CN103177728B - Voice signal denoise processing method and device - Google Patents
Voice signal denoise processing method and device Download PDFInfo
- Publication number
- CN103177728B CN103177728B CN201110432747.0A CN201110432747A CN103177728B CN 103177728 B CN103177728 B CN 103177728B CN 201110432747 A CN201110432747 A CN 201110432747A CN 103177728 B CN103177728 B CN 103177728B
- Authority
- CN
- China
- Prior art keywords
- background noise
- voice signal
- similarity
- noise
- reduction process
- 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
Landscapes
- Mobile Radio Communication Systems (AREA)
- Noise Elimination (AREA)
Abstract
The invention discloses a kind of voice signal denoise processing method and device, the method comprising the steps of: when mobile terminal carries out voice call, obtain the voice signal that this mobile terminal produces in stipulated time section; Judge whether the voice signal obtained only comprises background noise; If the determination result is YES, then according to this voice signal obtained, background noise storehouse is upgraded; If judged result is no, then, in each background noise of background noise library storage, choose the background noise carrying out noise reduction process, and according to the background noise selected, noise reduction process is carried out to the voice signal obtained.When technical solution of the present invention solves that in prior art, mobile terminal carries out voice signal noise reduction process according to the background noise prestored, good noise reduction process effect can not be produced, have impact on the problem of the voice call quality of user.
Description
Technical field
The present invention relates to voice communication technology field, particularly relate to voice signal denoise processing method and device.
Background technology
Along with the development of the communications industry, utilize mobile terminal (such as mobile phone) to carry out voice call to become the more and more important communication of people, exchange way, because people usually can be in more noisy environment, such as user is on bus, on road side or in market, if user uses mobile terminal to carry out voice call under this noisy environment, then usually can be subject to the interference of noise, such as, when user A is positioned at road side, voice call is carried out by mobile terminal and user B, in the voice signal that user B receives, not only comprise the signal of normal talking, the noise signal produced when simultaneously also can comprise vehicle travels, this noise signal can be called background noise.Using mobile terminal to carry out the speech quality of voice call to improve user, needing to carry out carrying out noise reduction process to voice signal in the process of voice call at mobile terminal.
In prior art, carry out at mobile terminal being specially the method that voice signal carries out noise reduction process in the process of voice call:
The background noise that some are common is prestored in mobile terminal, after this mobile terminal collects the voice signal of user, before the voice signal collected being sent to call opposite end, first in each background noise prestored, select the background noise similar to the voice signal collected, then according to the background noise selected, noise reduction process is carried out to the voice signal collected, the background noise that namely cancellation selects in the voice signal collected.
Therefore, in prior art, mobile terminal is when carrying out noise reduction process to the voice signal collected, in each background noise prestored, select the background noise similar to the voice signal collected, then according to the background noise selected, noise reduction process is carried out to the voice signal collected, this just may occur not comprising the background noise similar to the voice signal collected in each background noise prestored in mobile terminal, thus mobile terminal cannot select suitable background noise carries out noise reduction process situation to the voice signal collected, when therefore adopting the method for prior art to carry out noise reduction process to voice signal, good noise reduction process effect can not be produced, have impact on the voice call quality of user.
Summary of the invention
The embodiment of the present invention provides a kind of voice signal denoise processing method and device, during in order to solve that in prior art, mobile terminal carries out voice signal noise reduction process according to the background noise prestored, good noise reduction process effect can not be produced, have impact on the problem of the voice call quality of user.
Embodiment of the present invention technical scheme is as follows:
A kind of voice signal denoise processing method, the method comprises: when mobile terminal carries out voice call, obtain the voice signal that this mobile terminal produces in stipulated time section; Judge whether the voice signal obtained only comprises background noise; If the determination result is YES, then according to this voice signal obtained, background noise storehouse is upgraded; If judged result is no, then, in each background noise of background noise library storage, choose the background noise carrying out noise reduction process, and according to the background noise selected, noise reduction process is carried out to the voice signal obtained.
A kind of voice signal noise reduction process device, comprising: voice signal obtains unit, for when mobile terminal carries out voice call, obtains the voice signal that this mobile terminal produces in stipulated time section; Background noise judging unit, for judging whether the voice signal that voice signal acquisition unit obtains only comprises background noise; Background noise storehouse updating block, for when the judged result of background noise judging unit is for being, this voice signal obtaining unit acquisition according to voice signal upgrades background noise storehouse; Background noise chooses unit, for when the judged result of background noise judging unit is no, in each background noise of background noise library storage, chooses the background noise carrying out noise reduction process; Noise reduction processing unit, the background noise gone out for choosing unit selection according to background noise carries out noise reduction process to the voice signal that voice signal obtains unit acquisition.
In embodiment of the present invention technical scheme, when carrying out voice signal noise reduction process, not directly carry out voice signal noise reduction process according to the background noise prestored in mobile terminal, but when mobile terminal carries out voice call, first obtain the voice signal that this mobile terminal produces within a period of time, then judge whether the voice signal obtained only comprises background noise, if the determination result is YES, then according to this voice signal, background noise storehouse is upgraded, if judged result is no, then in each background noise of background noise library storage, choose the background noise carrying out noise reduction process, and according to the background noise selected, noise reduction process is carried out to the voice signal obtained.Therefore, embodiment of the present invention technical scheme chooses the background noise carrying out noise reduction process in each background noise of background noise library storage, the background noise wherein stored in background noise storehouse is when carrying out noise reduction process to the voice signal of mobile terminal, real-time update is carried out according to voice signal, therefore efficiently avoid in each background noise prestored in mobile terminal and do not comprise the background noise similar to the voice signal collected, thus mobile terminal cannot select suitable background noise carries out noise reduction process situation to the voice signal collected, therefore, it is possible to obtain good noise reduction process effect, improve the voice call quality of user.
Accompanying drawing explanation
Fig. 1 is in the embodiment of the present invention one, voice signal denoise processing method schematic flow sheet;
Fig. 2 is in the embodiment of the present invention two, voice signal denoise processing method specific implementation schematic flow sheet;
Fig. 3 is in the embodiment of the present invention two, voice signal noise reduction process apparatus structure schematic diagram.
Embodiment
Below in conjunction with each accompanying drawing, the main of embodiment of the present invention technical scheme is realized principle, embodiment and set forth in detail the beneficial effect that should be able to reach.
Embodiment one
As shown in Figure 1, be the voice signal denoise processing method schematic flow sheet that the embodiment of the present invention one proposes, its concrete treatment scheme is as follows:
Step 11, when mobile terminal carries out voice call, obtains the voice signal that this mobile terminal produces in stipulated time section.
Wherein, obtain the voice signal that this mobile terminal produces in stipulated time section, can carry out when arriving noise reduction process time point, stipulated time section can be the time period that this noise reduction process time point is corresponding.
In the embodiment of the present invention one, equipment voice signal being carried out to noise reduction process can be mobile terminal, that wherein can be produced self by voice signal sender mobile terminal, the to be sent voice signal of call opposite end of giving carries out noise reduction process, such as, mobile terminal is when carrying out voice call with other mobile terminal, after producing voice signal, before voice signal being sent to voice signal take over party mobile terminal, first noise reduction process is carried out to voice signal; Also noise reduction process can be carried out by voice signal take over party mobile terminal to the voice signal that voice signal sender mobile terminal produces, such as, mobile terminal is when carrying out voice call with other mobile terminal, after receiving the voice signal (this voice signal is produced by voice signal sender mobile terminal) of voice signal sender mobile terminal transmission, before the voice signal received is played to user, first noise reduction process is carried out to the voice signal received.
In addition, in the embodiment of the present invention one, equipment voice signal being carried out to noise reduction process can also be network equipment, when carrying out voice call between mobile terminal, this network equipment obtains the voice signal that voice signal sender mobile terminal sends to voice signal take over party mobile terminal, then noise reduction process is carried out to the voice signal obtained, the voice signal after noise reduction process is sent to voice signal take over party mobile terminal.
Wherein, this network equipment can be placed in base station, and also can arrange separately, the present invention is not specifically limited this.
Due to mobile terminal carry out voice call time, duration may be longer, in order to ensure the continuity of voice call, therefore need in voice call process, every length at regular intervals, noise reduction process will be carried out to the voice signal in this time span, and then play to user or send to voice signal take over party mobile terminal.Wherein, in voice call process, can be, but not limited to periodically carry out noise reduction process to the voice signal in this cycle.
In the embodiment of the present invention one, noise reduction process frequency (such as execution per second voice signal noise reduction process several times) can be preset, from mobile terminal carries out voice call, each noise reduction process time point just can be determined based on the noise reduction process frequency preset, when arriving each noise reduction process time point, time period between current noise reduction process time point and a upper noise reduction process time point, be the time period that current noise reduction process time point is corresponding, such as, the noise reduction process frequency preset is 10 times/second, represent execution per second 10 voice signal noise reduction process, the sart point in time that mobile terminal carries out voice call is 9 points, based on the noise reduction process frequency preset, the corresponding relation of each noise reduction process time point and corresponding time period can be obtained as shown in Table 1.
Table one:
| Noise reduction process time point | Time period |
| 9: 0.1 seconds | 0.1 second, 9 point ~ 9 |
| 9: 0.2 seconds | 9: 0.1 seconds ~ 9: 0.2 seconds |
| 9: 0.3 seconds | 9: 0.2 seconds ~ 9: 0.3 seconds |
| … | … |
When arriving each noise reduction process time point in table one, just perform the operation voice signal in the corresponding time period being carried out to noise reduction process.
Step 12, judges whether the voice signal obtained only comprises background noise.
When judging whether the voice signal obtained only comprises background noise, first can judge whether the sound spectrum pitch maximal value of voice signal obtained is less than the sound spectrum threshold values that presets (such as, the sound spectrum threshold values preset is 30 decibels), if the determination result is YES, then can confirm that the voice signal obtained only comprises background noise, if judged result is no, then can confirm that the voice signal obtained not only comprises background noise.
Step 13, if the judged result of step 12 is yes, then upgrades background noise storehouse according to this voice signal obtained.
Owing to judging that the voice signal only comprising background noise may for the background noise accidentally occurred, the frequency that this background noise occurs in the communication process of mobile terminal is lower, now without the need to upgrading background noise storehouse according to this background noise, for this situation, the embodiment of the present invention one proposes, background noise occurrence number threshold values can be preset (such as, be set to 3 times), after judging that the voice signal obtained only comprises background noise, first determine in the time period between stipulated time section (namely this noise reduction process time point and described mobile terminal carry out the sart point in time of voice call), the occurrence number of this voice signal, then judge whether the occurrence number of this voice signal is greater than the background noise occurrence number threshold values preset, if the determination result is YES, then carry out operation background noise storehouse upgraded according to this voice signal obtained, if judged result is no, then do not process.
In addition, judge the voice signal only comprising background noise, also be likely stored in background noise storehouse, or store the background noise similar to this voice signal in background noise storehouse, at this moment just can not upgrade background noise storehouse, for this situation, the embodiment of the present invention one proposes, can after judging that the voice signal obtained only comprises background noise, first judge in background noise storehouse, whether to store this voice signal of acquisition or the background noise similar to the voice signal obtained, if judged result is no, then carry out operation background noise storehouse upgraded according to this voice signal obtained, if the determination result is YES, then do not process.Concrete:
After judging that the voice signal obtained only comprises background noise, determine the similarity of each background noise stored in this voice signal of obtaining and background noise storehouse respectively, then judge whether each similarity determined all is less than the first similarity threshold preset, if the determination result is YES, then think this voice signal or the background noise similar to the voice signal obtained of not storing acquisition in background noise storehouse, therefore this voice signal obtained is added in background noise storehouse, if judged result is no, then think this voice signal or the background noise similar to the voice signal obtained of having stored acquisition in background noise storehouse, therefore do not process.
In the embodiment of the present invention one, the background noise often occurred due to zones of different may be different, such as, the background noise often occurred in market is noisy sound of speech, the other background noise often occurred of road is the tucket of automobile, therefore, when preserving background noise in background noise storehouse, can by background noise and corresponding call area corresponding stored, as shown in Table 2.
Table two:
When now background noise storehouse being upgraded according to this voice signal obtained, call area (specifically can be set as the call area residing when this noise reduction process time point carries out voice call) residing when first can determine that described mobile terminal (i.e. voice signal sender mobile terminal) carries out voice call, and in each background noise of background noise library storage, determine each background noise that described call area is corresponding, then the similarity between this voice signal of acquisition and each background noise determined is determined respectively, judge whether each similarity determined all is less than the third phase preset and seemingly spends threshold value, if the determination result is YES, then this voice signal obtained is added in the background noise that described in background noise storehouse, call area is corresponding.
Wherein, when voice call can be carried out according to mobile terminal in call area, residing base station is determined, presets the corresponding relation between call area and base station, as shown in Table 3.
Table three:
The situation of background noise is only comprised for the voice signal obtained, owing to wherein not comprising sound of conversing normally, so can not noise reduction process be carried out to this voice signal, play-over to user or send to voice signal take over party mobile terminal, also the denoise processing method directly eliminated by this voice signal can be adopted, after this voice signal is directly eliminated, user can produce the illusion of dropped calls, after this voice signal can being eliminated, add white noise signal, then this white noise signal played to user or send to voice signal take over party mobile terminal.
Step 14, if the judged result of step 12 is no, then, in each background noise of background noise library storage, chooses the background noise carrying out noise reduction process, and carries out noise reduction process according to the background noise selected to the voice signal obtained.
Because the background noise stored in background noise storehouse has a lot, may be only had one of them or be severally only background noise voice signal being carried out to noise reduction process, so when carrying out noise reduction process to voice signal, in each background noise of background noise library storage, the background noise carrying out noise reduction process to be selected.
When choosing the background noise carrying out noise reduction process, arbitrarily can choose a defined amount background noise, also the similarity between each background noise of background noise library storage and this voice signal of acquisition can first be calculated respectively, then in each background noise of background noise library storage, the similarity of correspondence is greater than the background noise of the second similarity threshold preset, is chosen for the background noise carrying out noise reduction process.
If when preserving background noise in background noise storehouse, by background noise and call area corresponding stored, then in each background noise of background noise library storage, when choosing the background noise carrying out noise reduction process, only need choose in each background noise that the call area of mobile terminal is corresponding, concrete:
Call area (being specifically as follows the call area residing when this noise reduction process time point carries out voice call) residing when first determining that described mobile terminal carries out voice call, then in each background noise of background noise library storage, determine each background noise that described call area is corresponding, again in each background noise determined, choose the background noise carrying out noise reduction process.
Wherein, in each background noise determined, when choosing the background noise carrying out noise reduction process, arbitrarily can choose a defined amount background noise, also the similarity between each background noise and this voice signal of acquisition determined can first be calculated respectively, then, in each background noise determined, the similarity calculated is greater than the background noise of the 4th similarity threshold preset, is chosen for the background noise carrying out noise reduction process.
Due in background noise storehouse, the quantity of the background noise corresponding to same call area may be a lot, in order to reduce calculated amount, to accelerate noise reduction process speed, the embodiment of the present invention one proposes, in each background noise determined, when choosing the background noise carrying out noise reduction process, first in each background noise determined, occurrence number is selected to meet pre-conditioned background noise, then the similarity between each background noise and this voice signal of acquisition selected is calculated respectively, again in each background noise selected, the similarity calculated is greater than the background noise of the 5th similarity threshold preset, be chosen for the background noise carrying out noise reduction process.
Wherein, described pre-conditioned can be, but not limited to for: occurrence number is greater than predetermined threshold value, or the top n that occurrence number is maximum, and N is the number preset.
In the embodiment of the present invention one, above-mentioned first similarity threshold, the second similarity threshold, third phase seemingly spend threshold value, the 4th similarity threshold can be identical with the 5th similarity threshold, also can be different, and the present invention is not specifically limited this.
Due to the background noise selected from background noise storehouse, each voice signal that mobile terminal carries out judging in voice call process only to comprise background noise may not be comprised, and each voice signal that mobile terminal carries out judging in voice call process only to comprise background noise can be used for carrying out noise reduction process, therefore, the embodiment of the present invention one proposes, when noise reduction process being carried out to the voice signal obtained according to the background noise selected, can first determine before this voice signal of acquisition (be specifically as follows this noise reduction process time point and described mobile terminal starts to carry out in the time period between the time point of voice call), judge each voice signal only comprising background noise, then in each voice signal determined, calculate the similarity between any two voice signals, again according to the similarity calculated, merging treatment is carried out to each voice signal determined, and according to the voice signal obtained after merging treatment and the background noise selected, noise reduction process is carried out to the voice signal obtained.
Wherein, according to the similarity calculated, the method that merging treatment carries out to each voice signal determined can be, but not limited to for: judge whether the similarity calculated is greater than default value, if judged result is for being greater than, then retain any one in two voice signals, as the voice signal obtained after merging treatment.
Because the background noise carrying out noise reduction process selected may have multiple, so when carrying out noise reduction process, noise reduction process can be carried out successively according to each background noise.
When noise reduction process being carried out to the voice signal obtained according to background noise, the method of directly being subtracted each other at the sound spectrum pitch in corresponding moment with this background noise at the sound spectrum pitch in each moment by this voice signal can be taked, such as, this voice signal is 12 decibels at the sound spectrum pitch of 0.01 second, this background noise is 2 decibels at the sound spectrum pitch of 0.01 second, then after noise reduction process this voice signal be 10 decibels at the sound spectrum pitch of 0.01 second.
In order to the sound spectrum pitch of the voice signal after noise reduction process can not be too low, after can taking often to carry out noise reduction process according to a background noise, again the sound spectrum pitch in each for this voice signal moment is multiplied by a numerical value, numerical value before making its sound spectrum pitch maximal value return to noise reduction process, or the numerical value before making the mean value of the sound spectrum pitch in this voice signal each moment return to noise reduction process.
From above-mentioned processing procedure, in embodiment of the present invention technical scheme, when carrying out voice signal noise reduction process, not directly carry out voice signal noise reduction process according to the background noise prestored in mobile terminal, but when mobile terminal carries out voice call, first obtain the voice signal that this mobile terminal produces within a period of time, then judge whether the voice signal obtained only comprises background noise, only background noise is comprised if judge, then according to this voice signal, background noise storehouse is upgraded, not only background noise is comprised if judge, then in each background noise of background noise library storage, choose the background noise carrying out noise reduction process, and according to the background noise selected, noise reduction process is carried out to the voice signal obtained.Therefore, embodiment of the present invention technical scheme chooses the background noise carrying out noise reduction process in each background noise of background noise library storage, the background noise wherein stored in background noise storehouse is when carrying out noise reduction process to the voice signal of mobile terminal, real-time update is carried out according to voice signal, therefore efficiently avoid in each background noise prestored in mobile terminal and do not comprise the background noise similar to the voice signal collected, thus mobile terminal cannot select suitable background noise carries out noise reduction process situation to the voice signal collected, therefore, it is possible to obtain good noise reduction process effect, improve the voice call quality of user.
Provide embodiment specifically below.
Embodiment two
As shown in Figure 2, be the voice signal denoise processing method specific implementation schematic flow sheet that the embodiment of the present invention two proposes, its concrete treatment scheme is as follows:
Step 21, when mobile terminal carries out voice call, if arrive noise reduction process time point, then obtains the voice signal that this mobile terminal produces within the time period that this noise reduction process time point is corresponding.
Step 22, judges whether the voice signal obtained only comprises background noise.
Step 23, if judge, the voice signal obtained only comprises background noise, then determine the call area that described mobile terminal (i.e. voice signal sender mobile terminal) is residing when this noise reduction process time point carries out voice call.
Step 24, background noise corresponding to the call area determined step 23 in background noise storehouse according to this voice signal obtained upgrades.
Step 25, if judge, the voice signal obtained not only comprises background noise, then determine the call area that described mobile terminal (i.e. voice signal sender mobile terminal) is residing when this noise reduction process time point carries out voice call.
Step 26, in each background noise that the call area that the step 25 of background noise library storage is determined is corresponding, determines each background noise that described call area is corresponding.
Step 27, in each background noise that step 26 is determined, chooses the background noise carrying out noise reduction process.
Step 28, carries out noise reduction process according to the background noise that step 27 selects to the voice signal obtained.
Step 29, plays to user by the voice signal after noise reduction process or sends to voice signal take over party mobile terminal.
Embodiment three
Accordingly, the embodiment of the present invention three provides a kind of voice signal noise reduction process device, and its structure as shown in Figure 3, comprising:
Voice signal obtains unit 31, for when mobile terminal carries out voice call, obtains the voice signal that this mobile terminal produces in stipulated time section;
Background noise judging unit 32, for judging whether the voice signal that voice signal acquisition unit 31 obtains only comprises background noise;
Background noise storehouse updating block 33, for when the judged result of background noise judging unit 32 is for being, this voice signal obtaining unit 31 acquisition according to voice signal upgrades background noise storehouse;
Background noise chooses unit 34, for when the judged result of background noise judging unit 32 is no, in each background noise of background noise library storage, chooses the background noise carrying out noise reduction process;
Noise reduction processing unit 35, the background noise selected for choosing unit 34 according to background noise carries out noise reduction process to the voice signal that voice signal obtains unit 31 acquisition.
Preferably, background noise judging unit 32 specifically comprises:
Sound spectrum pitch judgment sub-unit, for judging whether the sound spectrum pitch maximal value of the voice signal that voice signal acquisition unit 31 obtains is less than the sound spectrum threshold values preset;
Background noise confirms subelement, for when the judged result of sound spectrum pitch judgment sub-unit is for being, confirm that the voice signal that voice signal acquisition unit 31 obtains only comprises background noise, and when the judged result of sound spectrum pitch judgment sub-unit is no, confirm that the voice signal that voice signal acquisition unit 31 obtains not only comprises background noise.
Preferably, background noise storehouse updating block 33 specifically comprises:
First similarity computation subunit, obtains for calculating voice signal respectively the similarity of each background noise stored in the voice signal and background noise storehouse that unit 31 obtains;
First similarity judgment sub-unit, for judging whether each similarity that the first similarity computation subunit calculates all is less than the first similarity threshold preset;
First voice signal adds subelement, for when the judged result of the first similarity judgment sub-unit is for being, voice signal is obtained the voice signal that unit 31 obtains and being added in background noise storehouse.
Preferably, background noise is chosen unit 34 and is specifically comprised:
Second similarity computation subunit, each background noise and voice signal for calculating background noise library storage respectively obtain the similarity between the voice signal of unit 31 acquisition;
First background noise chooses subelement, for in each background noise of background noise library storage, the similarity of the correspondence the second similarity computation subunit calculated is greater than the background noise of the second similarity threshold preset, and is chosen for the background noise carrying out noise reduction process.
Preferably, background noise storehouse updating block 33 specifically comprises:
Subelement is determined in first call area, call area residing during for determining that described mobile terminal carries out voice call;
First background noise determination subelement, in each background noise of background noise library storage, determines each background noise that call area that the first call area determines that subelement is determined is corresponding;
Third phase seemingly spends computation subunit, the similarity between each background noise that voice signal and the first background noise determination subelement for calculating the acquisition of voice signal acquisition unit 31 are respectively determined;
Second similarity judgment sub-unit, for judging whether third phase is all less than like each similarity that degree computation subunit calculates the third phase preset and seemingly spends threshold value;
Second voice signal adds subelement, for when the judged result of the second similarity judgment sub-unit is for being, voice signal is obtained voice signal that unit 31 obtains and being added in background noise corresponding to the call area of determining that subelement is determined in the first call area in background noise storehouse.
More preferably, background noise is chosen unit and is specifically comprised:
Subelement is determined in second call area, call area residing during for determining that described mobile terminal carries out voice call;
Second background noise determination subelement, in each background noise of background noise library storage, determines each background noise that call area that the second call area determines that subelement is determined is corresponding;
Second background noise chooses subelement, in each background noise of determining at the second background noise determination subelement, chooses the background noise carrying out noise reduction process.
More preferably, the second background noise is chosen subelement and is specifically comprised:
First similarity calculation module, obtains similarity between voice signal that unit 31 obtains for calculating each background noise that the second background noise determination subelement determines and voice signal respectively;
First background noise chooses module, for in each background noise of determining at the second background noise determination subelement, the similarity that first similarity calculation module calculates is greater than the background noise of the 4th similarity threshold preset, is chosen for the background noise carrying out noise reduction process.
More preferably, the second background noise is chosen subelement and is specifically comprised:
Second background noise chooses module, in each background noise of determining at the second background noise determination subelement, selects occurrence number to meet pre-conditioned background noise;
Second similarity calculation module, chooses each background noise that model choice goes out and voice signal obtain similarity between voice signal that unit 31 obtains for calculating the second background noise respectively;
3rd background noise chooses module, for choosing in each background noise that model choice goes out at the second background noise, the similarity that second similarity calculation module calculates is greater than the background noise of the 5th similarity threshold preset, is chosen for the background noise carrying out noise reduction process.
Preferably, noise reduction processing unit 35 specifically comprises:
3rd background noise determination subelement, for determining before this voice signal of acquisition, each voice signal of what background noise judging unit 32 was judged only comprise background noise;
4th Similarity Measure subelement, in each voice signal of determining at the 3rd background noise determination subelement, calculates the similarity between any two voice signals;
Merging treatment subelement, for the similarity gone out according to the 4th Similarity Measure subunit computes, carries out merging treatment to each voice signal that the 3rd background noise determination subelement is determined;
Noise reduction process subelement, for choosing the background noise that unit 34 selects according to the voice signal obtained after merging treatment subelement merging treatment and background noise, voice signal voice signal being obtained to unit 31 acquisition carries out noise reduction process.
Although describe the preferred embodiments of the present invention, those skilled in the art once obtain the basic creative concept of cicada, then can make other change and amendment to these embodiments.So claims are intended to be interpreted as comprising preferred embodiment and falling into all changes and the amendment of the scope of the invention.
Obviously, those skilled in the art can carry out various change and modification to the present invention and not depart from the spirit and scope of the present invention.Like this, if these amendments of the present invention and modification belong within the scope of the claims in the present invention and equivalent technologies thereof, then the present invention is also intended to comprise these change and modification.
Claims (16)
1. a voice signal denoise processing method, is characterized in that, comprising:
When mobile terminal carries out voice call, obtain the voice signal that this mobile terminal produces in stipulated time section;
Judge whether the voice signal obtained only comprises background noise;
If judge, the described voice signal obtained only comprises background noise, then upgrade background noise storehouse according to this voice signal obtained, specifically comprise: the similarity calculating each background noise stored in this voice signal of acquisition and background noise storehouse respectively; Judge whether each similarity calculated all is less than the first similarity threshold preset; If judge, each similarity calculated all is less than described first similarity threshold preset, be then added in background noise storehouse by this voice signal obtained;
If judge, the described voice signal obtained not only comprises background noise, then, in each background noise of background noise library storage, choose the background noise carrying out noise reduction process, and carries out noise reduction process according to the background noise selected to the voice signal obtained.
2. the method for claim 1, is characterized in that, judges whether the voice signal obtained only comprises background noise, specifically comprises:
Judge whether the sound spectrum pitch maximal value of the voice signal obtained is less than the sound spectrum threshold values preset;
If the determination result is YES, then confirm that the voice signal obtained only comprises background noise;
If judged result is no, then confirm that the voice signal obtained not only comprises background noise.
3. the method for claim 1, is characterized in that, in each background noise of background noise library storage, chooses the background noise carrying out noise reduction process, specifically comprises:
Calculate the similarity between each background noise of background noise library storage and this voice signal of acquisition respectively;
In each background noise of background noise library storage, the similarity of correspondence is greater than the background noise of the second similarity threshold preset, is chosen for the background noise carrying out noise reduction process.
4. the method for claim 1, is characterized in that, this voice signal according to obtaining upgrades background noise storehouse, specifically comprises:
Call area residing when determining that described mobile terminal carries out voice call; And
In each background noise of background noise library storage, determine each background noise that described call area is corresponding;
Calculate the similarity between this voice signal of acquisition and each background noise determined respectively;
Judge whether each similarity calculated all is less than the third phase preset and seemingly spends threshold value;
If the determination result is YES, then this voice signal obtained is added in the background noise that described in background noise storehouse, call area is corresponding.
5. method as claimed in claim 4, is characterized in that, in each background noise of background noise library storage, choose the background noise carrying out noise reduction process, specifically comprise:
Call area residing when determining that described mobile terminal carries out voice call; And
In each background noise of background noise library storage, determine each background noise that described call area is corresponding;
In each background noise determined, choose the background noise carrying out noise reduction process.
6. method as claimed in claim 5, is characterized in that, in each background noise determined, choose the background noise carrying out noise reduction process, specifically comprise:
Calculate the similarity between each background noise and this voice signal of acquisition determined respectively;
In each background noise determined, the similarity calculated is greater than the background noise of the 4th similarity threshold preset, is chosen for the background noise carrying out noise reduction process.
7. method as claimed in claim 5, is characterized in that, in each background noise determined, choose the background noise carrying out noise reduction process, specifically comprise:
In each background noise determined, occurrence number is selected to meet pre-conditioned background noise;
Calculate the similarity between each background noise and this voice signal of acquisition selected respectively;
In each background noise selected, the similarity calculated is greater than the background noise of the 5th similarity threshold preset, is chosen for the background noise carrying out noise reduction process.
8. the method for claim 1, is characterized in that, carries out noise reduction process, specifically comprise according to the background noise selected to the voice signal obtained:
Determine, before this voice signal of acquisition, to judge each voice signal only comprising background noise;
In each voice signal determined, calculate the similarity between any two voice signals;
According to the similarity calculated, merging treatment is carried out to each voice signal determined;
According to the voice signal obtained after merging treatment and the background noise selected, noise reduction process is carried out to the voice signal obtained.
9. a voice signal noise reduction process device, is characterized in that, comprising:
Voice signal obtains unit, for when mobile terminal carries out voice call, obtains the voice signal that this mobile terminal produces in stipulated time section;
Background noise judging unit, for judging whether the voice signal that voice signal acquisition unit obtains only comprises background noise;
Background noise storehouse updating block, for when the judged result of background noise judging unit is for being, this voice signal obtaining unit acquisition according to voice signal upgrades background noise storehouse, specifically comprising: the first similarity computation subunit, obtaining for calculating voice signal respectively the similarity of each background noise stored in the voice signal and background noise storehouse that unit obtains; First similarity judgment sub-unit, for judging whether each similarity that the first similarity computation subunit calculates all is less than the first similarity threshold preset; First voice signal adds subelement, and for when the judged result of the first similarity judgment sub-unit is for being, voice signal voice signal being obtained unit acquisition is added in background noise storehouse;
Background noise chooses unit, for when the judged result of background noise judging unit is no, in each background noise of background noise library storage, chooses the background noise carrying out noise reduction process;
Noise reduction processing unit, the background noise gone out for choosing unit selection according to background noise carries out noise reduction process to the voice signal that voice signal obtains unit acquisition.
10. device as claimed in claim 9, it is characterized in that, background noise judging unit specifically comprises:
Sound spectrum pitch judgment sub-unit, for judging whether the sound spectrum pitch maximal value of the voice signal that voice signal acquisition unit obtains is less than the sound spectrum threshold values preset;
Background noise confirms subelement, for when the judged result of sound spectrum pitch judgment sub-unit is for being, confirm that the voice signal that voice signal acquisition unit obtains only comprises background noise, and when the judged result of sound spectrum pitch judgment sub-unit is no, confirm that the voice signal that voice signal acquisition unit obtains not only comprises background noise.
11. devices as claimed in claim 9, it is characterized in that, background noise is chosen unit and is specifically comprised:
Second similarity computation subunit, each background noise and voice signal for calculating background noise library storage respectively obtain the similarity between the voice signal of unit acquisition;
First background noise chooses subelement, for in each background noise of background noise library storage, the similarity of the correspondence the second similarity computation subunit calculated is greater than the background noise of the second similarity threshold preset, and is chosen for the background noise carrying out noise reduction process.
12. devices as claimed in claim 9, it is characterized in that, background noise storehouse updating block specifically comprises:
Subelement is determined in first call area, call area residing during for determining that described mobile terminal carries out voice call;
First background noise determination subelement, in each background noise of background noise library storage, determines each background noise that call area that the first call area determines that subelement is determined is corresponding;
Third phase seemingly spends computation subunit, the similarity between each background noise that voice signal and the first background noise determination subelement for calculating the acquisition of voice signal acquisition unit are respectively determined;
Second similarity judgment sub-unit, for judging whether third phase is all less than like each similarity that degree computation subunit calculates the third phase preset and seemingly spends threshold value;
Second voice signal adds subelement, for when the judged result of the second similarity judgment sub-unit is for being, voice signal is obtained the voice signal that unit obtains and being added in background noise corresponding to the call area of determining that subelement is determined in the first call area in background noise storehouse.
13. devices as claimed in claim 12, it is characterized in that, background noise is chosen unit and is specifically comprised:
Subelement is determined in second call area, call area residing during for determining that described mobile terminal carries out voice call;
Second background noise determination subelement, in each background noise of background noise library storage, determines each background noise that call area that the second call area determines that subelement is determined is corresponding;
Second background noise chooses subelement, in each background noise of determining at the second background noise determination subelement, chooses the background noise carrying out noise reduction process.
14. devices as claimed in claim 13, it is characterized in that, the second background noise is chosen subelement and is specifically comprised:
First similarity calculation module, obtains similarity between voice signal that unit obtains for calculating each background noise that the second background noise determination subelement determines and voice signal respectively;
First background noise chooses module, for in each background noise of determining at the second background noise determination subelement, the similarity that first similarity calculation module calculates is greater than the background noise of the 4th similarity threshold preset, is chosen for the background noise carrying out noise reduction process.
15. devices as claimed in claim 13, it is characterized in that, the second background noise is chosen subelement and is specifically comprised:
Second background noise chooses module, in each background noise of determining at the second background noise determination subelement, selects occurrence number to meet pre-conditioned background noise;
Second similarity calculation module, chooses each background noise that model choice goes out and voice signal obtain similarity between voice signal that unit obtains for calculating the second background noise respectively;
3rd background noise chooses module, for choosing in each background noise that model choice goes out at the second background noise, the similarity that second similarity calculation module calculates is greater than the background noise of the 5th similarity threshold preset, is chosen for the background noise carrying out noise reduction process.
16. devices as claimed in claim 9, it is characterized in that, noise reduction processing unit specifically comprises:
3rd background noise determination subelement, for determining before this voice signal of acquisition, each voice signal of what background noise judging unit was judged only comprise background noise;
4th Similarity Measure subelement, in each voice signal of determining at the 3rd background noise determination subelement, calculates the similarity between any two voice signals;
Merging treatment subelement, for the similarity gone out according to the 4th Similarity Measure subunit computes, carries out merging treatment to each voice signal that the 3rd background noise determination subelement is determined;
Noise reduction process subelement, for choosing the background noise that unit selection goes out according to the voice signal obtained after merging treatment subelement merging treatment and background noise, voice signal voice signal being obtained to unit acquisition carries out noise reduction process.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110432747.0A CN103177728B (en) | 2011-12-21 | 2011-12-21 | Voice signal denoise processing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110432747.0A CN103177728B (en) | 2011-12-21 | 2011-12-21 | Voice signal denoise processing method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103177728A CN103177728A (en) | 2013-06-26 |
| CN103177728B true CN103177728B (en) | 2015-07-29 |
Family
ID=48637534
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110432747.0A Active CN103177728B (en) | 2011-12-21 | 2011-12-21 | Voice signal denoise processing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103177728B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108171118A (en) * | 2017-12-05 | 2018-06-15 | 东软集团股份有限公司 | Signal of blinking data processing method, device, readable storage medium storing program for executing and electronic equipment |
| TWI640982B (en) * | 2017-07-21 | 2018-11-11 | 宏達國際電子股份有限公司 | Method, device, and non-transitory computer readable storage medium for processing audio signal |
Families Citing this family (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104378774A (en) * | 2013-08-15 | 2015-02-25 | 中兴通讯股份有限公司 | Voice quality processing method and device |
| CN106971714A (en) * | 2016-01-14 | 2017-07-21 | 芋头科技（杭州）有限公司 | A kind of speech de-noising recognition methods and device applied to robot |
| CN106971720A (en) * | 2016-01-14 | 2017-07-21 | 芋头科技（杭州）有限公司 | A kind of robot voice recognition methods for updating noise database and device |
| CN106971715A (en) * | 2016-01-14 | 2017-07-21 | 芋头科技（杭州）有限公司 | A kind of speech recognition equipment applied to robot |
| CN106971716A (en) * | 2016-01-14 | 2017-07-21 | 芋头科技（杭州）有限公司 | A kind of robot noise database updates and speech recognition equipment, method |
| CN108520754B (en) * | 2018-04-09 | 2021-01-12 | 广东思派康电子科技有限公司 | Noise reduction conference machine |
| CN108766453A (en) * | 2018-05-24 | 2018-11-06 | 江西午诺科技有限公司 | Voice de-noising method, device, readable storage medium storing program for executing and mobile terminal |
| CN109346099B (en) * | 2018-12-11 | 2022-02-08 | 珠海一微半导体股份有限公司 | Iterative denoising method and chip based on voice recognition |
| CN110198374A (en) * | 2019-05-30 | 2019-09-03 | 深圳市趣创科技有限公司 | A kind of mobile phone speech noise-reduction method and device based on error correction learning rules |
| CN111508474B (en) * | 2019-08-08 | 2021-04-06 | 马上消费金融股份有限公司 | Voice interruption method, electronic equipment and storage device |
| CN111698232B (en) * | 2020-06-03 | 2021-09-10 | 腾讯科技（深圳）有限公司 | Data processing method, data processing device, computer equipment and storage medium |
| CN113053404A (en) * | 2021-03-22 | 2021-06-29 | 三一重机有限公司 | Method and device for interaction between inside and outside of cab |
| CN114882482B (en) * | 2021-10-12 | 2023-03-17 | 北京九章云极科技有限公司 | Seal anti-counterfeiting identification method and device |
| CN117351993B (en) * | 2023-12-04 | 2024-02-13 | 方图智能(深圳)科技集团股份有限公司 | Audio transmission quality evaluation method and system based on audio distribution |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1390349A (en) * | 1999-11-15 | 2003-01-08 | 诺基亚有限公司 | Noise suppression |
| CN1447963A (en) * | 2000-08-21 | 2003-10-08 | 康奈克森特***公司 | Method for noise robust classification in speech coding |
| CN101056322A (en) * | 2006-04-13 | 2007-10-17 | 中兴通讯股份有限公司 | A device and method for overlapping the background sound at the mobile communication terminal |
| CN101197870A (en) * | 2006-12-05 | 2008-06-11 | 乐金电子(中国)研究开发中心有限公司 | Mobile terminal with adjustable speech quality |
| US7680653B2 (en) * | 2000-02-11 | 2010-03-16 | Comsat Corporation | Background noise reduction in sinusoidal based speech coding systems |
| CN101989431A (en) * | 2009-07-31 | 2011-03-23 | 深圳富泰宏精密工业有限公司 | Device and method for eliminating noise |
-
2011
- 2011-12-21 CN CN201110432747.0A patent/CN103177728B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1390349A (en) * | 1999-11-15 | 2003-01-08 | 诺基亚有限公司 | Noise suppression |
| US7680653B2 (en) * | 2000-02-11 | 2010-03-16 | Comsat Corporation | Background noise reduction in sinusoidal based speech coding systems |
| CN1447963A (en) * | 2000-08-21 | 2003-10-08 | 康奈克森特***公司 | Method for noise robust classification in speech coding |
| CN101056322A (en) * | 2006-04-13 | 2007-10-17 | 中兴通讯股份有限公司 | A device and method for overlapping the background sound at the mobile communication terminal |
| CN101197870A (en) * | 2006-12-05 | 2008-06-11 | 乐金电子(中国)研究开发中心有限公司 | Mobile terminal with adjustable speech quality |
| CN101989431A (en) * | 2009-07-31 | 2011-03-23 | 深圳富泰宏精密工业有限公司 | Device and method for eliminating noise |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI640982B (en) * | 2017-07-21 | 2018-11-11 | 宏達國際電子股份有限公司 | Method, device, and non-transitory computer readable storage medium for processing audio signal |
| CN108171118A (en) * | 2017-12-05 | 2018-06-15 | 东软集团股份有限公司 | Signal of blinking data processing method, device, readable storage medium storing program for executing and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103177728A (en) | 2013-06-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103177728B (en) | Voice signal denoise processing method and device | |
| CN110246515B (en) | Echo cancellation method and device, storage medium and electronic device | |
| EP3882913A1 (en) | Echo elimination method and terminal | |
| CN102568494B (en) | Optimized method, device and system for eliminating echo | |
| CN101026659B (en) | Method for realizing echo time delay positioning | |
| CN107360530B (en) | Echo cancellation testing method and device | |
| CN105979197A (en) | Remote conference control method and device based on automatic recognition of howling sound | |
| US9491545B2 (en) | Methods and devices for reverberation suppression | |
| CN110992923B (en) | Echo cancellation method, electronic device, and storage device | |
| CN105162555B (en) | A kind of code rate adjustment method and its terminal | |
| WO2004015961A2 (en) | Estimating bulk delay in a telephone system | |
| CN105657154B (en) | Select the method, apparatus and terminal of voice-grade channel | |
| CN104038610A (en) | Adjusting method and apparatus of conversation voice | |
| CN107770683A (en) | A kind of detection method and device of echo scene subaudio frequency acquisition state | |
| CN108091343A (en) | A kind of echo cancel method and device | |
| CN103997561A (en) | Communication apparatus and voice processing method therefor | |
| CN103249061A (en) | Method and system for testing voice quality | |
| CN104967913A (en) | Audio file playing control method and apparatus thereof | |
| US8983057B1 (en) | Step size control for acoustic echo cancellation | |
| CN104349426A (en) | Method and device for establishing conversation connection | |
| CN103354588A (en) | Determination method, apparatus and system for recording and playing sampling rate | |
| CN111613235A (en) | Echo cancellation method and device | |
| CN104717351A (en) | Terminal audio channel processing method and device and terminal | |
| CN105657203A (en) | Noise reduction method and system in voice communication of intelligent equipment | |
| CN112489679A (en) | Evaluation method and device for acoustic echo cancellation algorithm and terminal equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |