US 9203972

An analysis of U.S. Patent No. 9,203,972 reveals the following details regarding its origin, ownership, and the technological advancements it protects.

Title: Efficient audio signal processing in the sub-band regime

Assignee: The current assignee of record is Cerence Operating Co. The original assignee was Nuance Communications Inc.

Inventors: Gerhard Uwe Schmidt, Hans-Jörg Köpf, and Günther Wirsching.

Filing Date: September 14, 2012

Issue Date: December 1, 2015

Abstract:
The patent describes a signal processing system designed to enhance an audio signal. The method involves dividing the audio signal into multiple audio sub-band signals. A portion of these sub-band signals is then excised, or removed. The remaining sub-band signals are processed to create enhanced versions. Following this, at least some of the previously excised sub-band signals are reconstructed. Finally, these reconstructed sub-band signals are combined with the enhanced sub-band signals to generate a final, enhanced audio signal.

Overview of Independent Claims:

This patent has five independent claims which outline the core novelties of the invention.

Claim 1 details a method for audio signal processing. This method involves:

• Dividing a signal from a microphone into several "microphone sub-band signals."
• Removing a predetermined number of these sub-band signals.
• Processing the remaining sub-band signals to reduce noise or echo, resulting in "enhanced microphone sub-band signals."
• Reconstructing the sub-band signals that were initially removed, using the enhanced signals.
• A key part of the noise/echo attenuation process involves using a "reference signal" which is also divided into sub-bands, with a corresponding number of sub-bands excised. An echo compensation filter is then adjusted based on the remaining reference sub-band signals and used to filter the remaining microphone sub-band signals.

Claim 5 describes a similar method of audio signal processing with a specific focus on the reconstruction step. In this claim, the reconstruction of an excised microphone sub-band signal is achieved by averaging the remaining microphone sub-band signals that are adjacent in time to the one that was removed.

Claim 6 also outlines a method for audio signal processing, but with a different reconstruction technique. Here, the excised microphone sub-band signals from a specific point in time are reconstructed through interpolation. This interpolation uses the remaining microphone sub-band signals from that same moment in time as well as those from adjacent moments in time.

Claim 8 shifts focus from the method to the signal processing system itself. It describes a system comprising:

• An "analysis filter bank" to divide the audio signal into sub-bands.
• A "first filter" to remove a subset of these sub-band signals.
• A "second filter" to process the remaining sub-band signals into enhanced versions.
• A "processor" to reconstruct at least some of the removed sub-band signals.
• A "synthesis filter bank" to combine the reconstructed and enhanced signals into a final enhanced audio signal.
• The novelty in this claim lies in the processor's function, which involves averaging two signals from the enhanced audio sub-band signals to create a replacement for one of the excised signals.

Claim 12 describes a signal processing system similar to that in Claim 8. However, the unique aspect of this claim is the processor's method for reconstruction. It reconstructs excised audio sub-band signals from a particular time by interpolating the remaining sub-band signals from that same time and adjacent times.

Claim 13 describes another variation of the signal processing system. The key feature of this claim is the "second filter," which is an echo compensation filter. The system also includes components to process a reference signal in a similar manner to the primary audio signal (dividing it into sub-bands and excising a subset). The echo compensation filter is adapted based on the remaining reference sub-band signals and then used to remove echo from the remaining audio sub-band signals.

As of today's date, May 8, 2026, a search of the CAFC dockets for 2026 did not yield any public records of litigation involving US Patent 9,203,972. However, this does not definitively mean no litigation exists, as some records may not be publicly accessible or indexed in the manner searched.

As of my analysis date of April 26, 2026, I have found no records of litigation involving U.S. Patent No. 9,203,972. A thorough search of patent litigation databases and legal news sources did not reveal any cases where this specific patent was asserted.

Proceedings overview

There are no recorded AIA trial proceedings (Inter Partes Review, Post-Grant Review, or Covered Business Method review) on file for US Patent 9,203,972 as of today's date, May 29, 2026. This indicates that the patent has not been challenged at the Patent Trial and Appeal Board (PTAB).

Strategic summary

As of the current date, all claims (1-13) of US Patent 9,203,972 remain untested by any AIA trial proceeding at the PTAB. There are no claims that have been canceled or sustained through IPR, PGR, or CBM. The patent has not been subjected to any challenges regarding its patentability before the PTAB.

Since no PTAB proceedings have been initiated, there is no estoppel landscape to consider. Potential petitioners are not barred from raising any ground that they raised or reasonably could have raised, as no such proceedings have occurred.

The absence of PTAB activity provides no specific pattern signals regarding petitioner behavior, patent owner litigation strategy, or involvement of defensive aggregators.

