signal '872 Patent, abstract The system continuously updates the beamformer's weights to emphasize microphone signals with higher energy, which are presumed to contain the desired speech '872 Patent, col. 2:10-18 It further employs adaptive noise cancellation logic that generates a noise estimate from the input signals and subtracts it from the beamformed signal to remove residual noise '872 Patent, abstract
  • Asserted Claims: The complaint asserts at least one exemplary independent claim Compl. ¶86 Independent claims 1 and 21 are present in the patent.
  • Accused Features: The complaint accuses Sony’s WH-1000XM6 and WF-1000XM6 noise-cancelling headphones, which are alleged to use multi-microphone arrays for noise reduction and voice capture Compl. ¶84 Compl. ¶90

III. The Accused Instrumentality

Product Identification

  • The complaint identifies the "Accused Products" as a broad class of Sony products with voice processing functionality Compl. ¶1 Specific exemplary products include the Sony PlayStation 5 with DualSense Wireless Controller, Sony WH-1000XM6 and WF-1000XM6 noise-cancelling headphones, Sony's XAV Series Wireless Auto Media Receivers, and other Sony microphone and audio systems Compl. ¶1

Functionality and Market Context

  • The infringement allegations center on the voice control and audio processing features of these products. For the PlayStation 5, this includes the "voice command functionality" that allows users to control the console with their voice Compl. ¶38 fn. 4
  • For the accused headphones, the relevant functionality is the "AI-based beamforming and noise reduction" systems, which use microphone arrays to isolate the user's voice from background noise for phone calls and interaction with voice assistants Compl. ¶50 fn. 5
  • No probative visual evidence provided in complaint.

IV. Analysis of Infringement Allegations

The complaint references claim-chart exhibits for each asserted patent Compl. ¶34 Compl. ¶46 Compl. ¶58 Compl. ¶72 Compl. ¶86, but these exhibits were not included with the pleading. Therefore, a summary of the narrative infringement theory is provided in lieu of a claim-chart table.

'073 Patent Infringement Allegations

The complaint alleges that the Sony PlayStation 5's voice command system infringes the ’073 Patent Compl. ¶32 The narrative theory suggests that when a user issues a voice command, the PlayStation 5 utilizes a hybrid processing architecture involving both on-device and cloud-based resources to perform speech recognition and natural language understanding. The infringement allegation centers on the system's need to arbitrate between the faster on-device result and the potentially more accurate cloud-based result. The complaint implies that the PlayStation 5 performs a method of receiving the local result, analyzing its features (such as a confidence score), and then deciding whether to act immediately or wait for the cloud result, thereby mapping its function onto the method claimed in the ’073 patent.

'899 Patent Infringement Allegations

The complaint alleges that Sony’s noise-cancelling headphones (e.g., WH-1000XM6) infringe the ’899 Patent Compl. ¶44 These products are marketed as featuring "AI-based beamforming and noise reduction" Compl. ¶50 fn. 5 The infringement theory posits that the headphone's microphone array and processing system perform the claimed method of "babble noise suppression." This likely involves the system analyzing an incoming audio signal to compute a likelihood, or "soft" probability, that desired user speech is present versus background babble. This likelihood score is then allegedly used to dynamically control the level of noise suppression applied, thereby isolating the user's voice for calls or voice assistant queries in a manner that maps onto the claims of the ’899 patent.

Identified Points of Contention

  • Scope Questions: For the ’073 Patent, a key dispute may arise over the definition of "arbitrator" and the specific logic of "determining whether to... wait." The question for the court may be whether this requires a specific, sequential decision-making process before the cloud result is available, or if it could also read on a system that processes both locally and in the cloud in parallel and simply selects the result with the highest score after both are complete.
  • Technical Questions: For the ’899 Patent, the dispute will likely center on technical evidence. The question will be whether Sony's proprietary "AI-based" noise reduction performs the specific technical steps of the asserted claims. For instance, does the accused system actually calculate a "speech detection result indicating a likelihood" based on signal "kurtosis" as taught in the patent's specification, or does it employ an alternative, non-infringing machine learning model to achieve a similar end result?

