CN108712585A - Call terminal, the initial method of filter, device, equipment and storage medium - Google Patents
Call terminal, the initial method of filter, device, equipment and storage medium Download PDFInfo
- Publication number
- CN108712585A CN108712585A CN201810438294.4A CN201810438294A CN108712585A CN 108712585 A CN108712585 A CN 108712585A CN 201810438294 A CN201810438294 A CN 201810438294A CN 108712585 A CN108712585 A CN 108712585A
- Authority
- CN
- China
- Prior art keywords
- filter
- loud speaker
- filtering parameter
- signal
- sef
- 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
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M9/00—Arrangements for interconnection not involving centralised switching
- H04M9/08—Two-way loud-speaking telephone systems with means for conditioning the signal, e.g. for suppressing echoes for one or both directions of traffic
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03H—IMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
- H03H21/00—Adaptive networks
- H03H21/0012—Digital adaptive filters
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03H—IMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
- H03H21/00—Adaptive networks
- H03H21/0012—Digital adaptive filters
- H03H21/0043—Adaptive algorithms
Landscapes
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Telephone Function (AREA)
- Cable Transmission Systems, Equalization Of Radio And Reduction Of Echo (AREA)
Abstract
The invention discloses a kind of call terminal, the initial method of filter, device, equipment and storage medium, which includes senior filter, sef-adapting filter and controller；Senior filter is used for the filtering parameter according to senior filter, inhibits the first loud speaker echo signal in the first sound pick-up typing signal to obtain the first local voice signal exported to communication network；The initial value of the filtering parameter of senior filter is the optimum filtering parameter that training experiment filter obtains；Sef-adapting filter updates the filtering parameter of itself for iteration；Controller is used to, when the rejection of sef-adapting filter is better than the rejection of senior filter, the filtering parameter of senior filter is revised as the current filtering parameter of sef-adapting filter.The embodiment of the present invention realizes that call terminal starts to be provided with good echo inhibition function, optimization user's communication experience in call.
Description
Technical field
The present embodiments relate to the initialization of Audio Signal Processing technology more particularly to a kind of call terminal, filter
Method, apparatus, equipment and storage medium.
Background technology
Two-way real time phone call terminal such as mobile phone, bluetooth headset, tablet computer etc. is required for having echo inhibition function,
The echo formed due to loud speaker broadcasting opposite end voice that local sound pick-up is enrolled is eliminated, then local language is exported to opposite end
Sound, otherwise peer user can be heard oneself just finish if, influence dialogue experience.
Due to the quick and various characteristic of the difference and acoustic echo of the acoustic construction of distinct device, rational echo suppression
Device processed all has to rely on an adaptive algorithm, therefore is also required for certain training time that could really come into force.It is common to do
Method be call establish after just start train echo suppressor, and such training process completion before, echo suppressor control
The ability of echo processed is all very limited, and user is caused to remain to hear the echo of oneself within a period of time that call has just been established.This
One section of echo suppressor algorithm training time is also referred to as the convergence time of echo suppressor.
Fig. 1 shows the structure of an echo suppressor.The downlink voice signal 101 that the distal end wherein received is come can quilt
It is input to loud speaker, and is input into sef-adapting filter 111, the reference signal as sef-adapting filter 111.With reference to figure 2
Sound pick-up typing voice schematic diagram, after loud speaker plays downlink voice signal 101, due to general call terminal loud speaker with
The distance between sound pick-up all will not be too far, sound pick-up will necessarily the sound that plays of typing loud speaker, to echogenicity signal
103.Speaking for local user also will be by sound pick-up typing (being labeled as local user's voice signal 104) simultaneously.Actual pickup
Device typing signal 105 is the superposition of echo signal 103 and local user's voice signal 104.When sef-adapting filter Complete Convergence
Afterwards, it will generate a prediction signal 106 very close with echo signal 103, therefore subtract in sound pick-up typing signal 105
After falling the prediction signal, so that it may to obtain the filter output signal 107 close to local user's voice signal 104, filter
Output signal 107 can carry out the processing such as nonlinear processor or noise compensation again, just obtain the local uplink for being output to distal end
Voice.
The normal workflow of sef-adapting filter includes the iterative process of training filtering parameter：From the initial shape of filter
State (filtering parameter is the state of initial value) prediction echo signal, feeds back to filter by prediction error (i.e. residual echo signal) later
Wave device calculates the new filter parameter of one group of reduction prediction error, then predicts next samsara again with the parameter having updated
Acoustical signal, then obtain new prediction error, then prediction error is fed back to filter calculates another group of reduction prediction error
New filter parameter.Training filter repeatedly, gradually makes filter parameter converge to optimum value.
Since filter original state and optimum value state difference are generally all very big, even if filter inhibits the ability of echo
It can be gradually increased, but inhibit the ability of echo still not good enough in this convergence process median filter, influence user's communication experience.
Invention content
The present invention provides a kind of call terminal, the initial method of filter, device, equipment and storage medium, to realize
Call terminal starts to be provided with good echo inhibition function, optimization user's communication experience in call, and prevents filtering parameter from receiving
Hold back time growth.
In a first aspect, an embodiment of the present invention provides a kind of call terminal, the call terminal includes the first loud speaker, the
One sound pick-up, senior filter, sef-adapting filter and controller；
The senior filter is used for according to the filtering parameter of the first loud speaker play signal and the senior filter, inhibits the
The first loud speaker echo signal in one sound pick-up typing signal is to obtain the first local voice signal exported to communication network；
Wherein, the initial value of the filtering parameter of the senior filter is the optimum filtering parameter that training experiment filter obtains, the examination
It tests filter to be configured in experimental rig, the experimental rig includes that performance parameter identical with first loud speaker second is raised
Sound device and performance parameter the second sound pick-up identical with first sound pick-up, the experiment filter is for inhibiting the second ten
The second loud speaker echo signal in sound device typing signal；
The sef-adapting filter is used for the filter according to the first loud speaker play signal and the sef-adapting filter
Wave parameter, iteration updates the filtering parameter of itself after inhibiting the first loud speaker echo signal in the first sound pick-up typing signal；
The controller is used to be better than the rejection of the senior filter when the rejection of the sef-adapting filter
When, the filtering parameter of the senior filter is revised as the current filtering parameter of the sef-adapting filter.
Second aspect, the embodiment of the present invention additionally provide a kind of initial method of filter, and this method includes：
Obtain the optimum filtering parameter of experiment filter；Wherein, the experiment filter is configured in experimental rig, described
Experimental rig includes the second loud speaker and the second sound pick-up, and the experiment filter is for inhibiting in the second sound pick-up typing signal
The second loud speaker echo signal, the optimum filtering parameter is as obtained from the training experiment filter；
The filtering parameter of senior filter in call terminal is initialized as the optimum filtering parameter；Wherein, described logical
Telephone terminal includes senior filter, sef-adapting filter, controller and the first loud speaker and the first sound pick-up, first loud speaker
Identical with the performance parameter of the second loud speaker, second loud speaker is identical with the performance parameter of second sound pick-up；
The senior filter is used for the filtering parameter according to the first loud speaker play signal and the senior filter, inhibits the first sound pick-up
The first loud speaker echo signal in typing signal is to obtain the first local voice signal exported to communication network；It is described adaptive
Answer filter for according to the filtering parameter of the first loud speaker play signal and the sef-adapting filter, inhibiting the first ten
Iteration updates the filtering parameter of itself after the first loud speaker echo signal in sound device typing signal；The controller is for working as institute
State sef-adapting filter rejection be better than the senior filter rejection when, by the filtering parameter of the senior filter
It is revised as the current filtering parameter of the sef-adapting filter.
The third aspect, the embodiment of the present invention additionally provide a kind of apparatus for initializing of filter, including：
Acquisition module, the optimum filtering parameter for obtaining experiment filter；Wherein, the experiment filter is configured at examination
In experiment device, the experimental rig includes the second loud speaker and the second sound pick-up, and the experiment filter is for inhibition the second ten
The second loud speaker echo signal in sound device typing signal, the optimum filtering parameter are by the training experiment filter
It obtains；
Initialization module is joined for the filtering parameter of the senior filter in call terminal to be initialized as the optimum filtering
Number；Wherein, the call terminal includes senior filter, sef-adapting filter, controller and the first loud speaker and the first sound pick-up,
First loud speaker is identical with the performance parameter of second loud speaker, second loud speaker and second sound pick-up
Performance parameter is identical；The senior filter is used for the filtering parameter according to the first loud speaker play signal and the senior filter,
Inhibit the first loud speaker echo signal in the first sound pick-up typing signal to obtain the exported to communication network first local language
Sound signal；The sef-adapting filter is used for the filtering according to the first loud speaker play signal and the sef-adapting filter
Parameter, iteration updates the filtering parameter of itself after inhibiting the first loud speaker echo signal in the first sound pick-up typing signal；Institute
When stating rejection of the controller for being better than the senior filter when the rejection of the sef-adapting filter, by the master
The filtering parameter of filter is revised as the current filtering parameter of the sef-adapting filter.
Fourth aspect, the embodiment of the present invention additionally provide a kind of equipment, including：
One or more processors；
Memory, for storing one or more programs,
When one or more of programs are executed by one or more of processors so that one or more of processing
Device realizes the initial method for the filter that any embodiment of the present invention provides.
5th aspect, the embodiment of the present invention additionally provide a kind of computer readable storage medium, are stored thereon with computer
Program realizes the initial method of the filter of any embodiment body of the present invention when the program is executed by processor.
The embodiment of the present invention is obtained by training in advance and the experiment filter in the experimental rig of call terminal similar performance
To the optimum filtering parameter of the experiment filter, it is configured to for inhibiting echo signal and exporting local language in call terminal
The filtering parameter initial value of the senior filter of sound signal so that call terminal has higher possibility just to be obtained when just starting call
It obtains good echo and inhibits function, while filtering parameter is updated using sef-adapting filter iteration, prevent from experiment filter
The time that optimum filtering parameter converges to the optimum filtering parameter of senior filter longer can cause to influence user experience.To solve
The problem that prior art call terminal has the speech quality of long period not good enough in the call incipient stage realizes that call terminal exists
Call starts to be provided with good echo inhibition function, optimization user's communication experience, and prevents filtering parameter convergence time from increasing.
Description of the drawings
Fig. 1 is the structural schematic diagram of the echo suppressor of the prior art；
Fig. 2 is the voice schematic diagram of sound pick-up typing in the prior art；
Fig. 3 is the structural schematic diagram of the call terminal in the embodiment of the present invention one；
Fig. 4 is exemplary the process schematic that filtering parameter converges to ideal filtering parameter from 0；
Fig. 5 is exemplary the mistake that filtering parameter converges to ideal filtering parameter from optimum filtering parameter trained in advance
Journey schematic diagram；
Fig. 6 is the flow chart of the initial method of the filter in the embodiment of the present invention two；
Fig. 7 is the structural schematic diagram of the apparatus for initializing of the filter in the embodiment of the present invention four；
Fig. 8 is the structural schematic diagram of the equipment in the embodiment of the present invention five.
Specific implementation mode
The present invention is described in further detail with reference to the accompanying drawings and examples.It is understood that this place is retouched
The specific embodiment stated is used only for explaining the present invention rather than limitation of the invention.It also should be noted that in order to just
Only the parts related to the present invention are shown in description, attached drawing rather than entire infrastructure.
The present invention can be described by functional module element and various processing steps herein.These function modules can be with
Any amount of hardware component or software unit realize specified function by configuring.For example, the present invention can use it is various
Integrated circuit component, such as memory component, digital signal processing unit, logic unit, data form etc., one or more micro-
Various functions are realized under the control of processor or other control devices.
Any specific implementation being described herein is intended merely to preferably introduce the present invention and its optimum working mode,
It shall not be construed as any type of limitation to the present invention.For simplicity, traditional technology such as signal processing, number
According to transmission, signaling, packet-switched transmission, network-control and other system functions (and each operating member of system) may
It will not be discussed in detail herein, but skilled practitioner should have them enough understandings.In addition, the various elements in figure
Between connecting line be the different components of exemplary connection functional relationship and/or physical connection.It should be noted that many is replaced
Generation or additional functional relationships or physical connection can be realized in an actual communication system applications.
In addition, term first, second, third, etc. in the specification and in the claims is only used for difference same technique spy
The description purpose of sign is not understood to indicate or imply relative importance or implicitly indicates the number of indicated technical characteristic
Amount, also not necessarily describes order or time sequencing.Term can be interchanged in a suitable case.Define as a result, " the
One ", the feature of " second " can explicitly or implicitly include at least one of the features.
Similarly, term " connection " is equally used in the specification and in the claims, should not be construed as being limited to directly
Connection.Therefore, expression " device A is connect with device B " should not necessarily be limited by device A in device or system and be directly connected to device B,
Meaning has path between device A and device B, this can be the path for including other devices or tool.
Embodiment one
Fig. 3 is the structural schematic diagram of the call terminal in the embodiment of the present invention one.The embodiment of the present invention one provides a kind of logical
Telephone terminal is optionally that the terminal of two-way real time phone call, such as phone, bluetooth headset, tablet computer or mobile phone can be achieved.Such as
Shown in Fig. 3, the echo Restrain measurement of the call terminal is based on double filter echo inhibits framework, including the first loud speaker
35, the first sound pick-up 34, senior filter 31, sef-adapting filter 32 and controller 33.Wherein, the first loud speaker 35 is the call
Electro-acoustic transducer device in terminal, for playing voice signal, in embodiments of the present invention, the voice that the first loud speaker 35 plays
Signal is known as the first loud speaker play signal.First sound pick-up 34 is in the call terminal for acquiring sound and being transmitted to rear end
The device of equipment, in embodiments of the present invention, the voice signal of the first sound pick-up 34 acquisition (or be typing) are known as the first ten
Sound device typing signal.Controller 33 can be built in sef-adapting filter 32 or in senior filter 31, can also be whole in call
Opposite other component is independently arranged in end.
Senior filter 31 is used for according to the filtering parameter of the first loud speaker play signal and senior filter 31, inhibits the first ten
The first loud speaker echo signal in sound device typing signal is to obtain the first local voice signal exported to communication network.
It should be noted that the filtering parameter of filter usually has multiple, it is denoted as one group of filtering parameter.Fig. 4, Fig. 5 are shown
The values of multiple possibilities of the filtering parameter in convergence process.Such as in Fig. 4, filter initial parameter be (0,0 ... ...,
0), may be after the first stage (0.2, -0.1 ..., 0.1), then after a period of time may value (0.5, -
0.1 ..., 0.3), then may be backward (0.7, -0.3 ..., 0.4), gradually can be only achieved optimum value (0.7, -
0.4,...,0.3).In claims of the present invention and specification, filtering parameter 0 refers to every in one group of filtering parameter
One parameter is all 0.
Above-mentioned communication network refers to call terminal when realizing two-way call, and transmission road is provided for two-way call voice messaging
The communication network of diameter, for example, Public Switched Telephone Network (the Public Switched Telephone where phone
Network, PSTN), the internet etc. where mobile phone or tablet computer.Usually, senior filter 31, which receives, comes from communication network
In downlink voice signal, which is played by the first loud speaker 35, then the downlink voice in communication network
Signal is the first loud speaker play signal.Senior filter 31 according to the 34 typing signal of the first sound pick-up of filtering parameter pair of itself into
Row echo inhibits obtained signal directly, or can become the first local voice letter exported to communication network after treatment
Number.In some preferred embodiments, senior filter 31 carries out echo to 34 typing signal of the first sound pick-up and inhibits obtained
Signal compensates (comfort noise by nonlinear processor (Non-Linear Processor, NLP) and/or comfort noise
Generator, CNG) become the first local voice signal exported to communication network afterwards, or be ascending voice signal.
First sound pick-up typing signal mainly includes the superposition of two groups of signals, and one group of signal is to be played to believe by the first loud speaker
The voice signal that the first sound pick-up 34 is transmitted to after energy loss is undergone number in air medium, since the part signal derives from
First loud speaker play signal, that is, derive from communication network in peer user, if the signal by after sound pick-up typing not by
Inhibit and pass back to peer user, peer user can be made to hear the sound itself sent out, therefore the signal is echo signal, at this
It is referred to as the first loud speaker echo signal in inventive embodiments；Another group of signal is the voice signal that local user sends out.It can be with
Understanding, since the complexity such as sound wave diverging, absorption, reflection or scattering can occur in air medium for the first loud speaker play signal
Acoustic propagation phenomenon, and the characteristic of loud speaker, sound pick-up in the market differs, therefore the first loud speaker echo signal and first
Relationship between loud speaker play signal is difficult to determine, senior filter 31 inhibits the signal obtained after the first loud speaker echo signal
The signal that not perfect local user sends out, only one dependent on senior filter 31 filtering parameter approximation.Therewith
Similarly, it if sef-adapting filter 32 inhibits the first loud speaker echo signal in the first sound pick-up typing signal, obtains
Signal is also the approximation of a filtering parameter dependent on sef-adapting filter 32.
The important improvement of the present invention compared with the existing technology is that the initial value of the filtering parameter of senior filter 31 is instruction
Practice the optimum filtering parameter that experiment filter obtains, the experiment filter is configured in experimental rig, the experimental rig packet
Include performance parameter and first loud speaker, 35 identical second loud speaker and performance parameter and 34 phase of the first sound pick-up
The second same sound pick-up, the experiment filter is for inhibiting the second loud speaker echo in the second sound pick-up typing signal to believe
Number.It is the two-way real time phone call terminal of same model that experimental rig, which can be with call terminal provided in this embodiment, either
The semi-finished product of the two-way real time phone call terminal of same model, can also be other meet that above-mentioned performance parameter requires for testing
Device, such as can not connect with communication network, can independent play-out voice signal, typing voice signal and after inhibiting again
The device of voice signal is exported in electrical signal form.Further, the distance between the second loud speaker and the second sound pick-up and
Relative position, the performance ginseng of second loud speaker identical as the distance between the first loud speaker and the first sound pick-up and relative position
Number includes the broadcast sound volume of loud speaker, and the performance parameter of the second sound pick-up includes the sensitivity of sound pick-up.
Preferably, experimental rig is the two-way real time phone call terminal with the same model of call terminal provided in this embodiment.
The product of same model, due to its sound pick-up at a distance from loud speaker with placement position, the broadcast sound volume of loud speaker, sound pick-up
The characteristics such as sensitivity are basically the same, so the obtained optimum filtering of experiment filter in training experimental rig in advance
Parameter is very likely to the ideal filtering parameter close to the senior filter 31 in the present embodiment, and Fig. 4 is exemplary filtering parameter
The process schematic for converging to ideal filtering parameter from 0, Fig. 5 are exemplary filtering parameter from optimum filtering trained in advance
Parameter converges to the process schematic of ideal filtering parameter, it can be seen that under normal circumstances, uses advance training experiment dress
The optimum filtering parameter of acquisition is set, call terminal can be allow just to obtain good call voice matter in the call incipient stage
Amount.
Even with the product of model, the otherness that can be also inevitably present between individual, it is possible to instruction in advance
Experienced filtering parameter is the configuration of mistake, and senior filter 31 is gradually restrained by adaptive process from a wrong configuration again
Can be that 0 to converge to optimum filtering parameter more very long than filtering parameter to correct configuration, therefore, in the present embodiment, call
Terminal further includes sef-adapting filter 32 and controller 33, and sef-adapting filter 32, which is used to be played according to first loud speaker, to be believed
Number and the sef-adapting filter 32 filtering parameter, inhibit the first sound pick-up typing signal in the first loud speaker echo signal
Iteration updates the filtering parameter of itself afterwards；Controller 33 is used to be better than senior filter 31 when the rejection of sef-adapting filter 32
Rejection when, the filtering parameter of senior filter 31 is revised as the current filtering parameter of sef-adapting filter 32.By double
Filter echo inhibits framework, senior filter 31 to be responsible for that echo, sef-adapting filter 32 is inhibited to be responsible for auto-adaptive parameter training, and
Be used as the parameter of senior filter 31 using advance trained optimal parameter, thus echo can just be obtained in call beginning it is appropriate
Control, after sef-adapting filter 32 is restrained, more preferably filtering parameter can be used by senior filter 31 automatically, realize main filter
Wave device 31 is smoothly transitted into optimum Working.
It should be noted that the initial value of the filtering parameter of sef-adapting filter 32 is preferably 0, the advantage that initial value is 0
It is that convergence time and parameter quality more balance, it, can when filtering parameter trained in advance is not particularly suited for the call terminal
Time and quality of the preferable parameter to senior filter 31 are conveyed with balance.The initial value of the filtering parameter of sef-adapting filter 32
Can be the obtained optimum filtering parameter of advance training experiment filter or other values, because sef-adapting filter 32 inhibits back
After sound to voice signal be not used to communication network export, can pass through improve sef-adapting filter 32 convergence speed
Degree can't cause the quality that senior filter 31 inhibits echo to glide (and if sef-adapting filter 32 to reduce convergence time
For generating the first local voice signal exported to communication network, filtering parameter rapid divergence can be made by improving convergence rate,
Echo control may be caused unstable), so, using the technical solution of the present embodiment, it is also equipped with the advantage for shortening convergence time.
Sef-adapting filter 32 is joined according to the filtering of the first loud speaker play signal and the sef-adapting filter 32
Number, iteration updates the filtering parameter of itself after inhibiting the first loud speaker echo signal in the first sound pick-up typing signal, specifically
It may comprise steps of：
According to the filtering parameter of the first loud speaker play signal and the sef-adapting filter 32, inhibit the first pickup
The first loud speaker echo signal in device typing signal is to obtain the second local voice signal；
Prediction error is calculated according to second local voice signal；It is adaptively filtered according to described in the prediction error transfer factor
The current filtering parameter of wave device 32.Wherein, it calculates prediction error and can be according to the specific method of error transfer factor filtering parameter
Technical means commonly used by such a person skilled in the art do not repeat one by one herein.
And evaluate the rejection of filter, that is, the quality of filtering parameter is evaluated, is usually weighing apparatus with the gain of filter
Amount standard, such as echo weaken gain (echo return loss enhancement, ERLE, or the width for echo reduction
Degree).The method that controller 33 is compared the rejection of sef-adapting filter 32 and the rejection of senior filter 31 can
To be：The echo for calculating the sef-adapting filter 32 and the senior filter 31 weakens gain, if sef-adapting filter 32
Echo weaken gain be more than the senior filter 31 echo weaken gain, then confirm the rejection of sef-adapting filter 32
Better than the rejection of the senior filter 31.
On the basis of said program, training experiment filter is preferably：Inhibit described the using the experiment filter
The second loud speaker echo signal in two sound pick-up typing signals, iteration update the filtering parameter of the experiment filter until institute
It states filtering parameter and converges to optimum filtering parameter.The preferred embodiment needs experiment filter to be iterated instruction in communication process
Practice, training result relies on actual Echo, can so that training result is more accurate.Those skilled in the art can also use
Other training methods are trained filtering parameter.It should be noted that can not include other filters in experimental rig,
It completes to train using single filter, can also be using double filter framework.
In conclusion the technical solution of the present embodiment, passes through the experimental rig of training and call terminal similar performance in advance
In experiment filter, obtain the optimum filtering parameter of the experiment filter, it is configured to for inhibiting back in call terminal
Acoustical signal and export local voice signal senior filter filtering parameter initial value so that call terminal has higher possibility
Good echo is just obtained when just starting call and inhibits function, while updating filtering parameter using sef-adapting filter iteration,
Prevent the time for the optimum filtering parameter for converging to senior filter from the optimum filtering parameter for testing filter from longer can lead to shadow
Ring user experience.There is the speech quality of long period not good enough in the call incipient stage to solve prior art call terminal
Problem realizes that call terminal starts to be provided with good echo inhibition function, optimization user's communication experience in call, and prevents from filtering
Wave parameter convergence time increases.
Embodiment two
Second embodiment of the present invention provides a kind of initial method of filter, Fig. 6 is two median filter of the embodiment of the present invention
Initial method flow chart.The present embodiment is applicable to initialize the filter in call terminal, particularly, fits
For to inhibiting the filter in the call terminal of framework to initialize with double filter echo.This method can be by filtering
The apparatus for initializing of device executes, which can be integrated in computer equipment.Specifically, the initialization side of filter
Method includes step 601 and 602：
Step 601, the optimum filtering parameter for obtaining experiment filter；Wherein, the experiment filter is configured at experiment dress
In setting, the experimental rig includes the second loud speaker and the second sound pick-up, and the experiment filter is for inhibiting the second sound pick-up
The second loud speaker echo signal in typing signal, the optimum filtering parameter are obtained by the training experiment filter
's.
Specifically, the training experiment filter includes：Second sound pick-up is inhibited to record using the experiment filter
Enter the second loud speaker echo signal in signal, iteration updates the filtering parameter of the experiment filter until the filtering parameter
Converge to optimum filtering parameter.The preferred embodiment needs experiment filter to be iterated training, training result in communication process
Actual Echo is relied on, can so that training result is more accurate.Those skilled in the art can also use other training
Method is trained filtering parameter.
The filtering parameter of senior filter in call terminal is initialized as the optimum filtering parameter by step 602；Its
In, the call terminal includes senior filter, sef-adapting filter, controller and the first loud speaker and the first sound pick-up, described
The performance parameter of first loud speaker and the second loud speaker is identical, the performance parameter of second loud speaker and second sound pick-up
It is identical；The senior filter is used for the filtering parameter according to the first loud speaker play signal and the senior filter, inhibits first
The first loud speaker echo signal in sound pick-up typing signal is to obtain the first local voice signal exported to communication network；Institute
Sef-adapting filter is stated for the filtering parameter according to the first loud speaker play signal and the sef-adapting filter, is inhibited
Iteration updates the filtering parameter of itself after the first loud speaker echo signal in first sound pick-up typing signal；The controller is used
When being better than the rejection of the senior filter when the rejection of the sef-adapting filter, by the filter of the senior filter
Wave parameter is revised as the current filtering parameter of the sef-adapting filter.
The filtering parameter of the senior filter of call terminal is initialized as the optimum filtering that training experimental rig obtains in advance
Parameter can allow call terminal just to obtain good call voice quality in the call incipient stage, in conjunction with double filter
Echo inhibits framework, and echo can just obtain control appropriate when conversing and starting, even more ideal after sef-adapting filter is restrained
Filtering parameter can be used automatically by senior filter, realize senior filter be smoothly transitted into optimum Working.
Further, the initial method of the filter further includes：
The filtering parameter of sef-adapting filter in the call terminal is initialized as 0.
The advantage that the filtering parameter of sef-adapting filter in call terminal is initialized as to 0 is convergence time and parameter
Quality more balances, and when filtering parameter trained in advance is not particularly suited for the call terminal, can balance the preferable parameter of conveying
Time to senior filter 31 and quality.It is of course also possible to which the filtering parameter of the sef-adapting filter in call terminal is initial
The obtained optimum filtering parameter of the filter of training experiment in advance or other values are turned to, because after sef-adapting filter inhibits echo
To voice signal be not used to communication network export, can by improve sef-adapting filter convergence rate to drop
Low convergence time can't cause the quality that senior filter inhibits echo to glide.
Further, the sef-adapting filter is specifically used for：According to the first loud speaker play signal and it is described from
The filtering parameter of adaptive filter inhibits the first loud speaker echo signal in the first sound pick-up typing signal to obtain second
Ground voice signal；Prediction error is calculated according to second local voice signal；According to adaptive described in the prediction error transfer factor
Answer the filtering parameter that filter is current.
The controller is additionally operable to calculate the sef-adapting filter and the echo of the senior filter weakens gain, if
The echo of sef-adapting filter weakens the echo weakening gain that gain is more than the senior filter, then confirms sef-adapting filter
Rejection is better than the rejection of the senior filter.
Preferably, the experimental rig and the two-way real time phone call terminal that the call terminal is same model.Same type
Number product, due to its sound pick-up at a distance from loud speaker with the sensitivity of placement position, the broadcast sound volume of loud speaker, sound pick-up
Etc. characteristics it is basically the same, so in advance training experimental rig in the obtained optimum filtering parameter of experiment filter it is non-
It often may be close to the ideal filtering parameter of the senior filter in the present embodiment.
The initial method of filter provided in this embodiment, the call terminal for initializing the offer of embodiment one, not
In the technical detail of the detailed description of the present embodiment, reference can be made to the call terminal that embodiment one is provided.
In conclusion the technical solution of the present embodiment, passes through the experimental rig of training and call terminal similar performance in advance
In experiment filter, obtain the optimum filtering parameter of the experiment filter, it is configured to for inhibiting back in call terminal
Acoustical signal and export local voice signal senior filter filtering parameter initial value so that call terminal has higher possibility
Good echo is just obtained when just starting call and inhibits function, while updating filtering parameter using sef-adapting filter iteration,
Prevent the time for the optimum filtering parameter for converging to senior filter from the optimum filtering parameter for testing filter from longer can lead to shadow
Ring user experience.There is the speech quality of long period not good enough in the call incipient stage to solve prior art call terminal
Problem realizes that call terminal starts to be provided with good echo inhibition function, optimization user's communication experience in call, and prevents from filtering
Wave parameter convergence time increases.
Embodiment three
The embodiment of the present invention three provides a kind of apparatus for initializing of filter, and Fig. 7 is the filter that the embodiment of the present invention three provides
The structural schematic diagram of the apparatus for initializing of wave device.The apparatus for initializing of the filter includes：
Acquisition module 701, the optimum filtering parameter for obtaining experiment filter；Wherein, the experiment filter configuration
In experimental rig, the experimental rig includes the second loud speaker and the second sound pick-up, and the experiment filter is for inhibition the
The second loud speaker echo signal in two sound pick-up typing signals, the optimum filtering parameter are by the training experiment filtering
Obtained from device；
Initialization module 702, for the filtering parameter of the senior filter in call terminal to be initialized as the best filter
Wave parameter；Wherein, the call terminal includes senior filter, sef-adapting filter, controller and the first loud speaker and the first ten
The performance parameter of sound device, first loud speaker and the second loud speaker is identical, second loud speaker and second sound pick-up
Performance parameter it is identical；The senior filter is used to be joined according to the filtering of the first loud speaker play signal and the senior filter
Number, it is local to obtain exported to communication network first to inhibit the first loud speaker echo signal in the first sound pick-up typing signal
Voice signal；The sef-adapting filter is used for the filter according to the first loud speaker play signal and the sef-adapting filter
Wave parameter, iteration updates the filtering parameter of itself after inhibiting the first loud speaker echo signal in the first sound pick-up typing signal；
The controller is used for when the rejection of the sef-adapting filter is better than the rejection of the senior filter, will be described
The filtering parameter of senior filter is revised as the current filtering parameter of the sef-adapting filter.
Further, the apparatus for initializing of the filter further includes：
Second initialization module, for the filtering parameter of the sef-adapting filter in the call terminal to be initialized as 0.
Further, the training experiment filter includes：Inhibit described the second ten using the experiment filter
The second loud speaker echo signal in sound device typing signal, iteration update the filtering parameter of the experiment filter until the filter
Wave parameter converges to optimum filtering parameter.
Further, the sef-adapting filter is specifically used for：According to the first loud speaker play signal and it is described from
The filtering parameter of adaptive filter inhibits the first loud speaker echo signal in the first sound pick-up typing signal to obtain second
Ground voice signal；Prediction error is calculated according to second local voice signal；According to adaptive described in the prediction error transfer factor
Answer the filtering parameter that filter is current.
Further, the controller is additionally operable to calculate the sef-adapting filter and the echo of the senior filter weakens
Gain confirms adaptive if the echo of sef-adapting filter weakens the echo weakening gain that gain is more than the senior filter
The rejection of filter is answered to be better than the rejection of the senior filter.
Preferably, the experimental rig and the two-way real time phone call terminal that the call terminal is same model.
The apparatus for initializing for the filter that the embodiment of the present invention is provided can perform what any embodiment of the present invention was provided
The initial method of filter has the corresponding function module of execution method and advantageous effect.
Example IV
Fig. 8 is a kind of structural schematic diagram for equipment that the embodiment of the present invention four provides, as shown in figure 8, the equipment includes place
Manage device 810, memory 820, input unit 830 and output device 840；In equipment the quantity of processor 810 can be one or
It is multiple, in Fig. 8 by taking a processor 810 as an example；Processor 810, memory 820, input unit 830 in equipment and output dress
Setting 840 can be connected by bus or other modes, in Fig. 8 for being connected by bus.
Memory 820 is used as a kind of computer readable storage medium, can be used for storing software program, computer can perform journey
Sequence and module, if the corresponding program instruction/module of the initial method of the filter in the embodiment of the present invention is (for example, filtering
Acquisition module 601 in the initial method of device and initialization module 602).Processor 810 is stored in memory by operation
Software program, instruction in 820 and module are realized to execute various function application and the data processing of equipment
The initial method for the filter stated.
Memory 820 can include mainly storing program area and storage data field, wherein storing program area can store operation system
Application program needed for system, at least one function；Storage data field can be stored uses created data etc. according to terminal.This
Outside, memory 820 may include high-speed random access memory, can also include nonvolatile memory, for example, at least one
Disk memory, flush memory device or other non-volatile solid state memory parts.In some instances, memory 820 can be into one
Step includes the memory remotely located relative to processor 810, these remote memories can pass through network connection to equipment.On
The example for stating network includes but not limited to internet, intranet, LAN, mobile radio communication and combinations thereof.
Input unit 830 can be used for receiving the number or character information of input, and generate with the user setting of equipment with
And the related key signals input of function control.Output device 840 may include that display screen etc. shows equipment.
Embodiment five
The embodiment of the present invention five additionally provides a kind of computer readable storage medium, is stored thereon with computer program, should
A kind of initial method of filter is realized when program is executed by processor, including：
Obtain the optimum filtering parameter of experiment filter；Wherein, the experiment filter is configured in experimental rig, described
Optimum filtering parameter is as obtained from the training experiment filter；
The filtering parameter of senior filter in call terminal is initialized as the optimum filtering parameter；Wherein, described logical
Telephone terminal includes senior filter, sef-adapting filter, controller and the first loud speaker and the first sound pick-up, the experimental rig packet
Include the second loud speaker and the second sound pick-up, the performance parameter of first loud speaker and the second loud speaker is identical, and described second raises
Sound device is identical with the performance parameter of second sound pick-up；The senior filter is used for according to the first loud speaker play signal and institute
The filtering parameter of senior filter is stated, inhibits the first loud speaker echo signal in the first sound pick-up typing signal to obtain to communication
First local voice signal of network output；The sef-adapting filter is used for according to the first loud speaker play signal and institute
State the filtering parameter of sef-adapting filter, iteration is more after inhibiting the first loud speaker echo signal in the first sound pick-up typing signal
The new filtering parameter of itself；The controller is used for the rejection when the sef-adapting filter better than the senior filter
When rejection, the filtering parameter of the senior filter is revised as the current filtering parameter of the sef-adapting filter.
Certainly, a kind of storage medium being stored with computer program that the embodiment of the present invention is provided, computer program
The method operation being not limited to the described above, can also be performed the initial method for the filter that any embodiment of the present invention is provided
In relevant operation.
By the description above with respect to embodiment, it is apparent to those skilled in the art that, the present invention
It can be realized by software and required common hardware, naturally it is also possible to which by hardware realization, but the former is more in many cases
Good embodiment.Based on this understanding, technical scheme of the present invention substantially in other words contributes to the prior art
Part can be expressed in the form of software products, which can be stored in computer readable storage medium
In, such as the floppy disk of computer, read-only memory (Read-Only Memory, ROM), random access memory (Random
Access Memory, RAM), flash memory (FLASH), hard disk or CD etc., including some instructions are with so that a computer is set
Standby (can be personal computer, server or the network equipment etc.) executes the method described in each embodiment of the present invention.
It is worth noting that, in the embodiment of the apparatus for initializing of above-mentioned filter, included each unit and module
It is only divided according to function logic, but is not limited to above-mentioned division, as long as corresponding function can be realized；
In addition, the specific name of each functional unit is also only to facilitate mutually distinguish, the protection domain being not intended to restrict the invention.
Note that above are only presently preferred embodiments of the present invention and institute's application technology principle.It will be appreciated by those skilled in the art that
The present invention is not limited to specific embodiments described here, can carry out for a person skilled in the art it is various it is apparent variation,
It readjusts and substitutes without departing from protection scope of the present invention.Therefore, although being carried out to the present invention by above example
It is described in further detail, but the present invention is not limited only to above example, without departing from the inventive concept, also
May include other more equivalent embodiments, and the scope of the present invention is determined by scope of the appended claims.
Claims (10)
1. a kind of call terminal, which is characterized in that the call terminal include the first loud speaker, the first sound pick-up, senior filter,
Sef-adapting filter and controller；
The senior filter is used for according to the filtering parameter of the first loud speaker play signal and the senior filter, inhibits the first ten
The first loud speaker echo signal in sound device typing signal is to obtain the first local voice signal exported to communication network；Its
In, the initial value of the filtering parameter of the senior filter is the optimum filtering parameter that training experiment filter obtains；The experiment
Filter is configured in experimental rig, and the experimental rig includes that performance parameter identical with first loud speaker second is raised one's voice
Device and performance parameter the second sound pick-up identical with first sound pick-up, the experiment filter is for inhibiting the second pickup
The second loud speaker echo signal in device typing signal；
The sef-adapting filter is used to be joined according to the filtering of the first loud speaker play signal and the sef-adapting filter
Number, iteration updates the filtering parameter of itself after inhibiting the first loud speaker echo signal in the first sound pick-up typing signal；
The controller is used for when the rejection of the sef-adapting filter is better than the rejection of the senior filter, will
The filtering parameter of the senior filter is revised as the current filtering parameter of the sef-adapting filter.
2. call terminal according to claim 1, which is characterized in that the filtering parameter of the sef-adapting filter it is initial
Value is 0.
3. call terminal according to claim 1 or 2, which is characterized in that the sef-adapting filter is specifically used for：According to
The filtering parameter of the first loud speaker play signal and the sef-adapting filter inhibits in the first sound pick-up typing signal
First loud speaker echo signal is to obtain the second local voice signal；Prediction is calculated according to second local voice signal to miss
Difference；According to the current filtering parameter of sef-adapting filter described in the prediction error transfer factor；
The controller is additionally operable to calculate the sef-adapting filter and the echo of the senior filter weakens gain, if adaptive
It answers the echo of filter to weaken the echo weakening gain that gain is more than the senior filter, then confirms the inhibition of sef-adapting filter
Performance is better than the rejection of the senior filter.
4. a kind of initial method of filter, which is characterized in that including：
Obtain the optimum filtering parameter of experiment filter；Wherein, the experiment filter is configured in experimental rig, the experiment
Device includes the second loud speaker and the second sound pick-up, and the experiment filter is used to inhibit the in the second sound pick-up typing signal
Two loud speaker echo signals, the optimum filtering parameter are as obtained from the training experiment filter；
The filtering parameter of senior filter in call terminal is initialized as the optimum filtering parameter；Wherein, the call is whole
End includes senior filter, sef-adapting filter, controller and the first loud speaker and the first sound pick-up, first loud speaker and institute
The performance parameter for stating the second loud speaker is identical, and second loud speaker is identical with the performance parameter of second sound pick-up；It is described
Senior filter is used for the filtering parameter according to the first loud speaker play signal and the senior filter, inhibits the first sound pick-up typing
The first loud speaker echo signal in signal is to obtain the first local voice signal exported to communication network；The adaptive filter
Wave device is used for the filtering parameter according to the first loud speaker play signal and the sef-adapting filter, inhibits the first sound pick-up
Iteration updates the filtering parameter of itself after the first loud speaker echo signal in typing signal；The controller is used for described in certainly
When the rejection of adaptive filter is better than the rejection of the senior filter, the filtering parameter of the senior filter is changed
For the current filtering parameter of the sef-adapting filter.
5. the initial method of filter according to claim 4, which is characterized in that the initial method of the filter
Further include：
The filtering parameter of sef-adapting filter in the call terminal is initialized as 0.
6. the initial method of filter according to claim 4 or 5, which is characterized in that the training experiment filter
Wave device includes：Inhibit the second loud speaker echo signal in the second sound pick-up typing signal using the experiment filter,
Iteration updates the filtering parameter of the experiment filter until the filtering parameter converges to optimum filtering parameter.
7. the initial methods of filter according to claim 4 or 5, which is characterized in that the experimental rig leads to described
Telephone terminal is the two-way real time phone call terminal of same model.
8. a kind of apparatus for initializing of filter, which is characterized in that including：
Acquisition module, the optimum filtering parameter for obtaining experiment filter；Wherein, the experiment filter is configured at experiment dress
In setting, the experimental rig includes the second loud speaker and the second sound pick-up, and the experiment filter is for inhibiting the second sound pick-up
The second loud speaker echo signal in typing signal, the optimum filtering parameter are obtained by the training experiment filter
's；
Initialization module, for the filtering parameter of the senior filter in call terminal to be initialized as the optimum filtering parameter；
Wherein, the call terminal includes senior filter, sef-adapting filter, controller and the first loud speaker and the first sound pick-up, institute
It is identical with the performance parameter of second loud speaker to state the first loud speaker, the property of second loud speaker and second sound pick-up
Energy parameter is identical；The senior filter is used for the filtering parameter according to the first loud speaker play signal and the senior filter, suppression
The first loud speaker echo signal in the first sound pick-up typing signal is made to obtain the first local voice exported to communication network
Signal；The sef-adapting filter is used to be joined according to the filtering of the first loud speaker play signal and the sef-adapting filter
Number, iteration updates the filtering parameter of itself after inhibiting the first loud speaker echo signal in the first sound pick-up typing signal；It is described
Controller is used for when the rejection of the sef-adapting filter is better than the rejection of the senior filter, by the main filter
The filtering parameter of wave device is revised as the current filtering parameter of the sef-adapting filter.
9. a kind of equipment, which is characterized in that the equipment includes：
One or more processors；
Memory, for storing one or more programs,
When one or more of programs are executed by one or more of processors so that one or more of processors are real
The now initial method of the filter as described in any in claim 4-7.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is by processor
The initial method of the filter as described in any in claim 4-7 is realized when execution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810438294.4A CN108712585A (en) | 2018-05-09 | 2018-05-09 | Call terminal, the initial method of filter, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810438294.4A CN108712585A (en) | 2018-05-09 | 2018-05-09 | Call terminal, the initial method of filter, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108712585A true CN108712585A (en) | 2018-10-26 |
Family
ID=63867824
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810438294.4A Pending CN108712585A (en) | 2018-05-09 | 2018-05-09 | Call terminal, the initial method of filter, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108712585A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024121962A1 (en) * | 2022-12-07 | 2024-06-13 | 日本電信電話株式会社 | Acoustic quality evaluation device, acoustic quality evaluation method, and program |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1113588A2 (en) * | 1999-12-07 | 2001-07-04 | Mitsubishi Denki Kabushiki Kaisha | Echo canceling apparatus |
| CN1653713A (en) * | 2002-05-21 | 2005-08-10 | 英特尔公司 | Echo canceller with double-talk detector |
| CN1925526A (en) * | 2005-09-01 | 2007-03-07 | 华为技术有限公司 | Device and method for echo elimination |
| CN102739286A (en) * | 2011-04-01 | 2012-10-17 | 中国科学院声学研究所 | Echo cancellation method used in communication system |
| CN103179296A (en) * | 2011-12-26 | 2013-06-26 | 中兴通讯股份有限公司 | Echo canceller and echo cancellation method |
-
2018
- 2018-05-09 CN CN201810438294.4A patent/CN108712585A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1113588A2 (en) * | 1999-12-07 | 2001-07-04 | Mitsubishi Denki Kabushiki Kaisha | Echo canceling apparatus |
| CN1653713A (en) * | 2002-05-21 | 2005-08-10 | 英特尔公司 | Echo canceller with double-talk detector |
| CN1925526A (en) * | 2005-09-01 | 2007-03-07 | 华为技术有限公司 | Device and method for echo elimination |
| CN102739286A (en) * | 2011-04-01 | 2012-10-17 | 中国科学院声学研究所 | Echo cancellation method used in communication system |
| CN103179296A (en) * | 2011-12-26 | 2013-06-26 | 中兴通讯股份有限公司 | Echo canceller and echo cancellation method |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024121962A1 (en) * | 2022-12-07 | 2024-06-13 | 日本電信電話株式会社 | Acoustic quality evaluation device, acoustic quality evaluation method, and program |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP5694063B2 (en) | Indoor communication system for vehicle cabin | |
| US9264807B2 (en) | Multichannel acoustic echo reduction | |
| EP1615463B1 (en) | Adaptive howling canceller | |
| CN108141502A (en) | Audio signal processing | |
| CN109937579A (en) | In-Ear active noise reduction earphone | |
| CN111213359B (en) | Echo canceller and method for echo canceller | |
| CN110769352B (en) | Signal processing method and device and computer storage medium | |
| CN113241085B (en) | Echo cancellation method, device, equipment and readable storage medium | |
| CN105453530A (en) | Noise reduction | |
| CN110992923B (en) | Echo cancellation method, electronic device, and storage device | |
| US11895470B2 (en) | Methods of processing of audio signals | |
| CN108712585A (en) | Call terminal, the initial method of filter, device, equipment and storage medium | |
| US9031836B2 (en) | Method and apparatus for automatic communications system intelligibility testing and optimization | |
| US8768406B2 (en) | Background sound removal for privacy and personalization use | |
| CN107666637A (en) | Self-adjustable active noise cancellation method, system and Headphone device | |
| US10481831B2 (en) | System and method for combined non-linear and late echo suppression | |
| JP2018064162A (en) | Sound quality evaluation device, sound quality evaluation method, and program | |
| Rombouts et al. | Generalized sidelobe canceller based combined acoustic feedback-and noise cancellation | |
| US12010486B2 (en) | Detection of feedback path change | |
| CN115620737A (en) | Voice signal processing device, method, electronic equipment and sound amplification system | |
| CN113225574B (en) | Signal processing method and device | |
| CN102117620A (en) | Method for transmitting coefficients of double-filter and device thereof | |
| JP5245884B2 (en) | Voice communication system and voice communication apparatus | |
| US7907737B2 (en) | Acoustic apparatus | |
| CN110972032B (en) | Method for eliminating sound and electronic device for executing method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181026 |
|
| RJ01 | Rejection of invention patent application after publication |