# US 12,770,634 B2 · Neural-inspired audio signal processor > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,770,634 B2 - Document ID: US12770634B2 - Status: Granted - Kind: B2 - Application number: 18/705,774 - Filed: 2022-10-28 - Published: 2025-02-06 - Granted: 2026-10-20 - Earliest claimed: 2021-11-05 - Assignees: Ucl Ventures Limited - Applicants: Ucl Ventures Limited - Inventors: Ifat Yasin; Vit Drga - Web page: https://patentgrep.com/patent/US12770634B2 ## CPC - G10L 21/0208: Noise filtering - A61B 5/372: Analysis of electroencephalograms - G10L 2021/02087: {the noise being separate speech, e.g. cocktail party} - G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING - A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D) ## Abstract An audio-signal processor (100) for filtering an audio signal-of-interest from an input audio signal (111) comprising a mixture of the signal-of-interest and background noise, the processor (100) comprising a frontend unit (120), the frontend unit (120) comprising a filterbank (121) comprised of an array of bandpass filters, a sound level estimator (123), and a memory (122) with one or more input-output, I/O, functions stored on said memory (122); wherein the frontend unit (120) is configured to receive: an unfiltered input audio signal (111), and human-derived neural-inspired feedback signals, NIFS, wherein the frontend unit (120) is further configured to: extract sound level estimates from an output of the one or more bandpass filters using the sound level estimator (123), modify the input-output, I/O, functions in response to the received sound level estimates and the NIFS, and determine an enhanced I/O functions, and store the enhanced I/O functions on the memory (122) and use the enhanced I/O functions to determine one or more modified filterbank parameters in response to the received NIFS and sound level estimates, apply the modified filterbank parameters either across one filter in the filterbank (121) or across a range of filters in the filterbank (121) within the frontend unit (120), and output a filtered audio signal (112) to a sound feature onset detector (130) for further processing. ## Claims ### Claim 1 An audio-signal processor for filtering an audio signal-of-interest from an input audio signal comprising a mixture of the audio signal-of-interest and background noise, the processor comprising a frontend unit, the frontend unit comprising a filterbank comprised of an array of bandpass filters, a sound level estimator, and a memory with one or more input-output, I/O, functions stored on said memory; wherein the frontend unit is configured to receive: an unfiltered input audio signal, and human-derived neural-inspired feedback signals, NIFS, wherein the frontend unit is further configured to: extract sound level estimates from an output of the one or more bandpass filters using the sound level estimator, modify the input-output, I/O, functions in response to the received sound level estimates and the NIFS, and determine enhanced I/O functions, and store the enhanced I/O functions on the memory and use the enhanced I/O functions to determine one or more modified filterbank parameters in response to the received NIFS and sound level estimates, wherein the one or more modified filterbank parameters include: i) a modified gain value and ii) a modified compression value for a given input audio signal, apply the modified filterbank parameters either across one filter of the filterbank or across a range of filters of the filterbank within the frontend unit, wherein the frontend unit is further configured to: apply the modified gain value and the modified compression value to the unfiltered input audio signal by way of modifying the input or parameters of a given filter or range of filters of the filterbank to determine a filtered output audio signal, and output the filtered audio signal to a sound feature onset detector for further processing. ### Claim 2 (depends on 1) The audio-signal processor of claim 1, wherein the processor further comprises: a Higher-Level Auditory Information, HLAI, processing module, comprising an internal memory, the HLAI processing module configured to receive human-derived brain-processing information and/or measurements derived by psychophysical/physiological assessments, and store it on its internal memory and, using said brain-processing information, the HLAI is further configured to simulate aspects of the following, which constitute aspects of the NIFS: brainstem-mediated, BrM, neural feedback information and cortical-mediated, CrM, neural feedback information. ### Claim 3 (depends on 2) The audio-signal processor of claim 2, wherein the HLAI processing module is further configured to derive the human-derived NIFS using said simulated and/or direct BrM and/or CrM neural feedback information and relay said NIFS to the frontend unit. ### Claim 4 (depends on 2) The audio-signal processor of claim 2, wherein the HLAI processing module is configured to receive the brain-processing information by direct means or by indirect means from higher-levels of auditory processing areas of a human brain, and wherein the brain-processing information is derived from any one of, or a combination of, the following: psychophysical data, physiological data, electrophysiological data, or electroencephalographic (EEG) data. ### Claim 5 (depends on 2) The audio-signal processor of claim 2, wherein the human-derived brain-processing information further comprises a range of time constants which define a build-up and a decay of gain with time derived from human-derived measurements; and wherein HLAI processing module is further configured to modify the human-derived NIFS using said time constants in response to the received brain-processing information and relay said NIFS to the frontend unit. ### Claim 6 (depends on 5) The audio-signal processor of claim 5, wherein the range of time constants comprise human-derived onset time build-up constants, τon, applied to the I/O functions stored in the frontend unit to modify the I/O functions and derive the enhanced I/O functions stored in the frontend unit to modify a rate of increase of the gain value, effects of which are subsequently applied to the filter or filters of the filterbank. ### Claim 7 (depends on 5) The audio-signal processor of claim 5, wherein the range of time constants comprise human-derived offset time decay constants, τoff, applied to the I/O functions stored in the frontend unit to modify the I/O functions and derive the enhanced I/O functions stored in the frontend unit to modify a rate of decrease of the gain value, effects of which are subsequently applied to the filter or filters of the filterbank. ### Claim 8 (depends on 2) The audio-signal processor of claim 2, further comprising a sound feature onset detector configured to receive the filtered output audio signal from the frontend unit and detect sound feature onsets, and wherein the sound feature onset detector is further configured to relay the sound feature onsets to the HLAI processing module, and the HLAI processing module is configured to store said sound feature onsets on its internal memory for determining the NIFS. ### Claim 9 (depends on 8) The audio-signal processor of claim 8, wherein the sound feature onset detector is further configured to relay the filtered output audio signal to the HLAI processing module and the HLAI processing module configured to store the filtered output audio signal on its internal memory. ### Claim 10 (depends on 2) The audio-signal processor of claim 2, further comprising a signal-to-noise ratio, SNR, estimator module configured to receive the filtered output audio signal from the frontend unit and determine a SNR of the mixture of the signal-of-interest and the background noise, and wherein the SNR estimator module is further configured to relay the SNR to the HLAI processing module, and the HLAI processing module is configured to store said SNR on its memory for determining the NIFS. ### Claim 11 (depends on 10) The audio-signal processor of claim 10, further comprising a machine learning unit comprising a decision device in data communication with the HLAI processing module, the decision device comprising an internal memory, the decision device is configured to receive data from the HLAI processing module and store it on its internal memory, wherein the decision device is configured to process the data and output a speech-enhanced filtered output audio signal. ### Claim 12 (depends on 11) The audio-signal processor of claim 11, further comprising a feature extraction, FE, module, said FE module is further configured to perform feature extraction on the filtered output audio signal, and the FE module is further configured to relay the extracted features to the machine learning unit, and the decision device is configured to store the extracted features in its internal memory. ### Claim 13 (depends on 12) The audio-signal processor of claim 12, wherein the SNR estimator module is configured to relay the filtered output audio signal to the FE module, the FE module is configured to relay the filtered output audio signal to the machine learning unit, and the decision device is configured to store the filtered output audio signal in its internal memory. ### Claim 14 (depends on 11) The audio-signal processor of claim 11, wherein the decision device is configured to process: the data received from the HLAI processing module, which includes the SNR values, the extracted features, sound feature onsets and attentional oscillations data, and outputs a speech-enhanced filtered output audio signal. ### Claim 15 (depends on 14) The audio-signal processor of claim 14, wherein the machine learning unit further comprises a machine learning algorithm stored on its internal memory, and wherein the decision device applies an output of the algorithm to the data received from the HLAI processing module, including the SNR values, the extracted features, and derives neural-inspired feedback parameters. ### Claim 16 (depends on 15) The audio-signal processor of claim 15, wherein the derived neural-inspired feedback parameters are relayed, from the decision device, to the HLAI processing module and the HLAI processing module is configured to store said neural-inspired feedback parameters on its memory for determining the NIFS. ### Claim 17 (depends on 1) The audio-signal processor of claim 1, wherein the modified gain values are a continuum of gain values, derived from human data that are in the following range: 10 to 60 dB. ### Claim 18 (depends on 1) The audio-signal processor of claim 1, wherein the modified compression values are a continuum of compression values that are in the following range: 0.1 to 1.0. ### Claim 19 (depends on 18) The audio-signal processor of claim 18, wherein the continuum of compression values are derived from human datasets and are dependent on input sound aspects, wherein the input sound aspects comprise a sound level and temporal characteristics which define the compression applied. ### Claim 20 (depends on 1) The audio-signal processor of claim 1, wherein the filter or filters of the filterbank within the frontend unit is further configured so as to modify a bandwidth of each of the one or more bandpass filters. ### Claim 21 (depends on 1) The audio-signal processor of claim 1, wherein the modified gain and compression values are: either applied to the input audio signal per bandpass filter in the one or more bandpass filters, or are applied to the input audio signal across some or all bandpass filters in the one or more bandpass filters. ### Claim 22 A method of filtering an audio signal-of-interest from an input audio signal comprising a mixture of the audio signal-of-interest and background noise, the method performed by a processor comprising a frontend unit, the frontend unit comprising a filterbank, the filterbank comprising one or more bandpass filters, a sound level estimator, and a memory with an input-output, I/O, functions stored on said memory, wherein the frontend unit is configured to perform the following method steps: i) receiving: an unfiltered input audio signal, human-derived neural-inspired feedback signals, NIFS, ii) extracting sound level estimates from an output of the one or more bandpass filters using the sound level estimator, iii) modifying the input-output, I/O, functions in response to the received sound level estimates and the NIFS, iv) determining an enhanced I/O function, v) storing the enhanced I/O function on said memory, vi) using the enhanced I/O function to determine one or more modified parameters to apply to the filter/filters of the filterbank in response to the received NIFS, wherein the one or more modified filterbank parameters include: i) a modified gain value and ii) a modified compression value for a given input audio signal, vii) applying the modified filterbank parameters either across one filter of the filterbank or across a range of filters of the filterbank within the frontend unit, wherein the frontend unit is further configured to: apply the modified gain value and the modified compression value to the unfiltered input audio signal by way of modifying the input or parameters of a given filter or range of filters of the filterbank to determine a filtered output audio signal, and viii) outputting the filtered audio signal to a sound feature onset detector for further processing.