CN109218917B - Automatic acoustic feedback monitoring and eliminating method in real-time communication system - Google Patents
Automatic acoustic feedback monitoring and eliminating method in real-time communication system Download PDFInfo
- Publication number
- CN109218917B CN109218917B CN201811340217.1A CN201811340217A CN109218917B CN 109218917 B CN109218917 B CN 109218917B CN 201811340217 A CN201811340217 A CN 201811340217A CN 109218917 B CN109218917 B CN 109218917B
- Authority
- CN
- China
- Prior art keywords
- howling
- curve
- suspected
- point
- judging
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
- H04R3/00—Circuits for transducers, loudspeakers or microphones
- H04R3/02—Circuits for transducers, loudspeakers or microphones for preventing acoustic reaction, i.e. acoustic oscillatory feedback
Abstract
The invention provides an automatic acoustic feedback monitoring and eliminating method in a real-time communication system, which is mainly characterized in that historical howling judgment information is initialized by acquiring audio data; carrying out time-frequency transformation on the received time domain signals to obtain corresponding frequency domain signals; determining a power value according to the frequency domain signal to generate a logarithmic amplitude spectrogram; overlapping the logarithmic amplitude spectrograms obtained by multiple sampling, dividing the power values into M power value intervals from large to small, sequencing the probabilities of the same power value intervals in all time periods appearing in different frequency bands, and forming a howling point suspected curve on the logarithmic amplitude spectrograms; according to different scenes, different howling point suspected curves and different judging methods are applied to judge the howling point of the current frame; generating a corresponding wave trap according to the judgment result of the current frame howling point; the method can flexibly use different suspected curves of the howling point as the judgment curves according to different use scenes, thereby meeting the requirements of more users.
Description
Technical Field
The invention belongs to the technical field of communication, and particularly relates to an automatic acoustic feedback monitoring and eliminating method in a real-time communication system.
Background
The acoustic feedback refers to a phenomenon that sound emitted from a loudspeaker is fed back to a microphone again through a feedback path and is picked up by the microphone, communication quality and subjective feeling are affected due to the existence of the acoustic feedback, even a sound or a power amplifier is burnt due to over-strong signals in serious cases, in order to inhibit the acoustic feedback, the peak power ratio and the power value of frequency domain data of audio frequency are detected, analyzed and calculated to judge a howling point in the prior art.
Disclosure of Invention
In order to solve the above technical problems, the present invention provides an automatic acoustic feedback monitoring and eliminating method in a real-time communication system.
The specific technical scheme of the invention is as follows:
an automatic acoustic feedback monitoring and eliminating method in a real-time communication system, characterized in that the method comprises the following steps:
s1: collecting audio data and initializing historical howling judgment information;
s2: carrying out time-frequency transformation on the received time domain signals to obtain corresponding frequency domain signals;
s3: determining a power value according to the frequency domain signal to generate a logarithmic amplitude spectrogram;
s4: overlapping logarithmic amplitude spectrograms obtained by multiple sampling, dividing power values into M power value intervals from large to small, calculating the probability P of the power value intervals M corresponding to N time periods in all the overlapped logarithmic amplitude spectrograms appearing in different frequency bands, sequencing the probabilities of the same power value intervals appearing in different frequency bands in all the time periods, and connecting the frequency bands corresponding to the highest probability in the same power value interval on the logarithmic amplitude spectrogram to form a howling point suspected curve;
s5: according to different scenes, different howling point suspected curves and different judging methods are applied to judge the howling point of the current frame;
s6: and generating a corresponding wave trap according to the judgment result of the current frame howling point.
According to the automatic acoustic feedback monitoring and eliminating method in the real-time communication system, the plurality of suspected howling point curves are generated, and different suspected howling point curves can be flexibly used as the judgment curves according to different use scenes, so that the requirements of more users are met, and the use experience of the users is improved.
Drawings
FIG. 1 is a flowchart of a method of automatic acoustic feedback detection and cancellation in a real-time communication system according to embodiment 1;
FIG. 2 is a schematic representation of the logarithmic amplitude spectrum of example 1;
FIG. 3 is a flowchart of a method of automatic acoustic feedback monitoring and cancellation in the real-time communication system of embodiment 2;
FIG. 4 is a flowchart of a method of automatic acoustic feedback monitoring and cancellation in the real-time communication system of embodiment 3;
FIG. 5 is a flowchart of a method of automatic acoustic feedback monitoring and cancellation in the real-time communication system of embodiment 4;
FIG. 6 is a flowchart of a method of automatic acoustic feedback monitoring and cancellation in the real-time communication system of embodiment 4;
FIG. 7 is a flowchart of a method of automatic acoustic feedback monitoring and cancellation in the real-time communication system of embodiment 5;
fig. 8 is a flowchart of a method of automatic acoustic feedback monitoring and cancellation in a real-time communication system according to embodiment 6.
Detailed Description
The invention is further described with reference to the following figures and examples, which are provided for the purpose of illustrating the general inventive concept and are not intended to limit the scope of the invention.
Example 1
As shown in fig. 1, there is provided an automatic acoustic feedback monitoring and eliminating method in a real-time communication system, the method comprising the steps of:
s1: collecting audio data and initializing historical howling judgment information;
s2: carrying out time-frequency transformation on the received time domain signals to obtain corresponding frequency domain signals;
s3: determining a power peak value according to the frequency domain signal, and generating a logarithmic amplitude spectrogram;
s4: overlapping logarithmic amplitude spectrograms obtained by multiple sampling, dividing power values into M power value intervals from large to small, calculating the probability P of the power value intervals M corresponding to N time periods in all the overlapped logarithmic amplitude spectrograms appearing in different frequency bands, sequencing the probabilities of the same power value intervals appearing in different frequency bands in all the time periods, and connecting the frequency bands corresponding to the highest probability in the same power value interval on the logarithmic amplitude spectrogram to form a howling point suspected curve;
s5: according to different scenes, different howling point suspected curves and different judging methods are applied to judge the howling point of the current frame;
s6: generating a corresponding wave trap according to the judgment result of the current frame howling point;
it should be noted that the logarithmic amplitude spectrogram is one of spectrograms, the amplitude of each spectral line in the logarithmic amplitude spectrogram is logarithmically calculated, and the unit of the ordinate is dB (decibel); the multiple sampling mentioned in the present invention does not limit the sampling method (including sampling interval time, sampling times, etc.), any sampling method that does not affect the realization of the technical effect of the present invention is within the protection scope of the present invention, for example, when the sampling rate is set as 2048 sampling points at 8KHZ, the length of the time period can be specifically set according to the need, the present invention is not specifically limited, for example, in fig. 2a-d, N is 0.1s, the generation of the wave limiter is also a conventional technique, as long as the method that can realize the technical effect of the present invention is within the protection scope of the present invention, for example, a wave limiter pool mechanism (i.e. a buffer ring of a wave limiter, which is obtained from the use of the wave limiter, and put back when not used), supports 20 wave limiters to be executed simultaneously, and supports cyclic use, i.e. when the same howling point appears again, the wave limiter in the wave limiter pool is directly used, and no new wave limiter is generated, the gain and bandwidth of the limiter are a function of the curve of its howling frequency.
As shown in fig. 2, when the number of sampling times is 4, fig. 2a, fig. 2b, fig. 2c, and fig. 2d respectively show logarithmic amplitude spectrograms obtained from the first to fourth sampling, taking a power value M1 and a time period of 0.5-0.6s as an example, three logarithmic amplitude spectrograms of which M1 at a time of 0.5-0.6s are located in 5-6HZ segments, and it is proved that the frequency segment with a frequency of 5-6 is the frequency segment with a highest probability of M1 at 0.5-0.6, similarly, the frequency segment with a frequency of 1-2 is the frequency segment with a highest probability of M1 at 0.7-0.8, the frequency segment with a frequency of 2-3 is the frequency segment with a highest probability of M1 at 0.9-1.0, the frequency segment with a frequency of 1-2 is the frequency segment with a highest probability of M1 at 1.1-1.2, and the maximum corresponding frequency segments are connected (the frequency value of the frequency value can be any point on the frequency segment, but the method of taking points over M time periods should be consistent), one can get fig. 2 e: a howling point suspected curve obtained with the power value of m 1;
according to the method, a plurality of suspected howling point curves can be obtained simultaneously, and when the method is used, a user can judge the howling points according to different scenes by taking the different suspected howling point curves as standards, so that the method for judging the howling points is more flexibly adjusted, the method is more reasonably applied to various scenes, and the call quality and the user experience are improved.
Example 2
As shown in fig. 3, step S4 further includes:
s41: judging the size of the number A with the highest ranking probability in the same power value interval on the logarithmic amplitude spectrogram, if A is 1, performing step S5, if A is more than 1, transmitting a selection request to a user side, and performing step S42;
s42: receiving and judging selection information fed back by a user side, ranking A pieces of probability with the same ranking from high to low in frequency, if the user side selects a high requirement, selecting the frequency with the highest ranking as a connection point of a howling point suspected curve, if the user selects a low requirement, selecting the frequency with the lowest ranking as a connection point of the howling point suspected curve, and if the user selects a medium requirement, selecting any frequency except the frequency with the highest ranking and the lowest ranking as a connection point of the howling point suspected curve;
as shown in fig. 2, taking a power value interval of m2 and a time period of 0.5-0.6 as an example, of four logarithmic amplitude spectrograms, m2 of two logarithmic amplitude spectrograms at time of 0.5-0.6s is located in a 3-4HZ segment, and m2 of the other two logarithmic amplitude spectrograms at time of 0.5-0.6s is located in a 1-2HZ segment, which proves that the frequency band with frequency of 5-6 and the frequency band with frequency of 1-2 have the same probability at 0.5-0.6, at this time, a selection request is sent to the client, and then one frequency is selected as a connection point of the howling point suspected curve according to selection information fed back by the client.
By the method, the connection point of the suspected howling point curve can be flexibly determined according to the requirements of the user, so that the method is more intelligent and further improves the user experience.
Example 3
As shown in fig. 4 and 5, step S5 includes:
s51: dividing howling point suspected curves on the logarithmic amplitude spectrogram into c1 and c2... cI from low to high, and setting howling elimination levels from high to low corresponding to different howling point suspected curves to require levels 1 and 2.. I;
s52: judging whether the user sends a howling request, if so, performing step S53, and if not, performing step S54;
s53: identifying a howling elimination requirement level requested by a user, and taking a howling suspected point curve corresponding to the howling elimination requirement level as a judgment curve;
s54: judging the howling elimination requirement of a use place, and taking a corresponding suspected curve of a howling point as a judgment curve;
s55: and finishing the current frame howling point judgment by applying the frequency bands corresponding to all the time on the judgment curve.
According to the method, one of the howling point suspected curves can be selected as a judgment curve based on the request of the user, the howling elimination requirement of a use place can be actively judged, and a proper howling point suspected curve is further selected as the judgment curve.
Example 4
As shown in fig. 6, step S54 includes the steps of:
s541: judging the number of people U, if U is larger than or equal to U1, taking a suspected howling point curve ci2 as a judgment curve, if U1 is larger than U2, performing step S542, if U is smaller than U2, taking a suspected howling point curve ci1 as a judgment curve, wherein ci1 is smaller than ci 2;
s542: acquiring volume, judging the volume W, if W is larger than or equal to W1, taking a suspected howling point curve ci3 as a judgment curve, if W1 is larger than or equal to W2, taking a suspected howling point curve ci4 as a judgment curve, and if W is smaller than W2, taking a suspected howling point curve ci5 as a judgment curve, wherein ci1 is lower than ci5, ci4 is lower than ci3, and ci2 is lower than ci 3;
the number of people can be judged by a sound channel sensor or a voiceprint sensor, and the volume can be judged by a volume sensor, and the invention is not particularly limited.
When the using scene is a particularly large conference room, the number of people is 120 and U1 is 100, the number of people is large at the moment, the requirement for eliminating howling is low, and therefore a high-height howling point suspected curve ci2 is directly adopted as a judgment curve; when the using scene is a very small conference room, the number of people is U8 < U2 < 10, the number of people is small at the moment, the howling elimination requirement is high, and therefore the low-height howling point suspected curve ci1 is directly used as a judgment curve; when the using scene is a conference room with a medium size, the number of people is U1-10-U-50-U2-10, the number of people is general, the volume of the using scene needs to be continuously judged, if the volume W is 100 decibel-W1-80 decibel, the howling elimination requirement is low, a suspected howling point curve ci3 is adopted as a judging curve, when the volume W1-80-W-60 is not less than W2-50, the howling requirement is general, a suspected howling point curve ci4 is adopted as a judging curve, and when the volume W is 30-2-50, the howling requirement is high, the suspected howling point curve ci5 is adopted as a judging curve.
The number of people and the volume of the using scene can be judged through the method by specifically limiting the step S54, and then the appropriate suspected curve of the howling point is adopted as the judgment curve, so that the judgment curve is automatically adjusted, and the method is more humanized.
Example 5
As shown in fig. 7, step S55 includes the steps of:
s551: monitoring audio data of a current frame in real time, performing time-frequency transformation on the audio data to obtain corresponding frequency domain signals, and determining a power value corresponding to each time period according to the frequency domain signals;
s552: setting the power value corresponding to each time period of the judgment curve as a power threshold, traversing all peak information of the current frame, and judging the howling point according to the power value corresponding to each time period and the power threshold;
s553: and updating the historical howling judgment information by using the current frame howling point judgment result.
The howling point of the current frame can be judged by performing time-frequency transformation on the current frame to obtain the corresponding power value and comparing the power value with the power threshold value.
Example 6
As shown in fig. 8, step S552 includes the steps of:
s5521: judging the power value corresponding to the time period of the current frame and the size of the power threshold, if the power value is larger than or equal to the power threshold, performing step S5522, and if the power value is smaller than the power threshold, judging that the frequency point is a non-howling point;
s5522: traversing and analyzing all peak value information, calculating and obtaining Q peak value power ratios Z, wherein Z is the ratio of the peak value power Z to the average power Zp, judging the size of the peak value power ratio Z, if Z is larger than or equal to a threshold value Z1, judging the frequency point as a howling point, and if Z is smaller than Z1, judging the frequency point as a non-howling point;
for example, (2-3) × 10 over a 0.5-0.6 time segment of the current frame3The power value on the frequency band is 5200, the power threshold is 5000, 5200 is larger than 5000, z is obtained by calculation to be 4, z1 is set to be 5, therefore z is smaller than z1, and the frequency point is determined to be a non-howling point.
By the method, the judgment method is more scientific and reasonable, the judgment accuracy is further improved, and the user experience is improved.
Claims (7)
1. A method for automatic acoustic feedback detection and cancellation in a real-time communication system, the method comprising the steps of:
s1: collecting audio data and initializing historical howling judgment information;
s2: carrying out time-frequency transformation on the received time domain signals to obtain corresponding frequency domain signals;
s3: determining a power value according to the frequency domain signal to generate a logarithmic amplitude spectrogram;
s4: overlapping logarithmic amplitude spectrograms obtained by multiple sampling, dividing power values into m power value intervals from large to small, calculating the probability P of the power value intervals m corresponding to N time periods in all the overlapped logarithmic amplitude spectrograms appearing in different frequency bands, sequencing the probabilities of the same power value intervals appearing in different frequency bands in all the time periods, and connecting the frequency bands corresponding to the highest probability in the same power value interval on the logarithmic amplitude spectrogram to form a howling point suspected curve;
s5: according to different scenes, different howling point suspected curves and different judging methods are applied to judge the howling point of the current frame;
s6: and generating a corresponding wave trap according to the judgment result of the current frame howling point.
2. The method for automatic acoustic feedback detection and cancellation in a real-time communication system according to claim 1, wherein the step S4 further comprises:
s41: judging the number A of the highest ranking probabilities in the same power value interval on the logarithmic amplitude spectrogram, if A =1, performing step S5, and if A > 1, sending a selection request to a user side, and performing step S42;
s42: receiving and judging selection information fed back by a user side, ranking A pieces of probability with the same ranking from high to low in frequency, if the user side selects a high requirement, selecting the frequency with the highest ranking as a connection point of the howling point suspected curve, if the user selects a low requirement, selecting the frequency with the lowest ranking as a connection point of the howling point suspected curve, and if the user selects a medium requirement, selecting any frequency except the frequency with the highest ranking and the lowest ranking as a connection point of the howling point suspected curve.
3. The method for automatic acoustic feedback detection and cancellation in a real-time communication system according to claim 1, wherein the step S5 includes:
s51: dividing howling point suspected curves on the logarithmic amplitude spectrogram into c1 and c2... cI from low to high, and setting howling elimination levels from high to low corresponding to different howling point suspected curves to require levels 1 and 2.. I;
s52: judging whether the user sends a howling request, if so, performing step S53, and if not, performing step S54;
s53: identifying a howling elimination requirement level requested by a user, and taking a howling suspected point curve corresponding to the howling elimination requirement level as a judgment curve;
s54: judging the howling elimination requirement of a use place, and taking a corresponding suspected curve of a howling point as a judgment curve;
s55: and finishing the current frame howling point judgment by applying the frequency bands corresponding to all the time on the judgment curve.
4. The method for automatic acoustic feedback detection and cancellation in a real-time communication system according to claim 2, wherein the step S54 comprises the steps of:
s541: judging the number of people U, if U is larger than or equal to U1, taking a suspected howling point curve ci2 as a judgment curve, if U1 is larger than U2, performing step S542, if U is smaller than U2, taking a suspected howling point curve ci1 as a judgment curve, wherein ci1 is smaller than ci 2;
s542: acquiring the volume, judging the volume W, if W is larger than or equal to W1, taking a howling point suspected curve ci3 as a judgment curve, if W1 is larger than or equal to W2, taking a howling point suspected curve ci4 as a judgment curve, and if W is smaller than W2, taking a howling point suspected curve ci5 as a judgment curve, wherein ci1 is lower than ci3 and lower than ci4 and lower than ci5 and lower than ci 2.
5. The method for automatic acoustic feedback detection and cancellation in a real-time communication system according to claim 2, wherein the step S55 comprises the steps of:
s551: monitoring audio data of a current frame in real time, performing time-frequency transformation on the audio data to obtain corresponding frequency domain signals, and determining a power value corresponding to each time period according to the frequency domain signals;
s552: setting the power value corresponding to each time period of the judgment curve as a power threshold, traversing all peak information of the current frame, and judging the howling point according to the power value corresponding to each time period and the power threshold;
s553: and updating the historical howling judgment information by using the current frame howling point judgment result.
6. The method for automatic acoustic feedback detection and cancellation in a real-time communication system according to claim 5, wherein the step S552 comprises the steps of:
s5521: judging the power value corresponding to the time period of the current frame and the size of the power threshold, if the power value is larger than or equal to the power threshold, performing step S5522 and step S5523, and if the power value is smaller than the power threshold, judging that the position of the sound source in the time period is a non-howling point;
s5522: traversing and analyzing all peak information, calculating and obtaining Q peak power ratios Z, wherein Z = the ratio of the peak power Z to the average power Zp, judging the magnitude of the peak power ratio Z, if Z is greater than or equal to a threshold value Z1, judging that the position of the sound source in the time period is a howling point, if Z is less than Z1, judging that the position of the sound source in the time period is a non-howling point,
s5523: judging the howling elimination requirement level i of an application place, traversing and analyzing all peak value information, calculating a peak power ratio and a peak value interval, and obtaining Q peak value intervals y, wherein if i is less than i1, the peak value interval y takes the peak value as the center, the left and right shift y1 points, if i1 is less than or equal to i2, the peak value interval y takes the peak value as the center, the left and right shift y2 points, if i is more than or equal to i2, the peak value interval y takes the peak value as the center, the left and right shift y3 points, wherein y1 is more than y2 and more than y 3.
7. The method for automatic acoustic feedback detection and cancellation in a real-time communication system of claim 1, wherein the step S6 is specifically performed by: judging the howling point in the current frame, immediately starting a wave limiter for filtering, continuing the filtering duration, and if the howling is confirmed to be finished, stopping filtering in advance.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811340217.1A CN109218917B (en) | 2018-11-12 | 2018-11-12 | Automatic acoustic feedback monitoring and eliminating method in real-time communication system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811340217.1A CN109218917B (en) | 2018-11-12 | 2018-11-12 | Automatic acoustic feedback monitoring and eliminating method in real-time communication system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109218917A CN109218917A (en) | 2019-01-15 |
| CN109218917B true CN109218917B (en) | 2020-07-03 |
Family
ID=64996287
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811340217.1A Active CN109218917B (en) | 2018-11-12 | 2018-11-12 | Automatic acoustic feedback monitoring and eliminating method in real-time communication system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109218917B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110213694B (en) * | 2019-04-16 | 2020-12-04 | 浙江大华技术股份有限公司 | Audio device and howling processing method thereof, and computer storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1398054A (en) * | 2001-07-16 | 2003-02-19 | 松下电器产业株式会社 | Whistler detection and suppresser thereof, its method and computer program products |
| CN102740189A (en) * | 2011-04-01 | 2012-10-17 | 中国科学院声学研究所 | Acoustic feedback inhibition method based on time reversal |
| CN103002392A (en) * | 2011-09-13 | 2013-03-27 | 联想(北京)有限公司 | Audio feedback processing method and intelligent device |
| CN103081511A (en) * | 2011-03-09 | 2013-05-01 | 松下电器产业株式会社 | Howling detection device, howling suppressing device and method of detecting howling |
| KR20150043858A (en) * | 2013-10-15 | 2015-04-23 | 한국전자통신연구원 | Apparatus and methdo for howling suppression |
| CN106373587A (en) * | 2016-08-31 | 2017-02-01 | 北京容联易通信息技术有限公司 | Automatic sound feedback detection and elimination method of real-time communication system |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070104335A1 (en) * | 2005-11-09 | 2007-05-10 | Gpe International Limited | Acoustic feedback suppression for audio amplification systems |
-
2018
- 2018-11-12 CN CN201811340217.1A patent/CN109218917B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1398054A (en) * | 2001-07-16 | 2003-02-19 | 松下电器产业株式会社 | Whistler detection and suppresser thereof, its method and computer program products |
| CN103081511A (en) * | 2011-03-09 | 2013-05-01 | 松下电器产业株式会社 | Howling detection device, howling suppressing device and method of detecting howling |
| CN102740189A (en) * | 2011-04-01 | 2012-10-17 | 中国科学院声学研究所 | Acoustic feedback inhibition method based on time reversal |
| CN103002392A (en) * | 2011-09-13 | 2013-03-27 | 联想(北京)有限公司 | Audio feedback processing method and intelligent device |
| KR20150043858A (en) * | 2013-10-15 | 2015-04-23 | 한국전자통신연구원 | Apparatus and methdo for howling suppression |
| CN106373587A (en) * | 2016-08-31 | 2017-02-01 | 北京容联易通信息技术有限公司 | Automatic sound feedback detection and elimination method of real-time communication system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109218917A (en) | 2019-01-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111489760B (en) | Speech signal dereverberation processing method, device, computer equipment and storage medium | |
| CN113766073B (en) | Howling detection in conference systems | |
| CA2527461C (en) | Reverberation estimation and suppression system | |
| CN111048119B (en) | Call audio mixing processing method and device, storage medium and computer equipment | |
| US9064501B2 (en) | Speech processing device and speech processing method | |
| CN109461455B (en) | System and method for eliminating howling | |
| CN108235181B (en) | Method for noise reduction in an audio processing apparatus | |
| JPH09503590A (en) | Background noise reduction to improve conversation quality | |
| CN103077727A (en) | Method and device used for speech quality monitoring and prompting | |
| EP2896126B1 (en) | Long term monitoring of transmission and voice activity patterns for regulating gain control | |
| CN103124165A (en) | Automatic gain control | |
| KR20180036778A (en) | Event detection for playback management in audio devices | |
| CN104796836A (en) | Binaural source enhancement | |
| JP2001202100A (en) | Reduction of exponential echo and noise in silent section | |
| CN104580764A (en) | Ultrasound pairing signal control in teleconferencing system | |
| CN109218917B (en) | Automatic acoustic feedback monitoring and eliminating method in real-time communication system | |
| US9837064B1 (en) | Generating spectrally shaped sound signal based on sensitivity of human hearing and background noise level | |
| WO2023051622A1 (en) | Method for improving far-field speech interaction performance, and far-field speech interaction system | |
| CN113573212B (en) | Sound amplifying system and microphone channel data selection method | |
| EP3566229B1 (en) | An apparatus and method for enhancing a wanted component in a signal | |
| US11500610B2 (en) | Transmission control for audio device using auxiliary signals | |
| US20240107252A1 (en) | Insertion of forced gaps for pervasive listening | |
| US20240048931A1 (en) | Orchestration of acoustic direct sequence spread spectrum signals for estimation of acoustic scene metrics | |
| US20240056757A1 (en) | Orchestration of acoustic direct sequence spread spectrum signals for estimation of acoustic scene metrics | |
| CN112151055B (en) | Audio processing method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |