CN-122027949-A - Howling suppression method and device and related products

Abstract

The embodiment of the invention provides a howling suppression method, a howling suppression device and related products, wherein the method comprises the steps of obtaining a first signal and a second signal, obtaining signal parameters of the first signal and the second signal, wherein the signal parameters comprise coherence data corresponding to N frequency band areas of the first signal and the second signal and energy data corresponding to the N frequency band areas of the first signal, the N frequency band areas are obtained by dividing a target frequency band according to the order from low frequency to high frequency, and if the first signal is determined to have no effective voice data based on the signal parameters, suppression gains are respectively set for the first signal in the N frequency band areas, the howling can be suppressed, and meanwhile audio information of other frequency band areas can be reserved.

Inventors

• CHENG ZISHENG
• HUANG RONGJUN

Assignees

• 珠海市杰理科技股份有限公司

Dates

Publication Date

20260512

Application Date

20251229

Claims (20)

1. 1.A howling suppression method, characterized by comprising: acquiring a first signal and a second signal, wherein the first signal is an audio signal acquired by a microphone, and the second signal is an audio signal output by audio playing equipment; The signal parameters of the first signal and the second signal are obtained, wherein the signal parameters comprise coherence data corresponding to the first signal and the second signal in N frequency band areas and energy data corresponding to the first signal in N frequency band areas, and the N frequency band areas are obtained by dividing a target frequency band from low frequency to high frequency; And if the first signal is determined to have no effective voice data based on the signal parameters, respectively setting suppression gains for the first signal in the N frequency band areas based on the signal parameters so as to respectively suppress howling in different frequency band areas.
2. 2. The howling suppression method according to claim 1, wherein said setting suppression gains for the first signal in the N frequency band regions, respectively, based on the signal parameters, comprises: judging whether the signal parameters simultaneously meet a voice energy threshold and a voice characteristic, wherein the voice energy threshold is a lower limit value of voice energy generated by a user when voice output is performed, and the voice characteristic is a signal characteristic that effective voice data exist in the first signal; If yes, determining that effective voice data exists in the first signal, and not carrying out howling suppression on the first signal; if not, determining that the first signal does not have valid voice data.
3. 3. A howling suppression method according to claim 2, characterized in that, The energy data comprises a target frequency band energy average value, wherein the target frequency band energy average value is the average energy of a first signal in the target frequency band; The judgment of meeting the voice energy threshold comprises judging whether the target frequency band energy mean value is larger than the voice energy threshold or not, and if yes, meeting the voice energy threshold.
4. 4. A howling suppressing method as claimed in claim 3, characterized in that, The coherence data comprises at least one of a coherence mean value, a maximum coherence mean value and a coherence maximum region of each frequency band region, wherein the maximum coherence mean value is the maximum value of the coherence mean values of the N frequency band regions, and the coherence maximum region is the frequency band region where the maximum coherence mean value is located; the energy data comprises at least one of a signal energy mean value of each frequency band region, a region maximum energy mean value and an energy maximum region, wherein the region maximum energy mean value is the maximum value of the signal energy mean values of the N frequency band regions, and the energy maximum region is the frequency band region where the region maximum energy mean value is located.
5. 5. The howling suppression method according to claim 4, characterized in that, The speech features include at least one of: the method comprises the steps that a coherence maximum area is not located in a first frequency band area; the maximum coherence mean value is smaller than a first coherence threshold value; The energy maximum area is positioned in the first frequency band area; And fourthly, the energy of the N frequency band regions gradually decays from low frequency to high frequency.
6. 6. The howling suppression method according to claim 5, characterized in that, The energy data also comprises target area number and a last target area sequence number, wherein the target area number is the number of areas with the signal energy mean value larger than a first energy threshold value, and the last target area sequence number is the sequence number of the last frequency band area with the signal energy mean value larger than the first energy threshold value; the energy of the N frequency band areas gradually decays from low frequency to high frequency, and the energy comprises the number of the target areas which is larger than or equal to the sequence number of the last target area.
7. 7. The howling suppression method according to claim 4, wherein the setting of suppression gains for the first signal in the N frequency band regions based on the signal parameters, respectively, comprises: If the energy maximum area is not located in the first frequency band area and the first signal energy average value is larger than the second energy threshold value, determining the energy maximum area as a howling area, setting the suppression gain of the howling area to be 0 so that the audio output of the howling area is in a mute state, wherein the first signal energy average value is the average energy of the first signal in the whole frequency domain range, and the second energy threshold value is the signal energy lower limit value of the howling phenomenon of the first signal.
8. 8. The howling suppression method according to claim 4, wherein the setting of suppression gains for the first signal in the N frequency band regions based on the signal parameters, respectively, comprises: determining the dynamic mapping smoothness of the coherence of the first signal and the second signal based on average coherence, a preset maximum mapping smoothness parameter and a preset minimum mapping smoothness parameter, wherein the average coherence is the average value of the coherence of the N frequency band areas; based on the dynamic mapping center position of the coherence and the dynamic mapping smoothness, performing coherence mapping on the first coherence of a kth frequency band region to obtain second coherence of the kth frequency band region after mapping; and determining the suppression gain of the kth frequency band region based on a preset dynamic gain smoothing value and a dynamic gain mapping center value, the second coherence and the energy data of the kth frequency band region.
9. 9. The howling suppression method according to claim 8, characterized in that, The average coherence, the preset maximum mapping smoothness parameter and the minimum mapping smoothness parameter comprise: ; In the formula, Representing the smoothness of the dynamic map, And Respectively representing a preset maximum mapping smoothness parameter and a minimum mapping smoothness parameter, Represents average coherence; The coherence-based dynamic mapping center position and the dynamic mapping smoothness perform coherence mapping on the first coherence of the kth frequency band region to obtain a second coherence of the kth frequency band region after mapping, and the method comprises the following steps: In the formula, Representing a second coherence of the k-th band region, A first coherence representing a region of a kth frequency band, A dynamic mapping center position representing coherence; The determining the suppression gain of the kth frequency band region based on the preset dynamic gain smoothing value and the dynamic gain mapping center value, the second coherence, and the energy data of the kth frequency band region includes: In the formula, And Respectively representing a preset dynamic gain smoothing value and a dynamic gain mapping center value, Energy data representing a region of a kth frequency band, Representing the suppression gain in the k-th band region.
10. 10. A howling suppressing apparatus, comprising: the signal acquisition unit is configured to acquire a first signal and a second signal, wherein the first signal is an audio signal acquired by a microphone, and the second signal is an audio signal output by audio playing equipment; the data acquisition unit is configured to acquire signal parameters of the first signal and the second signal, wherein the signal parameters comprise coherence data corresponding to N frequency band areas of the first signal and the second signal and energy data corresponding to N frequency band areas of the first signal, and the N frequency band areas are obtained by dividing a target frequency band from a low frequency to a high frequency; And the gain setting unit is configured to set suppression gains for the first signal in the N frequency band areas respectively based on the signal parameters if the first signal is determined to have no effective voice data based on the signal parameters, so as to suppress howling in different frequency band areas respectively.
11. 11. Howling suppression apparatus as claimed in claim 10, characterized in that the apparatus further comprises: The voice judging unit is configured to judge whether the signal parameters simultaneously meet a voice energy threshold and a voice characteristic, wherein the voice energy threshold is a lower limit value of voice energy generated by a user when voice output is carried out, the voice characteristic is a signal characteristic that effective voice data exist in the first signal, if yes, the first signal is determined to have the effective voice data, howling suppression is not carried out on the first signal, and if not, the first signal is determined to have no effective voice data.
12. 12. Howling suppressing apparatus as claimed in claim 11, characterized in that, The energy data comprises a target frequency band energy average value, wherein the target frequency band energy average value is the average energy of a first signal in the target frequency band; the voice judging unit is further configured to judge whether the target frequency band energy average value is larger than a voice energy threshold value, and if yes, the voice energy threshold value is met.
13. 13. Howling suppressing apparatus as claimed in claim 12, characterized in that, The coherence data comprises at least one of a coherence mean value, a maximum coherence mean value and a coherence maximum region of each frequency band region, wherein the maximum coherence mean value is the maximum value of the coherence mean values of the N frequency band regions, and the coherence maximum region is the frequency band region where the maximum coherence mean value is located; the energy data comprises at least one of a signal energy mean value of each frequency band region, a region maximum energy mean value and an energy maximum region, wherein the region maximum energy mean value is the maximum value of the signal energy mean values of the N frequency band regions, and the energy maximum region is the frequency band region where the region maximum energy mean value is located.
14. 14. Howling suppressing apparatus as claimed in claim 13, characterized in that, The speech features include at least one of: the method comprises the steps that a coherence maximum area is not located in a first frequency band area; the maximum coherence mean value is smaller than a first coherence threshold value; The energy maximum area is positioned in the first frequency band area; And fourthly, the energy of the N frequency band regions gradually decays from low frequency to high frequency.
15. 15. Howling suppressing apparatus as claimed in claim 14, characterized in that, The energy data also comprises target area number and a last target area sequence number, wherein the target area number is the number of areas with the signal energy mean value larger than a first energy threshold value, and the last target area sequence number is the sequence number of the last frequency band area with the signal energy mean value larger than the first energy threshold value; the energy of the N frequency band areas gradually decays from low frequency to high frequency, and the energy comprises the number of the target areas which is larger than or equal to the sequence number of the last target area.
16. 16. The howling suppression method according to claim 13, wherein the gain setting unit is further configured to: If the energy maximum area is not located in the first frequency band area and the energy average value of the first signal is larger than a second energy threshold value, determining the energy maximum area as a howling area, setting the suppression gain of the howling area to be 0 so that the audio output of the howling area is in a mute state, wherein the first signal energy average value is the average energy of the first signal in the whole frequency domain range, and the second energy threshold value is the signal energy lower limit value of the howling phenomenon of the first signal.
17. 17. The howling suppression method according to claim 13, wherein the gain setting unit is further configured to: determining the dynamic mapping smoothness of the coherence of the first signal and the second signal based on average coherence, a preset maximum mapping smoothness parameter and a preset minimum mapping smoothness parameter, wherein the average coherence is the average value of the coherence of the N frequency band areas; based on the dynamic mapping center position of the coherence and the dynamic mapping smoothness, performing coherence mapping on the first coherence of a kth frequency band region to obtain second coherence of the kth frequency band region after mapping; and determining the suppression gain of the kth frequency band region based on the preset dynamic gain smoothing value and the dynamic gain mapping center value, the second coherence and the energy data of the kth frequency band region.
18. 18. Microphone, characterized by a howling suppression of the first signal by a method according to any of the claims 1-9, or by a howling suppression device according to any of the claims 10-17.
19. 19. A computer readable storage medium having stored thereon a computer program, which when executed by a processor, implements a howling suppression method as claimed in any one of claims 1 to 9.
20. 20. An electronic device comprising a storage medium storing a computer program, wherein the computer program, when executed by a processor, implements a howling suppression method as claimed in any one of claims 1 to 9.

Description

Howling suppression method and device and related products Technical Field The present invention relates to the technical field of bluetooth communications, and in particular, to a howling suppression method, a howling suppression device, a microphone, a computer-readable storage medium, and an electronic apparatus. Background With the popularization of intelligent outdoor karaoke sound equipment, users can enjoy high-quality music anytime and anywhere and can enjoy the music of impulse karaoke according to preference, and the free karaoke form also brings new problems, such as the phenomenon of howling of the sound equipment when the users need to hold a microphone to cut songs or adjust sound effects close to the sound equipment. In order to solve this problem, in one prior art, an interrupt button is provided on the microphone, and when a user needs to approach the sound, pressing the button can temporarily interrupt the pickup of the microphone, and disconnect the loop signal of the microphone, so as to avoid the howling phenomenon generated by the microphone. For the scheme of setting the interrupt button, the user is required to actively operate, and in practical application, the user is difficult to ensure that the user can operate consciously when approaching the sound box each time. Disclosure of Invention Based on the above-mentioned current situation, a main object of the present invention is to provide a method, an apparatus and a related product for suppressing howling, which are capable of suppressing the howling frequency band region and simultaneously retaining audio information of other frequency band regions by performing frequency domain division on a target frequency band and setting different suppression gains for different frequency band regions. In order to achieve the above purpose, the technical scheme adopted by the invention is as follows: the first aspect of the embodiment of the invention provides a device Acquiring a first signal and a second signal, wherein the first signal is an audio signal acquired by a microphone, and the second signal is an audio signal output by audio playing equipment; The signal parameters of the first signal and the second signal are obtained, wherein the signal parameters comprise coherence data corresponding to the first signal and the second signal in N frequency band areas and energy data corresponding to the first signal in N frequency band areas, and the N frequency band areas are obtained by dividing a target frequency band from low frequency to high frequency; And if the first signal is determined to have no effective voice data based on the signal parameters, respectively setting suppression gains for the first signal in the N frequency band areas based on the signal parameters so as to respectively suppress howling in different frequency band areas. Preferably, before the setting of the suppression gains for the first signal in the N frequency band regions based on the signal parameters, the method includes: Judging whether the signal parameters simultaneously meet a voice energy threshold and a voice characteristic, wherein the voice energy threshold is a lower limit value of voice energy generated by a user when voice output is carried out, the voice characteristic is a signal characteristic that effective voice data exist in the first signal, if so, determining that the effective voice data exist in the first signal, not carrying out howling suppression on the first signal, and if not, determining that the effective voice data do not exist in the first signal. Preferably, the energy data comprises a target frequency band energy average value, the target frequency band energy average value is the average energy of the first signal in the target frequency band, and the judgment of meeting the voice energy threshold comprises judging whether the target frequency band energy average value is larger than the voice energy threshold or not, and if yes, meeting the voice energy threshold. The coherence data comprises at least one of a coherence mean value, a maximum coherence mean value and a coherence maximum region of each frequency band region, wherein the maximum coherence mean value is the maximum value of the coherence mean values of the N frequency band regions, the coherence maximum region is the frequency band region where the maximum coherence mean value is located, the energy data comprises at least one of a signal energy mean value, a region maximum energy mean value and an energy maximum region of each frequency band region, the region maximum energy mean value is the maximum value of the signal energy mean values of the N frequency band regions, and the energy maximum region is the frequency band region where the region maximum energy mean value is located; Preferably, the voice features comprise at least one of a first feature, a second feature, an energy maximum region and a fourth feature, wherein the coherence maximum region is