V. Key Claim Terms for Construction

'073 Patent

  • The Term: "determining whether to select the first speech recognition result or to wait for a second speech recognition result" (from claim 1).
  • Context and Importance: This term defines the core decision-making step of the claimed arbitration process. Its construction is critical because it will determine whether the claim covers only systems that make an explicit, intermediate decision to "wait," or if it could also cover systems that process in parallel and select a result later.
  • Intrinsic Evidence for a Broader Interpretation: The claim language itself does not specify the precise timing or mechanism of the determination, which may support an interpretation that covers any process that logically decides between using a local result now or a cloud result later.
  • Intrinsic Evidence for a Narrower Interpretation: The patent’s detailed description and figures describe a specific embodiment with a "short-circuit" classifier that makes an initial decision, and a separate "system classifier" that is used only if the system waits '073 Patent, Fig. 2 '073 Patent, col. 5:37-51 This two-step process could be used to argue that the "determining" step requires a discrete, upfront decision-making stage, narrowing the claim's scope.

'899 Patent

  • The Term: "soft speech detector" (from claim 11).
  • Context and Importance: Practitioners may focus on this term because it distinguishes the invention from conventional "hard" detectors that provide a binary (yes/no) output. The case may turn on whether the accused headphones' system generates a non-binary, probabilistic "likelihood" of speech that is then used to dynamically control noise suppression, as the claim requires.
  • Evidence for a Broader Interpretation: The specification describes the output of the soft speech detector as a "speech detection result indicating a likelihood (i.e., a probability value) of a presence of the foreground speech components" '899 Patent, col. 6:15-18 This general language could support a broader definition covering any system that produces a probabilistic speech-presence score.
  • Evidence for a Narrower Interpretation: The patent extensively details an embodiment where this likelihood is calculated as a function of specific features, namely "kurtosis" and a "cepstral maximum feature" '899 Patent, col. 7:50-53 '899 Patent, col. 12:21-27 A defendant could argue that the term "soft speech detector" is implicitly limited by the specification to a detector that uses these or technically similar features.

VI. Other Allegations

Indirect Infringement

  • The complaint alleges both induced and contributory infringement for all asserted patents. Inducement is primarily based on allegations that Sony disseminates product information and user manuals that instruct and encourage end-users to operate the Accused Products in an infringing manner Compl. ¶38 Compl. ¶50 Contributory infringement is based on allegations that Sony provides technology that is especially made for infringement and is not a staple article of commerce with substantial non-infringing uses Compl. ¶40 Compl. ¶52

Willful Infringement

  • The complaint alleges willful infringement based on both pre-suit and post-suit knowledge. Pre-suit knowledge is alleged to arise from a multi-year commercial relationship between Sony and Cerence's predecessor (Nuance), prior litigation between the parties involving related technologies, and a specific "Licensing Presentation" in November 2025 where Plaintiff allegedly identified some of the asserted patents to Defendants Compl. ¶¶24-29 Compl. ¶36

VII. Analyst’s Conclusion: Key Questions for the Case

  • A core issue will be one of technical and evidentiary proof: For patents covering specific algorithms like the ’899 Patent (“Babble Noise Suppression”), can the Plaintiff demonstrate through discovery and reverse engineering that Sony’s proprietary “AI-based” systems perform the specific technical steps required by the claims—such as calculating signal kurtosis to create a soft speech likelihood—or will analysis show that Sony achieves a similar result via a different, non-infringing technical pathway?
  • A second central issue will be one of claim construction and scope: For process-oriented patents like the ’073 Patent (“Hybrid Arbitration System”), the case may turn on whether the claimed “arbitration” process is construed narrowly to require a specific, sequential decision-making logic, or broadly enough to read on systems that might process local and cloud results in parallel before selecting a final output.
  • A third key question, particularly relevant for willfulness and potential damages enhancement, will be the nature and timing of Defendant's knowledge: Given the extensive prior relationship and litigation history alleged in the complaint, the court will need to determine precisely what Sony knew about the asserted patents, and when, to assess whether its conduct rose to the level of objective recklessness required for a finding of willful infringement.