CN106303878A - One is uttered long and high-pitched sounds and is detected and suppressing method - Google Patents
One is uttered long and high-pitched sounds and is detected and suppressing method Download PDFInfo
- Publication number
- CN106303878A CN106303878A CN201510263917.5A CN201510263917A CN106303878A CN 106303878 A CN106303878 A CN 106303878A CN 201510263917 A CN201510263917 A CN 201510263917A CN 106303878 A CN106303878 A CN 106303878A
- Authority
- CN
- China
- Prior art keywords
- pitched sounds
- frame
- uttering long
- long
- doubtful
- 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.)
- Pending
Links
Landscapes
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
This application discloses one to utter long and high-pitched sounds detection and suppressing method, including: the voice signal of the present frame of reception is converted to frequency domain, is not that the voice signal of 0 is uttered long and high-pitched sounds detection to the signal amplitude in frequency domain of uttering long and high-pitched sounds；Described frequency domain speech signal is carried out energy measuring identification efficient voice signal；Described efficient voice signal uttered long and high-pitched sounds detection according to algorithm of uttering long and high-pitched sounds, determine doubtful frame of uttering long and high-pitched sounds；Calculate doubtful frame medium and low frequency energy and the ratio of high-frequency energy of uttering long and high-pitched sounds, using ratio higher than setting the doubtful frame of uttering long and high-pitched sounds of the first flase drop ratio as flase drop frame, and/or, add up the average harmonic number of doubtful frame of uttering long and high-pitched sounds, using average harmonic number higher than setting the doubtful frame of uttering long and high-pitched sounds of the second flase drop threshold value as flase drop frame；Will not belong to the doubtful frame of uttering long and high-pitched sounds of flase drop frame as actual frame of uttering long and high-pitched sounds, carry out chauvent's criterion.Application the application, it is possible to reduce the false drop rate of frame of uttering long and high-pitched sounds.
Description
Technical field
The application relates to communication technology, utters long and high-pitched sounds particularly to one and detects and suppressing method.
Background technology
The speech business form of sector terminal is mainly the business such as cluster, DMO, and such business mainly uses outer putting
Pattern.Owing to sector terminal major part is operated in the bigger open air of background noise or Factory Building, it is desirable to give great volume, therefore eventually
The up-downgoing volume gain of end is generally adjusted bigger, and after sound is amplified by loop gain, energy is constantly accumulated formation and maked a whistling sound
Cry, and utter long and high-pitched sounds and have a strong impact on the normal use of speech business, client perception is caused great discomfort, therefore to uttering long and high-pitched sounds
Scene carries out accurately identifying having a very big significance
Sector terminal is immature for scene Recognition solution of uttering long and high-pitched sounds at present, and a large amount of identifying schemes efficiency are low and identify
Inaccurate, particularly flase drop problem, this problem is primarily referred to as not existing utters long and high-pitched sounds phenomenon but algorithm of uttering long and high-pitched sounds is judged as field of uttering long and high-pitched sounds
Scape also carries out chauvent's criterion, and voice can be caused damage by chauvent's criterion, thus affects the voice quality of non-scene of uttering long and high-pitched sounds,
Affect client perception.
Summary of the invention
The application provides one to utter long and high-pitched sounds and detects and suppressing method, it is possible to reduce the false drop rate uttered long and high-pitched sounds.
For achieving the above object, the application adopts the following technical scheme that:
One is uttered long and high-pitched sounds and is detected and suppressing method, including:
The voice signal of the present frame of reception is converted to frequency domain, is not 0 to the signal amplitude in frequency domain of uttering long and high-pitched sounds
Voice signal carry out detection of uttering long and high-pitched sounds；
Described frequency domain speech signal is carried out energy measuring identification efficient voice signal；
Described efficient voice signal uttered long and high-pitched sounds detection according to algorithm of uttering long and high-pitched sounds, determine doubtful frame of uttering long and high-pitched sounds；
Calculate described doubtful frame medium and low frequency energy and the ratio of high-frequency energy of uttering long and high-pitched sounds, by described ratio higher than setting first by mistake
The doubtful frame of uttering long and high-pitched sounds of inspection ratio is as flase drop frame；
Will not belong to the doubtful frame of uttering long and high-pitched sounds of flase drop frame as actual frame of uttering long and high-pitched sounds, carry out chauvent's criterion.
It is preferred that after determining described doubtful frame of uttering long and high-pitched sounds, carry out chauvent's criterion before, the method farther includes: statistics
The average harmonic number of described doubtful frame of uttering long and high-pitched sounds, makees average harmonic number higher than the doubtful frame of uttering long and high-pitched sounds setting the second flase drop threshold value
For flase drop frame.
One is uttered long and high-pitched sounds and is detected and suppressing method, including:
The voice signal of the present frame of reception is converted to frequency domain, is not 0 to the signal amplitude in frequency domain of uttering long and high-pitched sounds
Voice signal carry out detection of uttering long and high-pitched sounds；
Described frequency domain speech signal is carried out energy measuring identification efficient voice signal；
Described efficient voice signal uttered long and high-pitched sounds detection according to algorithm of uttering long and high-pitched sounds, determine doubtful frame of uttering long and high-pitched sounds；
Add up the average harmonic number of described doubtful frame of uttering long and high-pitched sounds, by average harmonic number higher than setting the doubtful of the second flase drop threshold value
Utter long and high-pitched sounds frame as flase drop frame；
Will not belong to the doubtful frame of uttering long and high-pitched sounds of flase drop frame as actual frame of uttering long and high-pitched sounds, carry out chauvent's criterion.
It is preferred that after determining described doubtful frame of uttering long and high-pitched sounds, carry out chauvent's criterion before, the method farther includes: calculate
Described doubtful frame medium and low frequency energy and the ratio of high-frequency energy of uttering long and high-pitched sounds, by described ratio higher than setting the first flase drop ratio
Doubtful frame of uttering long and high-pitched sounds is as flase drop frame.
As seen from the above technical solution, in the application, the voice signal of reception is converted to frequency domain, is pointed to frequency of uttering long and high-pitched sounds
Frequency domain speech signal in the range of territory carries out detection of uttering long and high-pitched sounds；It is pointed to the frequency domain speech signal in the range of uttering long and high-pitched sounds and carries out energy
Detection identifies efficient voice signal；Efficient voice signal uttered long and high-pitched sounds detection according to algorithm of uttering long and high-pitched sounds, determine doubtful uttering long and high-pitched sounds
Frame；Calculate doubtful frame medium and low frequency energy and the ratio of high-frequency energy of uttering long and high-pitched sounds, by ratio higher than setting the first flase drop ratio
Doubtful frame of uttering long and high-pitched sounds as flase drop frame, and/or, add up the average harmonic number of doubtful frame of uttering long and high-pitched sounds, by average harmonic number higher than setting
The doubtful frame of uttering long and high-pitched sounds of fixed second flase drop threshold value is as flase drop frame；The doubtful frame of uttering long and high-pitched sounds that will not belong to flase drop frame is maked a whistling sound as reality
It is frame, carries out chauvent's criterion.By the way, on the one hand by the frequency range of signal and energy measuring determine into
The signal of capable detection of uttering long and high-pitched sounds, gets rid of and will not produce the voice signal uttered long and high-pitched sounds；On the other hand by low frequency and high-frequency energy
Ratio characteristic and/or harmonic characterisitic, identify flase drop frame from doubtful frame of uttering long and high-pitched sounds.Process, greatly of both above
Reduce greatly the false drop rate uttered long and high-pitched sounds.
Accompanying drawing explanation
Fig. 1 is the detection schematic flow sheet with suppressing method of uttering long and high-pitched sounds in the application；
Fig. 2 is existing to be carried out uttering long and high-pitched sounds the schematic diagram identified by peak-to-average force ratio；
Fig. 3 is the schematic diagram uttered long and high-pitched sounds and concentrate on high-frequency region；
Fig. 4 is the normal voice schematic diagram at low frequency region；
Frequency spectrum exemplary plot when Fig. 5 is to utter long and high-pitched sounds；
Fig. 6 is frequency spectrum exemplary plot during normal voice.
Detailed description of the invention
In order to make the purpose of the application, technological means and advantage clearer, below in conjunction with accompanying drawing to this Shen
Please be described in further details.
In the application, frequency domain is analyzed the difference with non-signal of uttering long and high-pitched sounds of uttering long and high-pitched sounds.If belonging to non-to utter long and high-pitched sounds under scene
Distinctive phenomenon, can extract feature, and to uttering long and high-pitched sounds, Rule of judgment retrains, and prevents flase drop problem.
Specifically, the application mainly by detection frequency domain thresholding of uttering long and high-pitched sounds, energy measuring, low-and high-frequency energy proportion judge with
And the mode such as harmonic wave identification reduces probability of miscarriage of justice.Below by embodiment, implementing of the application is described.
Fig. 1 is the detection schematic flow sheet with suppressing method of uttering long and high-pitched sounds in the application.As it is shown in figure 1, the method bag
Include:
Step 101, is converted to frequency domain by the voice signal of the present frame of reception, to the letter in frequency domain of uttering long and high-pitched sounds
Number amplitude be not 0 voice signal carry out detection of uttering long and high-pitched sounds, perform subsequent step.
By analyzing many groups of audio files of uttering long and high-pitched sounds, find that frequency of uttering long and high-pitched sounds commonly takes place in certain frequency domain,
Such as more than 1KHz, below 1KHz are fewer.And the speech energy of below 1KHz compares concentration, compare appearance
Easily there is flase drop, therefore can utter long and high-pitched sounds frequency domain (such as below 1KHz) for setting when uttering long and high-pitched sounds detection
Interior signal amplitude be 0 voice signal do not carry out uttering long and high-pitched sounds a detection, so can reduce probability of false detection of uttering long and high-pitched sounds, with
Time less on the impact of chauvent's criterion performance.Here, frequency domain of uttering long and high-pitched sounds may be greater than the detection of uttering long and high-pitched sounds of setting
Frequency domain thresholding.
Step 102, carries out energy judgement to the frequency domain speech signal uttered long and high-pitched sounds in frequency domain, identifies efficient voice letter
Number.
Identification of uttering long and high-pitched sounds is carried out in the prior art by peak-to-average force ratio.But, if being based entirely on the peak-to-average force ratio of frequency spectrum
Judge peak point of uttering long and high-pitched sounds, then some may be judged by accident into without end noise during voice and utter long and high-pitched sounds a little, thus to howl
The judgement being frame produces impact, as shown in Figure 2.Based on this, the application calculates the average of current speech signal
Energy work rate, if less than a certain thresholding, then it is assumed that this part signal belongs to non-effective voice signal, does not makes a whistling sound
Cry detection and suppression；If greater than or equal to respective threshold, it is believed that this part signal belongs to efficient voice signal, right
This part signal performs step 103 and subsequent step thereof.
Step 103, utters long and high-pitched sounds to efficient voice signal detection according to algorithm of uttering long and high-pitched sounds, determines doubtful frame of uttering long and high-pitched sounds.
This step can use existing algorithm of uttering long and high-pitched sounds carry out detection of uttering long and high-pitched sounds, and using detect utter long and high-pitched sounds frame as
Doubtful frame of uttering long and high-pitched sounds.
Step 104, calculates doubtful frame medium and low frequency energy and the ratio of high-frequency energy of uttering long and high-pitched sounds, by ratio higher than setting the
The doubtful frame of uttering long and high-pitched sounds of one flase drop ratio is as flase drop frame.
In general, utter long and high-pitched sounds and concentrate on the region (as shown in Figure 3) of high frequency, and voice be primarily present low frequency region (as
Shown in Fig. 4),.Vertical frame dimension frequency energy comparison of uttering long and high-pitched sounds is strong, and low frequency energy ratio is relatively low, and normal voice bag low frequency energy compares
By force, high-frequency energy ratio is relatively low, judges whether to belong to flase drop according to this feature of low-and high-frequency capacity volume variance in the application.
If the most doubtful frame of uttering long and high-pitched sounds meets following condition, then judge that the most doubtful frame of uttering long and high-pitched sounds is flase drop, do not carry out uttering long and high-pitched sounds pressing down
System:
Wherein Phigh_freqRepresent the average energy of high-frequency region, Plow_freqRepresent the average energy of low frequency region, high frequency
Dividing specifically to arrange according to different platform with low frequency region and carry out, (voice mainly collects to be traditionally arranged to be about 1KHZ
In at below 1KHZ), high_low_ratio_thd is low-and high-frequency ratio threshold, and value is according to different platform feature
Concrete setting.
Step 105, adds up the average harmonic number of doubtful frame of uttering long and high-pitched sounds, by average harmonic number higher than the second flase drop set
The doubtful frame of uttering long and high-pitched sounds of threshold value is as flase drop frame.
Analyze and utter long and high-pitched sounds and spectrum signature during normal voice, can be seen that normal voice is compared from spectral contrast and utter long and high-pitched sounds
There is obvious harmonic characteristic, be illustrated in figure 5 frequency spectrum exemplary plot when uttering long and high-pitched sounds, be illustrated in figure 6 normal
Frequency spectrum exemplary plot during voice.Based on this, for identifying flase drop frame, each doubtful frame of uttering long and high-pitched sounds can be added up
Average harmonic wave number, statistical can be arranged as required to.Such as take each 15 subbands about high frequency peaks point,
Every time from minimum point to peak or peak is designated as first harmonic to minimum point more than 10dB.If it is the most doubtful
The average harmonic number of frame of uttering long and high-pitched sounds exceedes certain thresholding Average_Harmony_Num_Thd, then judge the most doubtful
Frame of uttering long and high-pitched sounds belongs to flase drop, does not carry out chauvent's criterion.
Can identify flase drop by above-mentioned steps 104 and 105 is the speech frame of frame of uttering long and high-pitched sounds.Apply in reality
In, above-mentioned steps 104 and 105 can all use, or selects one way in which to use.When two steps
When all using, the sequencing of execution can be arranged as required to, it is also possible to carry out simultaneously.When by above-mentioned side
After formula determines flase drop frame, continue executing with step 106.
Step 106, will not belong to the doubtful frame of uttering long and high-pitched sounds of flase drop frame as actual frame of uttering long and high-pitched sounds, carries out chauvent's criterion.
So far, the detection of uttering long and high-pitched sounds in the application terminates with suppressing method flow process.By the process of above-mentioned the application,
Can effectively reduce algorithm probability of false detection of uttering long and high-pitched sounds, improve Detection accuracy of uttering long and high-pitched sounds, adaptive chauvent's criterion processes, and carries
High voice quality and Consumer's Experience.
The foregoing is only presently preferred embodiments of the present invention, not in order to limit the present invention, all in the present invention
Spirit and principle within, any modification, equivalent substitution and improvement etc. done, should be included in the present invention protect
Within the scope of protecting.
Claims (4)
1. utter long and high-pitched sounds detection and suppressing method, it is characterised in that including:
The voice signal of the present frame of reception is converted to frequency domain, is not 0 to the signal amplitude in frequency domain of uttering long and high-pitched sounds
Voice signal carry out detection of uttering long and high-pitched sounds；
Described frequency domain speech signal is carried out energy measuring identification efficient voice signal；
Described efficient voice signal uttered long and high-pitched sounds detection according to algorithm of uttering long and high-pitched sounds, determine doubtful frame of uttering long and high-pitched sounds；
Calculate described doubtful frame medium and low frequency energy and the ratio of high-frequency energy of uttering long and high-pitched sounds, by described ratio higher than setting first by mistake
The doubtful frame of uttering long and high-pitched sounds of inspection ratio is as flase drop frame；
Will not belong to the doubtful frame of uttering long and high-pitched sounds of flase drop frame as actual frame of uttering long and high-pitched sounds, carry out chauvent's criterion.
Method the most according to claim 1, it is characterised in that after determining described doubtful frame of uttering long and high-pitched sounds, make a whistling sound
Before crying suppression, the method farther includes: add up the average harmonic number of described doubtful frame of uttering long and high-pitched sounds, and average harmonic number is high
In setting the doubtful frame of uttering long and high-pitched sounds of the second flase drop threshold value as flase drop frame.
3. utter long and high-pitched sounds detection and suppressing method, it is characterised in that including:
The voice signal of the present frame of reception is converted to frequency domain, is not 0 to the signal amplitude in frequency domain of uttering long and high-pitched sounds
Voice signal carry out detection of uttering long and high-pitched sounds；
Described frequency domain speech signal is carried out energy measuring identification efficient voice signal；
Described efficient voice signal uttered long and high-pitched sounds detection according to algorithm of uttering long and high-pitched sounds, determine doubtful frame of uttering long and high-pitched sounds；
Add up the average harmonic number of described doubtful frame of uttering long and high-pitched sounds, by average harmonic number higher than setting the doubtful of the second flase drop threshold value
Utter long and high-pitched sounds frame as flase drop frame；
Will not belong to the doubtful frame of uttering long and high-pitched sounds of flase drop frame as actual frame of uttering long and high-pitched sounds, carry out chauvent's criterion.
Method the most according to claim 3, it is characterised in that after determining described doubtful frame of uttering long and high-pitched sounds, make a whistling sound
Before crying suppression, the method farther includes: calculate described doubtful frame medium and low frequency energy and the ratio of high-frequency energy of uttering long and high-pitched sounds,
Using described ratio higher than setting the doubtful frame of uttering long and high-pitched sounds of the first flase drop ratio as flase drop frame.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510263917.5A CN106303878A (en) | 2015-05-22 | 2015-05-22 | One is uttered long and high-pitched sounds and is detected and suppressing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510263917.5A CN106303878A (en) | 2015-05-22 | 2015-05-22 | One is uttered long and high-pitched sounds and is detected and suppressing method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106303878A true CN106303878A (en) | 2017-01-04 |
Family
ID=57632353
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510263917.5A Pending CN106303878A (en) | 2015-05-22 | 2015-05-22 | One is uttered long and high-pitched sounds and is detected and suppressing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106303878A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106981294A (en) * | 2017-02-17 | 2017-07-25 | 成都易慧通科技有限公司 | A kind of removing method of uttering long and high-pitched sounds judged based on infrasonic feature |
| CN108053833A (en) * | 2017-11-30 | 2018-05-18 | 广东欧珀移动通信有限公司 | Processing method, device, electronic equipment and the storage medium that voice is uttered long and high-pitched sounds |
| CN108449505A (en) * | 2018-03-12 | 2018-08-24 | 广东欧珀移动通信有限公司 | Voice communication data detection method, device, storage medium and mobile terminal |
| CN109637552A (en) * | 2018-11-29 | 2019-04-16 | 河北远东通信系统工程有限公司 | A kind of method of speech processing for inhibiting audio frequency apparatus to utter long and high-pitched sounds |
| CN109671445A (en) * | 2018-12-28 | 2019-04-23 | 广东美电贝尔科技集团股份有限公司 | A kind of suppressing method that audio system sound is uttered long and high-pitched sounds |
| CN110213694A (en) * | 2019-04-16 | 2019-09-06 | 浙江大华技术股份有限公司 | A kind of audio frequency apparatus and its processing method, the computer storage medium uttered long and high-pitched sounds |
| CN111418004A (en) * | 2018-06-11 | 2020-07-14 | 思睿逻辑国际半导体有限公司 | Techniques for howling detection |
| CN112004177A (en) * | 2020-09-03 | 2020-11-27 | 重庆瑞可利科技有限公司 | Howling detection method, microphone volume adjustment method and medium |
| CN112349295A (en) * | 2020-10-20 | 2021-02-09 | 浙江大华技术股份有限公司 | Howling detection method and device |
| CN112562717A (en) * | 2020-12-01 | 2021-03-26 | 广州华多网络科技有限公司 | Howling detection method, howling detection device, storage medium and computer equipment |
| CN113316075A (en) * | 2021-05-11 | 2021-08-27 | 紫光展锐(重庆)科技有限公司 | Howling detection method and device and electronic equipment |
Citations (7)
| 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 |
| CN1463422A (en) * | 2001-06-06 | 2003-12-24 | 三菱电机株式会社 | Noise suppressor |
| CN202475574U (en) * | 2012-03-12 | 2012-10-03 | 杭州艾力特音频技术有限公司 | Echo-cancelling talkback equipment |
| CN102740214A (en) * | 2011-04-01 | 2012-10-17 | 中国科学院声学研究所 | Howling suppression method based on feedback signal spectrum estimation |
| CN102986252A (en) * | 2011-04-11 | 2013-03-20 | 松下电器产业株式会社 | Hearing aid and method of detecting vibration |
| CN103391496A (en) * | 2013-07-16 | 2013-11-13 | 歌尔声学股份有限公司 | Howling inhibition method and device for ANR (Active Noise Reduction) earphones |
| CN103516440A (en) * | 2012-06-29 | 2014-01-15 | 华为技术有限公司 | Audio signal processing method and encoding device |
-
2015
- 2015-05-22 CN CN201510263917.5A patent/CN106303878A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1463422A (en) * | 2001-06-06 | 2003-12-24 | 三菱电机株式会社 | Noise suppressor |
| 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 |
| CN102740214A (en) * | 2011-04-01 | 2012-10-17 | 中国科学院声学研究所 | Howling suppression method based on feedback signal spectrum estimation |
| CN102986252A (en) * | 2011-04-11 | 2013-03-20 | 松下电器产业株式会社 | Hearing aid and method of detecting vibration |
| CN202475574U (en) * | 2012-03-12 | 2012-10-03 | 杭州艾力特音频技术有限公司 | Echo-cancelling talkback equipment |
| CN103516440A (en) * | 2012-06-29 | 2014-01-15 | 华为技术有限公司 | Audio signal processing method and encoding device |
| CN103391496A (en) * | 2013-07-16 | 2013-11-13 | 歌尔声学股份有限公司 | Howling inhibition method and device for ANR (Active Noise Reduction) earphones |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106981294A (en) * | 2017-02-17 | 2017-07-25 | 成都易慧通科技有限公司 | A kind of removing method of uttering long and high-pitched sounds judged based on infrasonic feature |
| CN108053833A (en) * | 2017-11-30 | 2018-05-18 | 广东欧珀移动通信有限公司 | Processing method, device, electronic equipment and the storage medium that voice is uttered long and high-pitched sounds |
| CN108449505A (en) * | 2018-03-12 | 2018-08-24 | 广东欧珀移动通信有限公司 | Voice communication data detection method, device, storage medium and mobile terminal |
| CN111418004A (en) * | 2018-06-11 | 2020-07-14 | 思睿逻辑国际半导体有限公司 | Techniques for howling detection |
| CN111418004B (en) * | 2018-06-11 | 2023-12-22 | 思睿逻辑国际半导体有限公司 | Techniques for howling detection |
| CN109637552A (en) * | 2018-11-29 | 2019-04-16 | 河北远东通信系统工程有限公司 | A kind of method of speech processing for inhibiting audio frequency apparatus to utter long and high-pitched sounds |
| CN109671445A (en) * | 2018-12-28 | 2019-04-23 | 广东美电贝尔科技集团股份有限公司 | A kind of suppressing method that audio system sound is uttered long and high-pitched sounds |
| CN110213694A (en) * | 2019-04-16 | 2019-09-06 | 浙江大华技术股份有限公司 | A kind of audio frequency apparatus and its processing method, the computer storage medium uttered long and high-pitched sounds |
| CN112004177A (en) * | 2020-09-03 | 2020-11-27 | 重庆瑞可利科技有限公司 | Howling detection method, microphone volume adjustment method and medium |
| CN112349295A (en) * | 2020-10-20 | 2021-02-09 | 浙江大华技术股份有限公司 | Howling detection method and device |
| CN112349295B (en) * | 2020-10-20 | 2023-03-31 | 浙江大华技术股份有限公司 | Howling detection method and device |
| CN112562717A (en) * | 2020-12-01 | 2021-03-26 | 广州华多网络科技有限公司 | Howling detection method, howling detection device, storage medium and computer equipment |
| CN112562717B (en) * | 2020-12-01 | 2023-08-18 | 广州华多网络科技有限公司 | Howling detection method and device, storage medium and computer equipment |
| CN113316075A (en) * | 2021-05-11 | 2021-08-27 | 紫光展锐(重庆)科技有限公司 | Howling detection method and device and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106303878A (en) | One is uttered long and high-pitched sounds and is detected and suppressing method | |
| CN109545188A (en) | A kind of real-time voice end-point detecting method and device | |
| CN104464722B (en) | Voice activity detection method and apparatus based on time domain and frequency domain | |
| US8422696B2 (en) | Apparatus and method for removing noise | |
| CN103646649A (en) | High-efficiency voice detecting method | |
| CN109643552A (en) | Robust noise estimation for speech enhan-cement in variable noise situation | |
| CA2699316A1 (en) | Apparatus and method for calculating bandwidth extension data using a spectral tilt controlled framing | |
| CN103632681B (en) | A kind of spectral envelope silence detection method | |
| CN106488052A (en) | One kind is uttered long and high-pitched sounds scene recognition method and equipment | |
| JP2010061151A (en) | Voice activity detector and validator for noisy environment | |
| CN104637489A (en) | Method and device for processing sound signals | |
| CN105427859A (en) | Front voice enhancement method for identifying speaker | |
| CN103730126B (en) | Noise suppressing method and noise silencer | |
| CN106098076A (en) | A kind of based on dynamic noise estimation time-frequency domain adaptive voice detection method | |
| CN106504760B (en) | Broadband ambient noise and speech Separation detection system and method | |
| CN1763844B (en) | End-point detecting method, apparatus and speech recognition system based on sliding window | |
| EP3136389B1 (en) | Noise detection method and apparatus | |
| CN105810201A (en) | Voice activity detection method and system | |
| CN105513614A (en) | Voice activation detection method based on noise power spectrum density Gamma distribution statistical model | |
| Wang et al. | Voice activity detection algorithm with low signal-to-noise ratios based on spectrum entropy | |
| CN108986844B (en) | Speech endpoint detection method based on speaker speech characteristics | |
| CN103021421A (en) | Multilevel screening detecting recognizing method for shots | |
| CN106486133B (en) | One kind is uttered long and high-pitched sounds scene recognition method and equipment | |
| CN102201230A (en) | Voice detection method for emergency | |
| Evangelopoulos et al. | Speech event detection using multiband modulation energy. |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170104 |