CN101222243B - Filter coefficient setting device and echo prevention device - Google Patents
Filter coefficient setting device and echo prevention device Download PDFInfo
- Publication number
- CN101222243B CN101222243B CN2008100028834A CN200810002883A CN101222243B CN 101222243 B CN101222243 B CN 101222243B CN 2008100028834 A CN2008100028834 A CN 2008100028834A CN 200810002883 A CN200810002883 A CN 200810002883A CN 101222243 B CN101222243 B CN 101222243B
- Authority
- CN
- China
- Prior art keywords
- signal
- aforementioned
- filter
- echo
- response
- 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.)
- Expired - Fee Related
Links
Images
Abstract
The invention relates to a filter coefficient setting device, comprising: a filter coefficient initial setting section which sets a predetermined filter coefficient for a first and a second FIR filters when an echo prevention device is started; a response signal obtaining section comprising a first response signal which generates a first signal and obtains output from the first FIR filter to an AD transducer, and a second response signal which generates a second signal and obtains output from the second FIR filter to the AD transducer; and a filter coefficient setting section for making a fourth simulation signal become a signal which is obtained after eliminating or attenuating a first simulation signal from a signal combined with the first simulation signal and a third simulation signal, and setting filter coefficient obtained according to the first response signal for the second FIR filter, and setting filter coefficient obtained according to the second response signal for the first FIR filter.
Description
Technical field
The present invention is relevant filter coefficient setting device and echo prevention device.
This case is according to the special 2007-050337 of hope of Japan that files an application in the special 2007-004814 of hope of the Japan that filed an application on January 12nd, 2007, on February 28th, 2007, advocates priority with the special 2007-058962 of hope of the Japan that filed an application on March 8th, 2007, and in this case, quotes its content.
Background technology
In recent years; Have in communication tools such as mobile phone that for example connects ear microphone (earphone microphone) or heldfree type phone, assembling is used for preventing the echo prevention device by the echo that produces from the sound coupling of loud speaker wraparound to microphone or the electric reflection on the circuit etc.
For example, in TOHKEMY 2006-304260 communique, disclosed and a kind ofly use digital processing and eliminate the echo prevention device of echo accurately.Figure 13 shows an example of using the echo prevention device that DSP100 is arranged.As shown in the figure, transmit and the analog signal of the sound that comes from the other side's side in order to show with mobile phone etc., be to input to AD (Analog/Digital; Analog/digital) transducer 101.Then, utilizing AD converter 101 to convert the signal of numeral into, is finite impulse response (FIR) (FIR) filter 102,103 in DSP100, imposes convolution operation (convolution) according to various filter factors and handles output again.From the signal of FIR filter 102 outputs, be to input to DA (Digital/Analog; Digital-to-analog) transducer 104.Then, utilizing DA transducer 104 to convert the signal of simulation into, is to export ear microphone to through output/input terminal 105, and inputs to the wherein square end of differential amplifier circuit 106.In addition, the signal from 103 outputs of FIR filter is to input to DA transducer 107.Then, the signal from 107 outputs of DA transducer is the opposing party's terminal that inputs to differential amplifier circuit 106.Then, the signal from differential amplifier circuit 106 outputs is to convert digital signal into and export DSP100 at AD converter 108.
Here, DSP100 is by the output of the AD converter 108 of output during pulse to DA transducer 104, obtains the impulse response to AD converter 108 from DA transducer 104.In addition, DSP100 is by the output of the AD converter 108 of output during pulse to DA transducer 107, obtains the impulse response to AD converter 108 from DA transducer 107.Then, suitably set the filter factor of FIR filter 102,103, can eliminate by the echo that signal was produced that inputs to AD converter 101 with differential amplifier circuit 106 therefrom according to those impulse responses.
From the signal of AD converter 108 output is output signal and the output that subtracting section 110 in DSP100 deducts adaptive filter 111.This adaptive filter 111 is under the state that has voice signal to send from the other side's side, filter factor to be changed adaptively so that the output signal that comes from subtracting section 110 becomes below the predetermined level, eliminates differential amplifier circuit 106 therefrom and does not eliminate most echo.Then, be after DSP100 output from the digital signal of subtracting section 110 output, convert analog signal at DA transducer 109, export as the output signal of echo prevention device.
Yet when echo prevention device started, the filter factor of FIR filter 102,103 became underrange.Therefore, according to carry out from user's indication impulse response obtain and suitably FIR filter 102,103 set filter factors during, do not carry out the elimination of echo, can make and make us uncomfortable echo and be transferred into the other side's side.In addition,, the user of echo prevention device makes us uncomfortable echo to the other side's side because can't knowing just transmission, thus do not notice sometimes and must obtain impulse response, and transmission is constantly made us uncomfortable echo to the other side's side.
In addition, be in echo prevention device after differential amplifier circuit 106 is eliminated echo, in DSP100, further carry out the elimination of echo.For example, have, just can't eliminate the situation of echo fully at differential amplifier circuit 106 when causing owing to the influences such as noise around electrical noise or the ear microphone wearer can't obtain impulse response accurately the time.Under this situation,, eliminate 106 echoes that can't eliminate of differential amplifier circuit effectively by the action of subtracting section in the DSP100 110 and adaptive filter 111.Yet, the situation that the action of subtracting section 110 and adaptive filter 111 makes echo increase is on the contrary but arranged when differential amplifying circuit 106 can be eliminated echo fully.
In addition,, after differential amplifier circuit 106 is eliminated echoes, in DSP100, further carry out the elimination of echo, have and fully do not eliminate echo and the remaining situation of echo is arranged though have in echo prevention device.But as far as the first speaker of the other side's side, to scarcely having inharmonious sense with the echo of giving orders or instructions almost to return simultaneously.
On the other hand, in echo prevention device, have because of the noise that comes from ear microphone or the influences such as electrical noise of circuit and cause in the situation of the no small noise of signal overlap that is sent to adversary's side.About in order to eliminate so countermeasure of noise, can consider and to export again through noise eliminator from the signal of subtracting section 110 outputs.
Yet,, again the data of accumulating are imposed the noise removing processing, thereby the delay corresponding to burden and treating capacity can take place because noise eliminator generally is to accumulate to a certain degree will import data.Therefore, though only have from the echo that signal contained of subtracting section 110 output a little, but as far as the first speaker of the other side's side, owing to can lately a little after giving orders or instructions hearing that echo produces inharmonious sense.
Summary of the invention
The wherein filter coefficient setting device of an embodiment of the present invention; It is the filter coefficient setting device of setting the filter factor of echo prevention device; And aforementioned echo prevention device possesses and has: the 1FIR filter is the input that receives the 1st digital signal, and exports the 2nd digital signal; The 2FIR filter is to receive the input of aforementioned the 1st digital signal simultaneously with aforementioned 1FIR filter, and exports the 3rd digital signal; The 1DA transducer is to convert aforementioned the 2nd digital signal to the 1st analog signal and output; The 2DA transducer is to convert aforementioned the 3rd digital signal to the 2nd analog signal and output; Output/input terminal is aforementioned the 1st analog signal of output or import the 3rd analog signal; Subtracting section, be output will aforementioned the 1st analog signal and the signal that combines of aforementioned the 3rd analog signal deduct aforementioned the 2nd analog signal and must the 4th analog signal; And AD converter, be that the analog signal conversion from aforementioned subtraction portion output is become digital signal and output; And this filter coefficient setting device is to possess: the filter factor initial setting section is when the starting of aforementioned echo prevention device, to the aforementioned the 1st and the predetermined filter factor of 2FIR filter configuration; Response signal obtains portion; Be to obtain output by the 1st signal is produced from aforementioned 1FIR filter; Arrive the 1st response signal of the output of aforementioned AD converter; And obtain output from aforementioned 2FIR filter by the 2nd signal is produced, to the 2nd response signal of the output of aforementioned AD converter; And filter coefficient setting portion; