# US 12,119,013 B2 · Acoustic crosstalk suppression device and acoustic crosstalk suppression method > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,119,013 B2 - Document ID: US12119013B2 - Status: Granted - Kind: B2 - Application number: 17/778,277 - Filed: 2020-11-16 - Published: 2022-12-29 - Granted: 2024-10-15 - Earliest claimed: 2019-11-21 - Assignees: Panasonic Intellectual Property Management Co., Ltd. - Applicants: Panasonic Intellectual Property Management Co., Ltd. - Inventors: Masanari Miyamoto; Naoya Tanaka; Hiromasa Ohashi - Web page: https://patentgrep.com/patent/US12119013B2 ## CPC - G10L 21/0208: Noise filtering - G10L 15/20: Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech (G10L21/02 takes precedence) - H04R 3/005: {for combining the signals of two or more microphones (specially adapted for hearing aids H04R25/407)} - H04R 3/02: for preventing acoustic reaction {, i.e. acoustic oscillatory feedback (specially adapted for hearing aids H04R25/453)} - G10L 15/22: Procedures used during a speech recognition process, e.g. man-machine dialogue - G10L 2021/02087: {the noise being separate speech, e.g. cocktail party} - G10L 2021/02166: {Microphone arrays; Beamforming} - H04R 1/406: {microphones} - H04R 2499/13: Acoustic transducers and sound field adaptation in vehicles - G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING - H04R: LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; ELECTRIC HEARING AIDS; PUBLIC ADDRESS SYSTEMS (producing sounds with frequency not determined by supply frequency G10K) ## Abstract An acoustic crosstalk suppression device includes a speaker estimation unit configured to estimate a main speaker based on voice signals collected by n units of microphones corresponding to n number of persons (n: an integer equal to or larger than 3); n units of filter update units each of which is configured to update a parameter of a filter configured to generate a suppression signal of a crosstalk component included in a voice signal of the main speaker; and a crosstalk suppression unit configured to suppress the crosstalk component by using a synthesis suppression signal generated by the maximum (n-1) units of filter update units corresponding to reference signals collected by the maximum (n-1) units of microphones. ## Claims ### Claim 1 An acoustic crosstalk suppression device, comprising: a speaker estimation unit configured to estimate a main speaker in a closed space based on voice signals collected by n units of microphones arranged in the closed space corresponding to n number of persons, n being an integer equal to or larger than 3; n units of filter update units each of which comprises a filter configured to generate a suppression signal of a crosstalk component caused by utterance of another speaker and included in a voice signal of the main speaker, and each of which is configured to update a parameter of the filter for suppressing the crosstalk component to store an update result of the parameter of the filter therein; a crosstalk suppression unit configured to suppress the crosstalk component included in the voice signal of the main speaker by using a synthesis suppression signal based on first suppression signals generated by a maximum of n-1 units of the filter update units corresponding to reference signals collected by a maximum of n-1 units of the microphones other than a microphone corresponding to the main speaker; a suppression amount estimation unit configured to estimate a suppression amount of the crosstalk component; and a reference signal acquisition unit configured to replace a first combination of the reference signals used to generate the synthesis suppression signal with a second combination of other reference signals in a case where the suppression amount is equal to or smaller than a first threshold, wherein the filter update units corresponding to the other reference signals after replacement regenerate second suppression signals based on an input of the other reference signals. ### Claim 2 (depends on 1) The acoustic crosstalk suppression device according to claim 1, further comprising: a synthesis unit configured to synthesize the first suppression signals generated by the maximum of the n-1 units of the filter update units. ### Claim 3 (depends on 1) The acoustic crosstalk suppression device according to claim 1, wherein the filter provided in each of the maximum of the n-1 units of the filter update units generates one of the first suppression signals by using a latest stored parameter. ### Claim 4 (depends on 1) The acoustic crosstalk suppression device according to claim 1, wherein a filter update unit corresponding to a reference signal collected by a microphone corresponding to the another speaker updates the parameter of the filter provided in the filter update unit by using the reference signal, and stores the update result of the parameter of the filter. ### Claim 5 (depends on 1) The acoustic crosstalk suppression device according to claim 1, further comprising: a main signal acquisition unit configured to replace with, as the voice signal of the main speaker, a second voice signal collected by another microphone disposed in a vicinity of the main speaker among the n units of the microphones in the case where the suppression amount is equal to or smaller than the first threshold. ### Claim 6 (depends on 1) The acoustic crosstalk suppression device according to claim 1, further comprising: a voice recognition unit configured to perform voice recognition on an utterance content of the main speaker based on the voice signal of the main speaker in which the crosstalk component is suppressed; and a main signal acquisition unit configured to switch to, as the voice signal of the main speaker, another voice signal collected by another microphone disposed in a vicinity of the main speaker among the n units of the microphones in a case where the voice recognition fails. ### Claim 7 (depends on 1) The acoustic crosstalk suppression device according to claim 1, further comprising: a voice recognition unit configured to perform voice recognition on an utterance content of the main speaker based on the voice signal of the main speaker in which the crosstalk component is suppressed, wherein the reference signal acquisition unit is further configured to replace the first combination of the reference signals used to generate the synthesis suppression signal with a third combination of the other reference signals in a case where the voice recognition fails. ### Claim 8 (depends on 1) The acoustic crosstalk suppression device according to claim 1, wherein the reference signal acquisition unit selects a combination of a maximum number of the reference signals collected by the maximum of the n-1 units of the microphones as the second combination of the other reference signals used to generate the synthesis suppression signal, and then reduces a number of selected combinations of the other reference signals to replace the second combination of the other reference signals with a third combination of the other reference signals. ### Claim 9 An acoustic crosstalk suppression method executed by an acoustic crosstalk suppression device, the acoustic crosstalk suppression method comprising: estimating a main speaker in a closed space based on voice signals collected by n units of microphones arranged in the closed space corresponding to n number of persons, n being an integer equal to or larger than 3; updating, by each of n units of filter update units, a parameter of a filter that generates a suppression signal of a crosstalk component caused by utterance of another speaker and included in a voice signal of the main speaker, and storing, by each of the n units of the filter update units, an update result of the parameter of the filter for suppressing the crosstalk component therein, suppressing the crosstalk component included in the voice signal of the main speaker by using a synthesis suppression signal based on first suppression signals generated by a maximum of n-1 units of the filter update units corresponding to reference signals collected by a maximum of n-1 units of the microphones other than a microphone corresponding to the main speaker; estimating a suppression amount of the crosstalk component; and replacing a first combination of the reference signals used to generate the synthesis suppression signal with a second combination of other reference signals in a case where the suppression amount is equal to or smaller than a first threshold, wherein the filter update units corresponding to the other reference signals after the replacing regenerate second suppression signals based on an input of the other reference signals. ### Claim 10 (depends on 7) The acoustic crosstalk suppression device according to claim 7, wherein the reference signal acquisition unit selects a combination of a maximum number of the reference signals collected by the maximum of the n-1 units of the microphones as the third combination of the other reference signals used to generate the synthesis suppression signal, and then reduces a number of selected combinations of the other reference signals to replace the third combination of the other reference signals with a fourth combination of the other reference signals. ### Claim 11 An acoustic crosstalk suppression device, comprising: a processor; and a memory including a program that, when executed by the processor, causes the processor to perform processing, the processing including: estimating a main speaker in a closed space based on voice signals collected by n units of microphones arranged in the closed space corresponding to n number of persons, n being an integer equal to or larger than 3; generating, via n units of filters, a suppression signal of a crosstalk component caused by utterance of another speaker and included in a voice signal of the main speaker, and updating, for each filter, a parameter of the filter for suppressing the crosstalk component to store an update result of the parameter of the filter; suppressing the crosstalk component included in the voice signal of the main speaker by using a synthesis suppression signal based on first suppression signals generated by a maximum of n-1 units of the filters corresponding to reference signals collected by a maximum of n-1 units of the microphones other than a microphone corresponding to the main speaker; estimating a suppression amount of the crosstalk component; and replacing a first combination of the reference signals used to generate the synthesis suppression signal with a second combination of other reference signals in a case where the suppression amount is equal to or smaller than a first threshold, wherein the filters corresponding to the other reference signals after replacement regenerate second suppression signals based on an input of the other reference signals.