CN101442696A - Method for filtering sound noise - Google Patents
Method for filtering sound noise Download PDFInfo
- Publication number
- CN101442696A CN101442696A CNA2007101938137A CN200710193813A CN101442696A CN 101442696 A CN101442696 A CN 101442696A CN A2007101938137 A CNA2007101938137 A CN A2007101938137A CN 200710193813 A CN200710193813 A CN 200710193813A CN 101442696 A CN101442696 A CN 101442696A
- Authority
- CN
- China
- Prior art keywords
- radio reception
- reception device
- coefficient
- voice signal
- noise
- 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
Images
Abstract
The invention discloses a method for filtering sound noise, which comprises the following steps: firstly, using a plurality of sound receivers of a sound receiver array to detect a sound signal transmitted by a sound source; secondly, analyzing a sound signal detected by each sound receiver, and selecting the first sound receiver with good sound-receiving effect from the sound receiver array and the second sound receiver with poor sound-receiving effect; and finally using the second sound signal received by the second sound receiver to filter the noise in the first sound signal received by the first sound receiver, thereby acquiring a sound signal of which the noise is filtered. Therefore, the ambient noise in the sound signal can be effectively filtered, and the speech quality is improved.
Description
Technical field
The present invention is relevant for a kind of noise filtering method, and particularly relevant for a kind of method of filtering microphone voice noise.
Background technology
Along with making rapid progress of electronics technology, various electronic products are constantly weeded out the old and bring forth the new, wherein portable electron devices such as mobile phone, personal digital assistant and notebook computer rely on its compact and be convenient for carrying advantage, and become one of instrument indispensable in modern's life gradually.Therefore in these portable electron devices, mobile phone has instant messaging, is not subjected to the characteristic of time, site limitation, also replaces black phone gradually, becomes the most convenient between the person to person, means of communication the most fast.On the other hand, personal digital assistant or notebook computer also can be by installing instant communication software, and provide the function of instant messaging, another approach that provides people to communicate by the transmission of network.
At present portable electron device is when carrying out instant communication function, and the usefulness of its radio reception device and stability have all obtained to promote, no matter be under rugged environment or be in high-speed mobile, can both keep the speech quality of clear and smooth.Yet, be exactly because the sound articulation of radio reception device improves, also can make its radio reception quality very easily be subjected to the influence of noise in the environment jointly, when first speaker is called, the noise of its one's voice in speech and surrounding environment all can pass in the radio reception device, and passes to by communicating to connect of portable electron device in those who answer's the ear.If in general indoor environment, environmental noise is less, influence to speech quality also can be very not big, if but the next door is the crowd of clamour or has the steam turbine car to cross, those who answer just is difficult to listen to such an extent that know the sound of first speaker, particularly when portable electron device is under the pattern that amplifies of Handless.
In order to address the above problem, existing a solution is a threshold value of setting a sound amplitude in portable electron device, when the amplitude of single-frequency in the voice signal that it received during greater than this threshold value, the voice signal of this frequency just can be transmitted, when if very little and its amplitude of noise surpasses threshold value, noise will be earlier by filtering to improve speech quality, but if noise is quite big and when surpassing threshold value, these noises still can be transmitted, and the result still can't solve the problem of noise jamming at all.
Summary of the invention
In view of this, the invention provides a kind of method of filtering sound noise, by selecting the preferable main radio reception device of radio reception, and the sound noise of the received voice signal filtering master radio reception device of other radio reception devices around utilizing, make the radio reception effect reach optimization.
For reaching above-mentioned or other purposes, the present invention proposes a kind of method of filtering sound noise, is applicable to electronic installation, and this method comprises the following steps: that a. utilizes radio reception device array received one signal that source of sound is sounded, and this radio reception device array comprises a plurality of radio reception devices; B. analyze voice signal, select the first preferable radio reception device of at least one radio reception in the radio reception device array, and the second relatively poor radio reception device of at least one radio reception.Wherein, the second relatively poor radio reception device of radio reception can comprise the radio reception device except that the first radio reception device in the radio reception device array; C. the noise in first voice signal that the second sound signal filtering first radio reception device that utilizes the second radio reception device to be received is received, the voice signal of acquisition filtering noise.
In one embodiment of this invention, above-mentioned step b also comprises the analysis voice signal, selects the first radio reception device of radio reception the best in the first radio reception device, and selects the second the poorest radio reception device of radio reception in the second radio reception device.Step c then also comprises the noise in first voice signal that the first radio reception device of second sound signal filtering radio reception the best that the second the poorest radio reception device that utilizes radio reception is received received, and obtains the voice signal of filtering noise.
In one embodiment of this invention, above-mentioned step c comprises first voice signal and second sound signal is multiplied by addition behind first coefficient and second coefficient individually, and with the voice signal of addition result as filtering noise.
In one embodiment of this invention, before above-mentioned step c, also comprise and judge that first coefficient and second coefficient whether in before being recorded, if having, then are written into first coefficient and second coefficient of precedence record, if nothing, then initialization first coefficient and second coefficient.
In one embodiment of this invention, after the above-mentioned steps c, also comprise steps d.: whether the tonequality of judging voice signal reaches a preset standard.Wherein,, then adjust the size of first coefficient and second coefficient, recomputate voice signal, and write down the size of first coefficient and second coefficient if the tonequality of voice signal does not reach preset standard; Otherwise,, then write down the size of first coefficient and second coefficient if the tonequality of voice signal reaches preset standard.
In one embodiment of this invention, the size of above-mentioned adjustment first coefficient and second coefficient, recomputate after the step of voice signal, also comprise repeated execution of steps c, d, till the tonequality of voice signal reaches this preset standard, or repeated execution of steps a-d, reselect the preferable first radio reception device and the second relatively poor radio reception device of radio reception of radio reception in the radio reception device array, and the calculating voice signal, till the tonequality of voice signal reaches preset standard.
In one embodiment of this invention, before the above-mentioned steps c, also comprise first initialization first coefficient and second coefficient, or be written into first coefficient and second coefficient of precedence record earlier.
In one embodiment of this invention, above-mentioned source of sound comprises the sound that is sent by the user, and above-mentioned radio reception device then comprises directional microphone (unidirectional microphone).
In one embodiment of this invention, the method of above-mentioned analysis voice signal comprises gauss hybrid models (Gaussian mixture model, GMM), minimize classification error (Minimum ClassificationError, MCE), vector quantization coding (Vector Quantization, VQ) and mixed Gaussian mixed model/SVMs (Hybrid Gaussian mixture model/Support vector machine, Hybrid GMM/SVM) one of them.
In one embodiment of this invention, above-mentioned electronic installation comprise mobile phone, personal digital assistant and notebook computer one of them.
The present invention is because of adopting the structure with radio reception device array received voice signal, at the beginning of the detection voice signal, select the preferable radio reception device of radio reception as main audio signal reception device earlier, and utilize the relatively poor detected voice signal of one or more radio reception devices of radio reception to come noise in the voice signal that filtering master audio signal reception device received, and effective filtering sound noise improves the radio reception quality.
For above-mentioned and other purposes, feature and advantage of the present invention can be become apparent, preferred embodiment cited below particularly, and cooperate appended graphicly, be described in detail below.
Description of drawings
Fig. 1 is the schematic diagram according to the filtering sound noise method shown in one embodiment of the invention.
Fig. 2 is the method flow diagram according to the filtering sound noise shown in one embodiment of the invention.
Fig. 3 is the method flow diagram according to the filtering sound noise shown in another embodiment of the present invention.
Fig. 4 is the method flow diagram according to the filtering sound noise shown in further embodiment of this invention.
The main element symbol description
100: electronic installation
110: processor
120: first speaker
S210-S230: each step of the method for the filtering sound noise of one embodiment of the invention
S310-S390: each step of the method for the filtering sound noise of another embodiment of the present invention
S410-S490: each step of the method for the filtering sound noise of further embodiment of this invention
Embodiment
The mode that tradition uses single group of radio reception device to receive voice signal very easily influences its radio reception quality because receiving the noise of surrounding environment, and because the radio reception scope of radio reception device has certain directivity, in case during user's this radio reception scope of deviation in driction of giving orders or instructions, to make that more the radio reception quality significantly reduces, even can't listen to.In view of the above, the present invention is promptly by setting up many group radio reception devices enlarging the radio reception scope in electronic installation, and the voice signal that is received with reference to each group radio reception device simultaneously comes filtering or sound-inhibiting noise, and reaches the purpose that promotes the radio reception quality.In order to make content of the present invention more clear, below the example that can implement according to this really as the present invention especially exemplified by embodiment.
Fig. 1 is the schematic diagram according to the filtering sound noise method shown in one embodiment of the invention.Please refer to Fig. 1, present embodiment system receives voice signal by the many groups of configuration radio reception device (comprising radio reception device A, B, C, D) in electronic installation 100 and form one group of radio reception device array, uses the radio reception scope that increases electronic installation 100.Wherein, electronic installation for example is portable electron devices such as mobile phone, personal digital assistant or notebook computer, and does not limit its scope.
Each radio reception device is after receiving the voice signal that first speaker 120 sent, promptly voice signal can be sent in the processor 110 of electronic installation 100, carrying out the filtering operation of sound noise, below promptly describe in detail for an embodiment with regard to the step of this filtering sound noise.
Fig. 2 is the method flow diagram according to the filtering sound noise shown in one embodiment of the invention.Please be simultaneously with reference to Fig. 1 and Fig. 2, present embodiment is applicable to as the electronic installation among the above-mentioned embodiment 100, by the received voice signal of processor 110 comprehensive a plurality of radio reception devices (comprising radio reception device A, B, C, D) of electronic installation 100, and the noise in the filtering sound signal, its step is as follows:
At first, in step S210, utilize radio reception device array detection by signal that source of sound is sounded.This source of sound for example is the sound that is sent by the user, or the sound that sends of other pronunciation devices, and does not limit its scope.
Then, in step S220, then further analyze detected voice signal, from radio reception device array, selecting the first preferable radio reception device of at least one radio reception, and the second relatively poor radio reception device of at least one radio reception.Wherein, the method of above-mentioned analysis voice signal comprises gauss hybrid models (Gaussianmixture model, GMM), minimize classification error (Minimum Classification Error, MCE), vector quantization coding (Vector Quantization, VQ) or mixed Gaussian mixed model/SVMs (Hybrid Gaussian mixture model/Support vector machine, HybridGMM/SVM) etc., and utilize these methods can in radio reception device array, select the preferable or radio reception device that gain is bigger of reception sound signal quality, with radio reception device as the main voice signal of reception, also can select simultaneously and receive the relatively poor or radio reception device that gains less of sound signal quality, and in order to the noise in the main voice signal of filtering.
For instance, in the radio reception device array of the electronic installation 100 of Fig. 1, apart from first speaker 120 nearer radio reception devices is radio reception device B and D, therefore around under the situation of ambient noise normal distribution, the radio reception quality of radio reception device B and D can be good than the radio reception quality of radio reception device A and C, therefore can be as the radio reception device that receives main voice signal, be the best with radio reception quality the most approaching and that face the radio reception device B of first speaker 120 again wherein.Relatively, radio reception device A and C are the relatively poor radio reception device of radio reception, and the voice signal that it received promptly can be used to the noise in the main voice signal of filtering, wherein again with back to first speaker 120 and the poorest apart from the radio reception quality of first speaker 120 radio reception device C farthest.
What deserves to be mentioned is that though be to select the best and the poorest radio reception device of at least one radio reception to carry out the filtering of sound noise in the foregoing, operational capability of installing and electrical power pattern decide the radio reception device quantity that participates in computing in the application of reality.For instance, electronic installation can select the preferable a plurality of radio reception devices of radio reception as the radio reception device that receives main voice signal, also selects the relatively poor a plurality of radio reception devices of radio reception (may in the radio reception device array except the preferable radio reception device of radio reception remaining radio reception device) to come the noise in the main voice signal of filtering simultaneously.On the other hand, for the limited hand-held device of electric weight, also can only in the preferable radio reception device of radio reception, select the radio reception device of a radio reception the best as the radio reception device that receives main voice signal, and in the relatively poor radio reception device of radio reception, select a poorest radio reception device of radio reception to come noise in the main voice signal of filtering, reaching the purpose of simplifying operation program and power consumption, more than two kinds of methods all can reach the effect of filtering sound noise of the present invention.
And analyze to select after primary and secondary wants the radio reception device, in step S230, the second sound signal that can utilize the second radio reception device to be received comes the noise in first voice signal that the filtering first radio reception device received, and the voice signal of acquisition filtering noise.In the process of user's conversation, in order to deal with contingent at any time conversation environment change, for example the user changes the conversation posture or moves in another conversation place, and above-mentioned steps will repeatedly be carried out, and reaches optimization in the hope of speech quality.
What deserves to be mentioned is, the present invention also is included in after the filtering noise, judges further whether voice signal reaches preset standard, and dynamically adjusts the size of above-mentioned first coefficient and second coefficient according to this, so that the effect of filtering noise of the present invention reaches best, describe in detail for an embodiment in addition with next.
Fig. 3 is the method flow diagram according to the filtering sound noise shown in another embodiment of the present invention.Please refer to Fig. 3, present embodiment is applicable to as the electronic installation among the above-mentioned embodiment 100, by the received voice signal of processor 110 comprehensive a plurality of radio reception devices of electronic installation 100, and the effective noise in the filtering sound signal, its step is as follows:
At first, in step S310, utilize radio reception device array detection by a signal that source of sound is sounded, and in step S320, then further analyze detected voice signal, from radio reception device array, selecting the first preferable radio reception device of at least one radio reception, and the second relatively poor radio reception device of at least one radio reception.Above step is all same or similar with the step S210 and the S220 of previous embodiment, so its detailed content does not repeat them here.
And it should be noted that present embodiment in step S330, and also comprising need judging whether initialization coefficient a, b, this coefficient is promptly corresponding to first coefficient in the previous embodiment and second coefficient, and in order to calculate the voice signal behind the filtering noise.Wherein, when electronic installation just begins to detect voice signal, promptly need given one group of coefficient a, b earlier, carry out the calculating of follow-up voice signal for electronic installation, therefore in step S330, need initialization coefficient a, b if judge, then enter step S340, carry out the initialization operation of coefficient a, b; Otherwise if judgement does not need initialization coefficient a, b, then representative before provided and had write down optimized coefficient a, b, and can enter step S350 this moment, downloads coefficient a, the b of precedence record.
Behind initialization or downloadable coefficients a, b, can utilize this downloadable coefficients a, b to calculate voice signal, and in step S360, be about to the first voice signal Pmic and second sound signal Smic and be not multiplied by addition behind coefficient a and the coefficient b, and with the voice signal MICin of addition result as filtering noise, its relational expression is as follows:
MIC
in＝a×P
mic+b×S
mic。(1)
After filtering noise, in step S370, judge further then whether voice signal reaches preset standard.Wherein, if judge that voice signal has reached preset standard, then represent at present given coefficient a, b to reach certain standard in order to the effect of filtering noise, therefore in step S390, this can be organized in the medium that coefficient a, b be recorded in electronic installation, wait is written into use as subsequent detection during to new voice signal, and the voice signal that utilizes this group coefficient a, b to be calculated then will export electronic installation to; And if judge that voice signal does not reach preset standard, then represent previous given coefficient a, b is not the best, then enter step S380 this moment, first by electronic installation with after the voice signal output, dynamically adjust coefficient a, size back and the storage of b, after waiting the receiving next record voice signal, can follow execution in step S310 again, S320, S330, S350, S360 and S370, judge once again in step S370 equally wherein whether voice signal arrives preset standard, if still no show preset standard then enters among the step S380 output sound signal again and again to coefficient a, b dynamically adjusts; And if this voice signal has arrived preset standard, then can enter S390 directly coefficient a, b to be stored, adjust coefficient a, b later in subsequent step, can also directly take.
Yet, described as previous embodiment, in order to deal with the conversation environment change, in step S380 or step S390, after usage factor a, b calculating and output sound signal, will return step S310, reselect the first radio reception device and the second the poorest radio reception device of radio reception of radio reception the best in the radio reception device array, and the calculating voice signal, till the tonequality of this voice signal reaches preset standard.
On the other hand, Fig. 4 is the method flow diagram according to the filtering sound noise shown in further embodiment of this invention.Please refer to Fig. 4, the step S410-S470 of present embodiment is all same or similar with the content of the step S310-S370 of previous embodiment, so its detailed content does not repeat them here.Yet present embodiment has been adjusted after the size of coefficient a, b in step S480, is to return step S460, recomputates after the tonequality of voice signal reaches preset standard, just enters step S490 packing coefficient a, b and voice signal is exported.If compare with previous embodiment, present embodiment can repeat coefficient a, b to be adjusted to the quality that can make the sound news and arrive the mode of just exporting after the preset standard, can make the tonequality of voice signal not need to adjust the transitional period, and can be adjusted to the desired preset standard of user quickly through one section.
In sum, in the method for filtering sound noise of the present invention, by selecting at least one to receive the preferable radio reception device of sound signal quality as the radio reception device that receives main voice signal, and select at least one to receive the relatively poor radio reception device of sound signal quality in order to the noise in the main voice signal of filtering, therefore effective filtering sound noise.If method of the present invention is applied on the call function of electronic installation, then can significantly promote speech quality, increase the comfortableness of conversation.
Though the present invention with preferred embodiment openly as above, so it is not in order to limit the present invention.Those of ordinary skill under any in the technical field under the situation that does not break away from the spirit and scope of the present invention, can carry out various changes and modification.Therefore, protection scope of the present invention is as the criterion with the scope of the claim that proposed.
Claims (10)
1. the method for a filtering sound noise is applicable to an electronic installation, and this method comprises the following steps:
A. a voice signal that utilizes a radio reception device array received one source of sound to be sent, wherein this radio reception device array comprises a plurality of radio reception devices;
B. analyze this voice signal, select the preferable at least one first radio reception device of radio reception in this radio reception device array, and the relatively poor at least one second radio reception device of radio reception; And
C. the noise at least one first voice signal that this at least one first radio reception device of at least one second sound signal filtering that utilizes this at least one second radio reception device to be received is received, this voice signal of acquisition filtering noise.
2. the method for filtering sound noise as claimed in claim 1, wherein the radio reception device that this at least one first radio reception device is radio reception the best in this radio reception device array in this step b..
3. the method for filtering sound noise as claimed in claim 1, wherein this at least one second radio reception device is the poorest radio reception device of radio reception in this radio reception device array in this step b..
4. the method for filtering sound noise as claimed in claim 1, wherein this at least one second radio reception device comprises the radio reception device except that this at least one first radio reception device in this radio reception device array.
5. the method for filtering sound noise as claimed in claim 1, wherein this step c comprises:
Should at least one first voice signal and this at least one second sound signal be multiplied by addition behind one first coefficient and one second coefficient individually, and with addition result this voice signal as filtering noise.
6. the method for filtering sound noise as claimed in claim 5 wherein before this step c, also comprises:
Judge that this first coefficient and this second coefficient whether in before being recorded, if having, then are written into this first coefficient and this second coefficient of precedence record, if nothing, then this first coefficient of initialization and this second coefficient.
7. the method for filtering sound noise as claimed in claim 5 wherein after this step c, also comprises:
Whether the tonequality of d. judging this voice signal reaches a preset standard;
D1. if do not reach this preset standard, then adjust the size of this first coefficient and this second coefficient, and write down the size of this first coefficient and this second coefficient; And
D2. if reach this preset standard, then write down the size of this first coefficient and this second coefficient.
8. the method for filtering sound noise as claimed in claim 7 wherein after this steps d 1, also comprises:
Repeat this step c, d, till the tonequality of this voice signal reaches this preset standard.
9. the method for filtering sound noise as claimed in claim 7 wherein after this steps d 1, also comprises:
Repeat this step a-d, till the tonequality of this voice signal reaches this preset standard.
10. the method for filtering sound noise as claimed in claim 1, the method for wherein analyzing this voice signal comprise gauss hybrid models (GMM), minimize classification error (MCE), vector quantization coding (VQ) and mixed Gaussian mixed model/SVMs (GMM/SVM) one of them.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA2007101938137A CN101442696A (en) | 2007-11-21 | 2007-11-21 | Method for filtering sound noise |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA2007101938137A CN101442696A (en) | 2007-11-21 | 2007-11-21 | Method for filtering sound noise |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101442696A true CN101442696A (en) | 2009-05-27 |
Family
ID=40726907
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNA2007101938137A Pending CN101442696A (en) | 2007-11-21 | 2007-11-21 | Method for filtering sound noise |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101442696A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103248727A (en) * | 2012-02-01 | 2013-08-14 | 华为终端有限公司 | Recording device |
| CN106356074A (en) * | 2015-07-16 | 2017-01-25 | 中华映管股份有限公司 | Audio processing system and audio processing method thereof |
-
2007
- 2007-11-21 CN CNA2007101938137A patent/CN101442696A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103248727A (en) * | 2012-02-01 | 2013-08-14 | 华为终端有限公司 | Recording device |
| CN106356074A (en) * | 2015-07-16 | 2017-01-25 | 中华映管股份有限公司 | Audio processing system and audio processing method thereof |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7961891B2 (en) | System and method of audio testing of acoustic devices | |
| US10269369B2 (en) | System and method of noise reduction for a mobile device | |
| CN102770909B (en) | Voice activity detection based on multiple speech activity detectors | |
| CN101313483B (en) | Configuration of echo cancellation | |
| CN101668085B (en) | Method for regulating voice output of mobile terminal and mobile terminal | |
| JP2019159307A (en) | Terminal bracket and far-field voice interaction system | |
| CN103650533A (en) | Generating a masking signal on an electronic device | |
| CN102197422A (en) | Audio source proximity estimation using sensor array for noise reduction | |
| CN105611014A (en) | Method and device for mobile terminal call voice noise reduction | |
| CN115762579A (en) | Sound processing method, device and equipment | |
| WO2014000476A1 (en) | Voice noise reduction method and device for mobile terminal | |
| CN106384597B (en) | Audio data processing method and device | |
| CN107395883A (en) | Voice signal adjusting method, communication terminal and computer-readable recording medium | |
| CN106569773A (en) | Terminal and voice interaction processing method | |
| CN107172313A (en) | Improve method, device, mobile terminal and the storage medium of hand-free call quality | |
| CN107682553B (en) | Call signal sending method and device, mobile terminal and storage medium | |
| CN115482830A (en) | Speech enhancement method and related equipment | |
| CN107277208B (en) | Communication method, first communication device and terminal | |
| CN106331258A (en) | Call methods, call apparatus and call system | |
| CN103873978A (en) | Information processing method and electronic equipment | |
| CN101753657A (en) | Method and device for reducing call noise | |
| CN101442696A (en) | Method for filtering sound noise | |
| CN108154886A (en) | Noise suppressing method and device, electronic device and computer readable storage medium | |
| CN104639719A (en) | Calling method and communication terminal | |
| CN110996208A (en) | Wireless earphone and noise reduction method thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20090527 |