RU2011143091A - ACOUSTIC ECHO SUPPRESSION DEVICE AND FRONT CONFERENCE COMMUNICATION DEVICE - Google Patents

Claims (15)

1. The acoustic echo canceller (210), including input interface means (230), providing the extraction of the downmix signal - downmix signal (310) from the input signal (300), which, in general, consists of a downmix signal (310) and service parametric information (320), collectively representing a multi-channel signal containing at least additional channels or more channels than is contained in the downmix signal; a computer (220) designed to calculate the transmittance (350) of the adaptive filter (240) based on the downmix signal (310) received by the computer (220), the microphone signal (340), or the signal derived from the microphone signal (720); adaptive filter (240), designed to receive filter coefficients (350) from the calculator (220) and use them to filter the microphone signal (340) or a signal derived from the microphone signal (720) to suppress the echo excited by the multi-channel signal in the microphone signal (340). 2. The acoustic echo suppressor (210) according to claim 1, in which the calculator (220) is designed to calculate filtering coefficients (350) by determining the first indicator of the reference spectral power density based on the down-mixed signal (310), by determining the second indicator reference power spectral density based on a microphone signal (340) or a signal derived from a microphone signal (720) by determining the expected echo filtering coefficients based on the first and second reference power density by determining the expected echo based on the first reference power spectral density and the expected echo filter coefficients and by determining filter coefficients (350) based on the expected echo filtering coefficients and the second reference power spectral density. 3. The acoustic echo canceller (210) according to claim 1, in which the calculator (220) is designed to calculate the first reference power spectrum based on

or

, Where

is the first indicator of the reference power spectral density, a _i (k, m) is the weight coefficient, B _i (k, m) is the ith channel of the down-mixed signal (310), where N is the number of channels in the down-mixed signal (310 ), where N is greater than or equal to 1, where k is the time index of the block and m is the frequency coefficient. 4. The acoustic echo canceller (210) according to claim 1, in which the calculator (220) is designed to calculate the second reference power spectrum based on

or

Where

is the second reference power spectrum, c _i (k, m) is the weight coefficient, A _i (k, m) is the ith channel of the signal mixed with decreasing (720), where K is the number of channels in the signal mixed with decreasing (720) and K is greater than or equal to 1, where k is the time index of the block and m is the frequency coefficient. 5. The acoustic echo suppression device (210) according to claim 1, in which the calculator (220), in addition to the above, is designed to calculate the expected echo filtering coefficients and evaluate the echo signal based on the first reference indicator of the power spectral density in the delayed version some amount of delay. 6. The acoustic echo canceller (210) according to claim 5, in which the calculator (220) is also designed to determine the delay by finding the correlation value for the set of possible delay values, by finding the gain from the echo prediction for the set of possible delay values and by finding the value of the set of possible delay values as a delay value with a maximum value among the established values of the gain from the echo prediction. 7. The acoustic echo suppression device (210) according to claim 1, in which the calculator (220) provides for the possibility of calculating the first modified power spectrum based on the first reference indicator of the power spectral density by subtracting the average value of the first reference power spectral density; in which the calculator (220) provides for the possibility of calculating the second modified power spectrum based on the second reference indicator of the power spectral density by subtracting the second average value of the second reference power spectral density; and as a part of which the calculator (220) provides for the possibility of calculating the expected echo filtering coefficients based on the first and second modified parameters of the power spectral density. 8. A method of suppressing an acoustic echo signal, including: extracting a downmix signal (310) from an input signal (300) consisting of a downmix signal (310) and overhead parametric information (320), where the downmix signal (310) and service parameter information (320) collectively represent a multi-channel signal containing at least additional channels or a number of channels greater than the number of channels in the downmix signal; calculating a filter transmittance (350) for adaptive filtering based on a down-mixed signal and a microphone signal or a signal derived from a microphone signal; adaptive filtering of the microphone signal (340) or a signal derived from the microphone signal (720), based on the filtering coefficients, in order to suppress the echo excited by the multi-channel signal in the microphone signal (340). 9. The method according to claim 8, further comprising decoding the down-mixed signal (310) and the service parameter information (320) with conversion to a plurality of speaker signals (330). 10. The front-end conference device (terminal) (200), comprising: an acoustic echo canceller (210) according to any one of claims 1 to 7; multi-channel decoder (250); at least one microphone unit (110), in which a multi-channel decoder (250) is designed to decode a down-mixed signal (310) and service parametric information (320) with conversion into a plurality of speaker signals (330); in which at least one microphone unit (110) provides a microphone signal (340). 11. The conference terminal (200) according to claim 10, in which the input interface means (230) is intended for further extraction of service parametric information (320); comprising a multi-channel decoder (250) includes a boost mixer (705) and a parametric processor (480), where the parametric processor (480) is configured to receive service parametric information (320) from the input interface (230) and with the possibility of generation the upmix control signal (707), and where the upmixer (705) is configured to receive the downmix signal (310) from the input interface means (230) and receive the upmix control signal from the parametric one processor and configured to generate the plurality of speaker signals (330) based on the downmix signal (310) and a control signal up-mixer (707). 12. The conference terminal (200) according to claim 10, comprising, in addition to the specified one, a multi-channel encoder (400), designed to encode multiple input audio signals (340; 410) with conversion to a new down-mixed signal (720) and new overhead parametric information (730), which together represent a plurality of input audio signals, of which a microphone signal (340) from at least one microphone unit (110) is contained in a plurality of input audio signals; also incorporating an acoustic echo canceller (210), configured to receive a new down-mixed signal (720) as a signal derived from a microphone signal. 13. The conference communication terminal (200) according to claim 10, comprising a plurality of microphone units (110) providing a plurality of input audio signals (330; 410). 14. A method for generating a plurality of speaker signals (330) and microphone signals (340), including: a method for suppressing (210) an acoustic echo according to claim 8; a multi-channel decoding step (250); a microphone signal receiving step (340), wherein in the multi-channel decoding step (250), the down-mixed signal (310) and the service parametric information (320) are decoded to obtain a plurality of speaker signals (330). 15. A computer program for implementing the method according to any one of claim 8 or 14, provided that it is executed using a processor.