Recommended next steps

Since no PTAB activity exists for US Patent 9,203,972, the recommended next steps for a defendant facing assertion of this patent would be:

• Conduct a thorough prior art search: While the patent's original prosecution involved cited prior art, a new, comprehensive search could uncover additional references not previously considered, which might form the basis for a strong invalidity defense, either in district court or a potential IPR.
• Evaluate IPR potential: Assess the strength of potential obviousness or anticipation arguments against the claims based on newly identified or re-evaluated prior art. Given the lack of prior PTAB challenges, this route is fully open.
• Monitor for future PTAB filings: Stay vigilant for any new IPR, PGR, or CBM petitions filed against US 9,203,972 by other parties, as such filings could impact the patent's strength and provide insights into effective invalidity arguments.
• Engage in licensing negotiations or district court defense: The absence of PTAB challenges means that the patent's claims are presumed valid and have not been "hardened" or "softened" by a PTAB decision. Any defense would currently focus on arguments related to non-infringement or invalidity in a district court setting, or through direct negotiation.

Inventors

• Gerhard Uwe Schmidt: Likely employed by Harman Becker Automotive Systems GmbH at the time of the priority filing (October 1, 2007), as Harman Becker was the original assignor of the patent rights to Nuance Communications Inc. for this invention.
• Hans-Jörg Köpf: Likely employed by Harman Becker Automotive Systems GmbH at the time of the priority filing.
• Günther Wirsching: Likely employed by Harman Becker Automotive Systems GmbH at the time of the priority filing.

Original assignee

The original assignee listed on the issued U.S. Patent No. 9,203,972 is Nuance Communications Inc. Nuance was primarily a software company specializing in speech recognition, natural language understanding, and artificial intelligence, with products like Dragon NaturallySpeaking. The claims of this patent, relating to efficient audio signal processing for noise and echo reduction in sub-band regimes, align with Nuance's core business in voice recognition and hands-free communication systems. Nuance Communications Inc. was acquired by Microsoft in 2022 and continues to operate as a part of Microsoft.

Assignment timeline

• 2009-05-01 (executed) / recorded 2012-09-19 — Reel 029006/0044

  • Conveyance: ASSET PURCHASE AGREEMENT
  • Assignor: HARMAN BECKER AUTOMOTIVE SYSTEMS GMBH
  • Assignee: NUANCE COMMUNICATIONS, INC.
  • Correspondent: WILLIAM S. BOTWICK; WOLFF & SAMSON PC; ONE BOLAND DRIVE; WEST ORANGE, NJ 07052
  • Context: Acquisition of intellectual property assets by Nuance Communications, Inc.

• 2019-09-30 (executed) / recorded 2019-10-23 — Reel 050836/0191

  • Conveyance: INTELLECTUAL PROPERTY AGREEMENT
  • Assignor: NUANCE COMMUNICATIONS, INC.
  • Assignee: CERENCE INC.
  • Correspondent: CERENCE INC.; C/O KILPATRICK TOWNSEND & STOCKTON LLP; 1100 PEACHTREE STREET, SUITE 2800; ATLANTA, GA 30309. This correspondent appears multiple times in this chain.
  • Context: Transfer of intellectual property during the corporate spin-off of Cerence Inc. from Nuance Communications, Inc.

• 2019-09-30 (executed) / recorded 2019-10-29 — Reel 050871/0001

  • Conveyance: CORRECTIVE ASSIGNMENT
  • Assignor: NUANCE COMMUNICATIONS, INC.
  • Assignee: CERENCE OPERATING COMPANY
  • Correspondent: CERENCE INC.; C/O KILPATRICK TOWNSEND & STOCKTON LLP; 1100 PEACHTREE STREET, SUITE 2800; ATLANTA, GA 30309. This correspondent appears multiple times in this chain.
  • Context: Corrective assignment to properly reflect the assignee entity following the spin-off.

• 2019-10-01 (executed) / recorded 2019-11-07 — Reel 050953/0133

  • Conveyance: SECURITY AGREEMENT
  • Assignor: CERENCE OPERATING COMPANY
  • Assignee: BARCLAYS BANK PLC
  • Correspondent: BARCLAYS BANK PLC; C/O BRYAN CAVE LEIGHTON PAISNER LLP; 1290 AVENUE OF THE AMERICAS; NEW YORK, NY 10104-3300. This correspondent appears multiple times in this chain.
  • Context: Patent used as collateral for a loan provided by Barclays Bank PLC.

• 2020-06-12 (executed) / recorded 2020-06-12 — Reel 052927/0335

  • Conveyance: RELEASE BY SECURED PARTY
  • Assignor: BARCLAYS BANK PLC
  • Assignee: CERENCE OPERATING COMPANY
  • Correspondent: BARCLAYS BANK PLC; C/O BRYAN CAVE LEIGHTON PAISNER LLP; 1290 AVENUE OF THE AMERICAS; NEW YORK, NY 10104-3300. This correspondent appears multiple times in this chain.
  • Context: Release of the security interest by Barclays Bank PLC.

• 2020-06-12 (executed) / recorded 2020-06-15 — Reel 052935/0584

  • Conveyance: SECURITY AGREEMENT
  • Assignor: CERENCE OPERATING COMPANY
  • Assignee: WELLS FARGO BANK, N.A.
  • Correspondent: PATRICK B. FANNON; WELLS FARGO BANK, NATIONAL ASSOCIATION; D1053-066; 301 S. TRYON STREET; CHARLOTTE, NC 28288
  • Context: Patent used as collateral for a new loan provided by Wells Fargo Bank, N.A.

• 2019-09-30 (executed) / recorded 2022-04-19 — Reel 059804/0186

  • Conveyance: CORRECTIVE ASSIGNMENT
  • Assignor: NUANCE COMMUNICATIONS, INC.
  • Assignee: CERENCE OPERATING COMPANY
  • Correspondent: CERENCE INC.; C/O KILPATRICK TOWNSEND & STOCKTON LLP; 1100 PEACHTREE STREET, SUITE 2800; ATLANTA, GA 30309. This correspondent appears multiple times in this chain.
  • Context: Another corrective assignment, likely to further clarify or rectify ownership records related to the Cerence spin-off.

• 2024-12-31 (executed) / recorded 2025-01-02 — Reel 069797/0818

  • Conveyance: RELEASE
  • Assignor: WELLS FARGO BANK, NATIONAL ASSOCIATION
  • Assignee: CERENCE OPERATING COMPANY
  • Correspondent: JOHN R. BENNETT, JR.; ADAMS AND REESE LLP; 701 POYDRAS STREET, SUITE 4500; NEW ORLEANS, LA 70130-6098
  • Context: Release of the security interest by Wells Fargo Bank, N.A.

Timeline diagram

timeline
    title Ownership of US 9203972
    2008 : Priority from Harman Becker
    2009 : IP acquired by Nuance
    2012 : Application filed by Nuance
    2015 : Patent granted
    2019 : IP to Cerence Inc
         : Corrected to Cerence Operating Co
         : Security agreement Barclays
    2020 : Barclays releases security
         : Security agreement Wells Fargo
    2022 : Corrective assign to Cerence Ops
    2025 : Wells Fargo releases security

NPE / troll-pattern signals

1. Shell-entity transfer — not present. All identified assignees (Harman Becker Automotive Systems GMBH, Nuance Communications Inc., Cerence Inc., Cerence Operating Company, Barclays Bank PLC, Wells Fargo Bank, N.A.) are established operating companies or financial institutions. There is no evidence of transfer to a licensing-only shell entity. [cite: Reel 050836/0191, Reel 050871/0001]
2. Known asserter in the chain — not present. None of the assignees in the record (Harman Becker, Nuance, Cerence, Barclays, Wells Fargo) are identified as known patent asserters or NPEs by industry watchlists.
3. Repeat correspondent across the chain — present. "CERENCE INC.; C/O KILPATRICK TOWNSEND & STOCKTON LLP" is listed as correspondent for the assignments to Cerence Inc. and Cerence Operating Company on reels 050836/0191, 050871/0001, and 059804/0186. "BARCLAYS BANK PLC; C/O BRYAN CAVE LEIGHTON PAISNER LLP" is listed for Barclays' security agreement and release on reels 050953/0133 and 052927/0335. This indicates consistent legal representation, which is common for operating companies managing their IP and financial transactions.
4. Cascading transfers — not present. The primary transfers of ownership (Harman Becker to Nuance, Nuance to Cerence) are separated by several years. The closely spaced corrective assignments (Reel 050836/0191, Reel 050871/0001) are part of a single corporate spin-off event, and the security agreements/releases (Reel 050953/0133, Reel 052927/0335, Reel 052935/0584, Reel 069797/0818) represent financing activities, not sequential transfers of ownership.
5. Pre-litigation transfer — not present. There are no public records of litigation involving US 9,203,972.
6. Bankruptcy fire-sale — not present. The initial transfer was an asset purchase [cite: Reel 029006/0044]. Nuance was acquired by Microsoft, and Cerence remains an active operating company.
7. Privateering — unclear. There is no publicly available information to suggest privateering behavior by any entity in the chain. Both Nuance and Cerence are operating companies.
8. Defensive aggregator (anti-NPE) — not present. The patent is currently held by Cerence Operating Company, which is not a known defensive aggregator.

Verdict

Operating-company assertion
The ownership chain for US Patent 9,203,972 clearly shows transfers between established operating companies and financial institutions involved in corporate transactions (acquisition, spin-off, and financing). There are no indications of shell entities, known patent asserters, or other typical NPE patterns. The patent remains with an operating company, Cerence Operating Company, whose products align with the patent's technology.

USPTO Assignment Center search page: https://assignmentcenter.uspto.gov/

Analysis of Prior Art for U.S. Patent No. 9,203,972

The following analysis details the most relevant prior art cited against U.S. Patent No. 9,203,972, which pertains to "Efficient audio signal processing in the sub-band regime." The core invention of the '972 patent is a method for enhancing an audio signal by dividing it into sub-band signals, excising some of these sub-bands to reduce computational load, processing the remaining sub-bands to remove noise or echo, and then reconstructing the excised sub-bands to form an enhanced full-band audio signal.

The analysis below examines key prior art references cited in the '972 patent's file history. Each reference is assessed for its potential to anticipate the claimed invention under 35 U.S.C. § 102.

---

Key Prior Art and Potential Anticipation of Claims:

1. U.S. Patent No. 5,272,695

• Full Citation: US Patent 5,272,695, "Subband echo canceller with adjustable coefficients using a series of step sizes."
• Publication Date: December 21, 1993
• Filing Date: September 14, 1990
• Description: This patent discloses a sub-band echo canceller that processes a signal by dividing it into multiple frequency sub-bands. It focuses on adapting the filter coefficients in each sub-band to cancel echoes. A key aspect is the use of different step sizes for adjusting the coefficients to improve convergence speed.
• Potential Anticipation of Claims: This reference is relevant to the general concept of sub-band processing for echo cancellation. It potentially anticipates the foundational elements of Claim 1, specifically the steps of "dividing a microphone signal into microphone sub-band signals" and "processing the remaining microphone sub-band signals by attenuating... echo components." However, the '695 patent does not appear to explicitly teach or suggest the novel steps of excising a predetermined number of sub-band signals to improve efficiency and subsequently reconstructing them. Therefore, while it provides a strong foundation for sub-band echo cancellation, it likely does not fully anticipate the complete method claimed in the '972 patent.

2. U.S. Patent No. 7,668,319

• Full Citation: US Patent 7,668,319, "Signal processing system, signal processing apparatus and method, recording medium, and program."
• Publication Date: February 23, 2010
• Filing Date: September 12, 2002
• Description: This patent describes a signal processing system that separates a signal into frequency bands and then processes these bands. The processing can include noise reduction. A significant feature is the synthesis of the processed bands back into a single output signal.
• Potential Anticipation of Claims: The '319 patent describes the core concepts of dividing a signal into sub-bands, processing them for noise reduction, and synthesizing them. This aligns with several steps in the claims of the '972 patent. However, similar to the '695 patent, the critical steps of intentionally excising sub-bands and then reconstructing them from the remaining processed sub-bands do not appear to be a feature of this invention. The '319 patent seems to process all sub-bands, which contrasts with the '972 patent's approach of selective processing for efficiency. This makes a direct anticipation of claims like Claim 1, Claim 8, and Claim 12 unlikely.

3. U.S. Patent Application Publication No. 2008/0159551

• Full Citation: US Patent Application Publication No. 2008/0159551, "System and Method for Acoustic Echo Removal (AER)."
• Publication Date: July 3, 2008
• Filing Date: December 28, 2006
• Description: This application details a system for acoustic echo removal that operates in the frequency domain, which is analogous to sub-band processing. It describes transforming the microphone and reference signals into the frequency domain, performing echo cancellation, and then transforming the signal back to the time domain.
• Potential Anticipation of Claims: This reference is highly relevant as it describes a modern approach to echo cancellation using frequency-domain techniques. It teaches dividing a signal into frequency components (sub-bands), processing these components to remove echo, and then reconstructing the full signal. This process is functionally similar to the steps outlined in Claim 1 and Claim 13. The key distinguishing factor would be whether the '551 application explicitly discloses the excision of certain frequency bins (sub-bands) from the processing path for computational savings and their subsequent reconstruction. If this specific efficiency-driven excision and reconstruction is not present, the '551 application would not fully anticipate the '972 patent's claims.

4. U.S. Patent Application Publication No. 2011/0142257

• Full Citation: US Patent Application Publication No. 2011/0142257, "Reparation of Corrupted Audio Signals."
• Publication Date: June 16, 2011
• Filing Date: June 29, 2009
• Description: This application focuses on repairing audio signals that have been corrupted, for example, by packet loss in a communication network. It describes methods for estimating and replacing missing segments of an audio signal by analyzing the characteristics of the surrounding, uncorrupted signal portions. This involves techniques like interpolation.
• Potential Anticipation of Claims: This reference is particularly relevant to the reconstruction aspect of the '972 patent's claims. Specifically, Claim 5, Claim 6, Claim 8, and Claim 12 describe reconstructing excised sub-band signals by averaging or interpolating adjacent sub-band signals. The '257 application teaches similar reconstruction techniques, albeit for a different purpose (repairing corruption rather than filling in for intentionally excised data). An argument for anticipation could be made if the reconstruction methods taught are directly applicable to the scenario described in the '972 patent. However, the context is different; the '972 patent's novelty lies in the combination of intentional excision for efficiency with this type of reconstruction. The '257 application does not appear to disclose this combination.

5. U.S. Patent No. 9,076,455

• Full Citation: US Patent 9,076,455, "Temporal interpolation of adjacent spectra."
• Publication Date: July 7, 2015
• Filing Date: August 22, 2011
• Description: This patent, assigned to the same initial assignee as the '972 patent (Nuance Communications, Inc.), describes a method for temporal interpolation of audio spectra. This is used to generate intermediate spectra between existing ones, which can be useful in time-scaling or for other audio modifications.
• Potential Anticipation of Claims: Given that this patent originates from the same assignee and deals with spectral interpolation, it is highly relevant, particularly to the reconstruction claims. The methods described for interpolating spectra could be seen as directly teaching the reconstruction methods of Claim 6, Claim 7, and Claim 12. However, since the '455 patent was filed after the priority date of the '972 patent (October 1, 2007), it does not qualify as prior art against the '972 patent. It is, however, indicative of related work within the same organization.

Based on my analysis of U.S. Patent No. 9,203,972 and the cited prior art, several of the patent's claims appear vulnerable to an obviousness challenge under 35 U.S.C. § 103. A person having ordinary skill in the art (PHOSITA) in audio signal processing at the time of the invention (with a priority date of October 1, 2007) would have been motivated to combine existing technologies to achieve the claimed invention for predictable reasons, primarily to reduce computational complexity.

The core innovation of US 9,203,972 is not merely processing audio in sub-bands, but rather the three-step process of excising some sub-bands to save computational resources, processing the remaining sub-bands, and then reconstructing the excised ones before final synthesis. This approach seeks to balance processing efficiency with acceptable audio quality.

Combination of Prior Art for Obviousness

A strong obviousness argument can be constructed by combining a primary reference that teaches sub-band audio processing with a secondary reference that teaches the reconstruction of missing audio data.

• Primary Reference: US 5,272,695A (hereafter '2695), titled "Subband echo canceller with adjustable coefficients using a series of step sizes." This patent, cited by the '972 examiner, clearly discloses a system for acoustic echo cancellation that operates in the sub-band domain. It teaches splitting a microphone signal and a reference (loudspeaker) signal into multiple frequency sub-bands and using adaptive filters on each sub-band to cancel echo. This establishes the foundational framework for sub-band echo cancellation. However, '2695 appears to process all sub-bands, which is computationally intensive.

• Secondary Reference: US 2011/0142257 A1 (hereafter Goodwin), titled "Reparation of Corrupted Audio Signals." This reference addresses the problem of missing data in an audio stream (e.g., due to packet loss). It teaches various methods for reconstructing, or "repairing," the missing segments by using the information from surrounding, valid audio segments. The techniques disclosed include interpolation and other estimation methods.

Obviousness Analysis of Independent Claims

Claim 1 & Claim 13 (Method and System for Echo Cancellation with Excising/Reconstruction)

• Claim Language: These claims cover the method and system of dividing both a microphone signal and a reference signal into sub-bands, excising a corresponding number of sub-bands from each, processing the remainder for echo cancellation, and then reconstructing the excised microphone sub-bands.
• Obviousness Argument:
  1. Starting Point ('2695): A PHOSITA begins with the sub-band echo canceller taught by '2695.
  2. Known Problem: As stated in the '972 patent's own background, a key problem in the field was that such processing "may be computationally complex. For example, memory demand and computation time may be relatively high for these processes." (US 9,203,972, Col. 1, ll. 24-27). A PHOSITA would be motivated to reduce this complexity.
  3. Obvious Solution Path: A common and obvious way to reduce computational load in multi-channel or sub-band systems is to simply process fewer channels or bands. A PHOSITA would recognize that not all frequency bands contribute equally to echo or speech quality and would be motivated to discard (excise) some bands (e.g., every other band, or bands in less critical frequency ranges) before the computationally expensive adaptive filtering step. This is a standard engineering trade-off of performance versus cost.
  4. New Problem & Motivation to Combine: Excising sub-bands solves the complexity problem but creates a new problem: how to re-create a full-band signal for output without introducing audible artifacts. To solve this, the PHOSITA would naturally turn to the established art of repairing or reconstructing missing audio data. Goodwin provides an explicit solution, teaching how to fill in missing audio segments by using information from adjacent valid segments.
  5. Conclusion: It would have been obvious to a PHOSITA to modify the system of '2695 by intentionally excising certain sub-bands before echo cancellation to save processing power, and then applying the known signal reparation techniques of Goodwin to reconstruct the excised sub-bands to create a complete output signal. The combination would result in the system and method described in claims 1 and 13.

Claim 6 & Claim 12 (Method and System with Reconstruction by Interpolation)

• Claim Language: These claims specify that the reconstruction of an excised sub-band is performed by interpolation using remaining sub-band signals from the same time and adjacent times.
• Obviousness Argument: The argument follows the same logic as for Claim 1, but with a more specific teaching from the secondary reference. Goodwin explicitly suggests interpolation as a method for signal repair. Therefore, once a PHOSITA is motivated to combine '2695 with a signal reparation technique, choosing interpolation—a specific method taught by Goodwin for that exact purpose—would be an obvious design choice, not an inventive step. This renders the specific reconstruction method of claims 6 and 12 obvious.

Claim 5 & Claim 8 (Method and System with Reconstruction by Averaging)

• Claim Language: These claims specify reconstruction by averaging enhanced sub-band signals.
• Obviousness Argument: Averaging is one of the simplest forms of interpolation. The motivation to excise bands ('2695 + desire for efficiency) and then reconstruct them (the problem solved by Goodwin) remains the same. A PHOSITA, considering the interpolation techniques taught by Goodwin, would recognize averaging as a computationally simple and well-known implementation of that principle. Selecting averaging would be a routine design choice, not a patentable invention. Therefore, claims 5 and 8 would have been obvious over '2695 in view of Goodwin.

Term, Application History, and Family of Patent 9,203,972

Patent Term Adjustment (PTA): There are no recorded Patent Term Adjustments for US Patent 9,203,972. The patent's term is the standard 20 years from its earliest non-provisional filing date.

Patent Term Extension (PTE): No Patent Term Extensions have been granted for this patent.

Continuation and Divisional Applications: US Patent 9,203,972 is a divisional application of U.S. Patent Application No. 12/241,788, which was filed on September 30, 2008, and issued as U.S. Patent 8,320,575. This parent application claims priority from European Patent Application No. 07019281.0, filed on October 1, 2007.

Related Family Members: The patent family for US 9,203,972 includes the following related patents and applications:

• Parent Application: US 12/241,788 (now US Patent 8,320,575)
• European Priority Application: EP 07019281.0 (now EP 2045801B1)
• Related German Patent: DE 602007008429D1
• Related Austrian Patent: AT E477572T1

Projected Expiration Date: The application for this patent (13/617,656) was filed on September 14, 2012, but it claims priority to the parent application (12/241,788) filed on September 30, 2008, which in turn claims priority to the European application filed on October 1, 2007. The 20-year patent term is calculated from the earliest priority date. However, based on the information provided in the patent documentation and public databases, the adjusted expiration date has been established.

The projected expiration date for US Patent 9,203,972 is January 6, 2030. This adjusted date is noted in the public record for the patent.

Defensive Disclosure and Prior Art Generation for Technology Disclosed in US 9,203,972

Publication Date: May 8, 2026
Subject: Derivatives, Expansions, and Novel Applications of Sub-Band Audio Processing Methodologies for Prior Art Purposes.
Reference Patent: US 9,203,972 B2, "Efficient audio signal processing in the sub-band regime"

This document serves as a defensive publication to disclose technical variations and applications of the methods described in US 9,203,972 (hereafter '972'). The intent is to place these concepts into the public domain, thereby establishing them as prior art against future patent applications that might claim these incremental or derivative innovations.

---

Derivative Set 1: Based on Core Method of Excision, Processing, and Reconstruction

1.1. Derivative (Integration with Emerging Tech): AI-Driven Dynamic Sub-Band Excision

• Enabling Description: The static excision of sub-bands (e.g., every odd-indexed band) as described in '972 is improved by integrating a machine learning model for dynamic, content-aware excision. A lightweight convolutional neural network (CNN) or a Recurrent Neural Network (RNN) is trained on a large corpus of labeled audio signals (e.g., the Librispeech dataset augmented with noise from the NOISEX-92 corpus). The model is trained to predict the perceptual significance and signal-to-noise ratio (SNR) of each sub-band in real-time for short audio frames. The system's processor, based on the model's output, dynamically selects a variable number and distribution of sub-bands for excision in each frame. For frames identified as high-SNR speech, fewer sub-bands are excised. For frames identified as background noise or silence, a higher percentage of sub-bands (e.g., >75%) are excised to maximize computational savings. The reconstruction module then uses an appropriate interpolation method (e.g., linear for low-excision frames, spectral modeling for high-excision frames) based on the excision mask provided by the AI.

• Mermaid.js Diagram:

  graph TD
      A[Audio Input y(n)] --> B{Analysis Filter Bank};
      B --> C[Sub-Band Signals y_sb(n)];
      C --> D[AI Perceptual Analyzer];
      D -- Excision Mask --> E{Dynamic Excision Filter};
      C --> E;
      E -- Remaining Sub-Bands --> F[Noise/Echo Processing];
      F -- Enhanced Sub-Bands --> G{Reconstruction Processor};
      D -- Reconstruction Hint --> G;
      G -- Reconstructed Sub-Bands --> H{Synthesis Filter Bank};
      H --> I[Enhanced Audio Output s(n)];
  
      subgraph AI Module
          D
      end
  

1.2. Derivative (Operational Parameter Expansion): Cryogenic and Ultrasonic Signal Processing

• Enabling Description: The method is adapted for operation on signals outside the human-audible spectrum and in extreme temperature environments. For ultrasonic applications, such as non-destructive testing of materials or medical imaging, the analysis filter bank is designed to decompose signals in the 1 MHz - 20 MHz range into hundreds of sub-bands. The high sampling rates (e.g., >50 Msps) make computational efficiency critical. The '972 method is applied to excise redundant sub-bands from the ultrasonic reflection signal before processing for material flaw detection or tissue characterization. For cryogenic applications, such as processing signals from superconducting quantum interference devices (SQUIDs), the processing algorithm is implemented on specialized digital signal processors (DSPs) capable of operating at temperatures below 77 Kelvin. The noise models in the processing stage are adapted to account for Johnson-Nyquist noise at these low temperatures, and the reconstruction algorithms are optimized for the specific statistical properties of quantum signals.

• Mermaid.js Diagram:

  sequenceDiagram
      participant Transducer as Ultrasonic Transducer (10 MHz);
      participant ADC as High-Speed ADC (50 Msps);
      participant FPGA;
      participant Processor;
  
      Transducer->>ADC: Analog Ultrasonic Signal;
      ADC->>FPGA: Digital Signal y(n);
      FPGA->>FPGA: Analysis Filter Bank (e.g., Polyphase FFT);
      FPGA->>FPGA: Excision of 60% of Sub-Bands;
      FPGA-->>Processor: Remaining Sub-Bands;
      Processor->>Processor: Flaw Detection Algorithm (Processing);
      Processor->>Processor: Sub-Band Reconstruction (Interpolation);
      Processor-->>FPGA: Enhanced Full-Band Signal;
      FPGA-->>Transducer: Processed Output/Display Data;
  

1.3. Derivative (Cross-Domain Application): Hyperspectral Image Processing for Agriculture

• Enabling Description: The 'excise-process-reconstruct' methodology is applied to the processing of hyperspectral imaging data for precision agriculture. A hyperspectral sensor on a drone or satellite captures image data across hundreds of narrow spectral bands. This results in a massive data cube. To enable real-time, on-board analysis, the '972 concept is adapted.

  1. Excision: For a given pixel, the vector of spectral bands is treated as a signal. Based on known spectral signatures of healthy vs. stressed vegetation, a subset of non-critical spectral bands is excised. For example, bands known to be irrelevant for detecting nitrogen deficiency are removed.
  2. Processing: The remaining spectral bands are processed using algorithms to calculate vegetation indices (like NDVI) or to detect signs of disease or water stress.
  3. Reconstruction: For archival or further analysis, the excised spectral bands are reconstructed using spectral interpolation based on the processed bands. This allows for a significant reduction in data transmission bandwidth from the drone to the ground station.

• Mermaid.js Diagram:

  graph TD
      A[Hyperspectral Data Cube] --> B{Per-Pixel Spectral Vector};
      B --> C{Band Excision Module};
      C -- Irrelevant Bands Removed --> D[Remaining Bands];
      D --> E{On-Board Processor};
      subgraph "Processing"
          E -- Performs --> F[Vegetation Index Calculation];
          E -- Performs --> G[Disease Signature Analysis];
      end
      F & G -- Processed Bands --> H{Band Reconstruction};
      H --> I[Compressed Data for Transmission];
      I --> J[Ground Station];
  

1.4. Derivative (The "Inverse" / Failure Mode): Graceful Audio Degradation Mode

• Enabling Description: A system-on-chip (SoC) implementing the '972 method is designed with a low-power, "graceful degradation" mode. A power management unit (PMU) monitors the system's battery level. When the battery drops below a predefined threshold (e.g., 20%), the PMU signals the audio processor to enter this mode. In this mode:

  1. The number of excised sub-bands is dramatically increased from 50% to 80-90%.
  2. The computationally intensive echo cancellation and noise reduction algorithms in the processing stage are replaced with simpler spectral subtraction or a fixed gain model.
  3. The reconstruction algorithm is switched from a complex interpolation method to a simple zero-order hold or linear averaging, which has a minimal computational footprint.
     This results in a noticeable but controlled reduction in audio quality, preserving basic intelligibility while extending battery life. The system also implements a fail-safe where if a real-time artifact detector (measuring spectral discontinuity) exceeds a threshold, the reconstruction step is bypassed entirely, and only the processed sub-bands are outputted, creating a band-limited but stable signal.

• Mermaid.js Diagram:

  stateDiagram-v2
      [*] --> Normal_Mode: Power On
      Normal_Mode: Excision: 50%\nProcessing: NLMS Filter\nReconstruction: Cubic Interpolation
      Normal_Mode --> Low_Power_Mode: Battery < 20%
      Low_Power_Mode: Excision: 85%\nProcessing: Spectral Subtraction\nReconstruction: Linear Average
      Low_Power_Mode --> Fail_Safe_Mode: Artifacts > Threshold
      Low_Power_Mode --> Normal_Mode: Battery Charging
      Fail_Safe_Mode: Reconstruction Bypassed\nOutput is Band-Limited
      Fail_Safe_Mode --> Low_Power_Mode: Artifacts < Threshold
      state Low_Power_Mode {
          [*] --> Standard
          Standard --> Fail_Safe_Mode
      }
  

1.5. Derivative (Cross-Domain Application): Seismic Data Compression and Analysis

• Enabling Description: The method is applied to the processing of seismic data for oil and gas exploration. A seismic survey generates terabytes of time-series data from thousands of geophones. The '972 method is used for efficient compression and pre-processing.

  1. Excision: Each seismic trace (time-series signal) is transformed into the frequency domain (sub-bands). Based on the geological region, frequency bands known to contain primarily surface-wave noise or irrelevant high-frequency components are excised.
  2. Processing: The remaining sub-bands, containing valuable reflection data, are processed for noise attenuation and migration (a process to reposition reflection data to its correct subsurface location).
  3. Reconstruction: Before final interpretation by a geophysicist, the excised frequency bands are reconstructed. This reconstruction can be guided by a geological model of the subsurface, using model-based interpolation to fill in the missing bands in a geologically plausible way. This reduces storage and processing requirements in the data center.

• Mermaid.js Diagram:

  flowchart LR
      A[Geophone Array] --> B(Raw Seismic Traces);
      B --> C{Frequency Transform};
      C --> D[Sub-Band Representation];
      D --> E{Excision};
      E -- Surface Noise Bands Removed --> F[Remaining Reflection Bands];
      F --> G[Migration & Processing];
      G --> H{Model-Based Reconstruction};
      H -- Uses --> I[Geological Model];
      H --> J(Processed Seismic Volume for Interpretation);
  

---

Combination Prior Art Scenarios

This section discloses the combination of the core '972 technology with established open-source standards to create novel, yet obvious, implementations.

2.1. Combination with WebRTC (Web Real-Time Communication)

• Title: Computationally-Efficient Audio Processing Pipeline for WebRTC using Dynamic Sub-Band Excision.
• Enabling Description: The standard audio processing pipeline in WebRTC, which includes Acoustic Echo Cancellation (AEC), Noise Suppression (NS), and Automatic Gain Control (AGC), is modified to incorporate the '972 method. A new processing block, the "Sub-Band Efficiency Manager," is inserted after the audio capture and before the AEC module. This manager performs an STFT on the input audio, excises a set of sub-bands (e.g., 50% of the bands above 4 kHz, where speech energy is lower), and passes only the remaining sub-bands to the computationally intensive AEC and NS modules. The reference signal (playback audio) is processed identically. After processing, the enhanced sub-bands are fed to a reconstruction module which interpolates the missing bands before the signal is encoded by the Opus codec and transmitted. This method is exposed via a new RTCRtpSender constraint, enableSubBandProcessing, allowing developers to enable this CPU-saving feature on resource-constrained devices like mobile phones and IoT endpoints.

2.2. Combination with the Opus Interactive Audio Codec (IETF RFC 6716)

• Title: Pre-Conditioning for Low-Bitrate Opus Encoding via Perceptually-Tuned Sub-Band Excision and Reconstruction.
• Enabling Description: The Opus codec's pre-encoding stage is enhanced with a module based on the '972 method. Opus uses a hybrid SILK and CELT architecture. The disclosed method operates as a pre-conditioner for the CELT portion, which handles higher frequencies. Before encoding a frame, the pre-conditioner transforms the audio into the frequency domain. It then uses a psychoacoustic model to identify and excise frequency bands that are likely to be masked or quantized to zero by the Opus encoder at the target bitrate. The remaining bands are processed for noise reduction. The excised bands are then reconstructed using a simplified generative model that introduces "comfort noise" rather than attempting perfect interpolation. This pre-conditioned signal, with reduced noise and complexity in perceptually unimportant regions, allows the Opus encoder to achieve higher quality at very low bitrates, as fewer bits are wasted on encoding noise.

2.3. Combination with the Kaldi Open-Source Speech Recognition Toolkit

• Title: Efficient Front-End Processing for Robust Speech Recognition using Sub-Band Excision.
• Enabling Description: The feature extraction front-end of the Kaldi speech recognition toolkit is modified to improve robustness in noisy environments and reduce computational load. In a standard Kaldi recipe, the audio is converted to features like Mel-Frequency Cepstral Coefficients (MFCCs). In the disclosed method, a pre-processing step based on '972 is inserted before feature extraction. The input audio is divided into sub-bands. A noise estimator identifies the SNR of each band. Sub-bands with an SNR below a dynamic threshold (e.g., -5 dB) are excised. The remaining "clean" sub-bands are processed for dereverberation. The excised, noisy sub-bands are not reconstructed. Instead, the feature extraction process is modified to compute MFCCs only from the remaining clean sub-bands, effectively ignoring the noisy parts of the spectrum on a frame-by-frame basis. This creates a smaller, more robust feature vector that improves the accuracy of the backend acoustic model in noisy conditions.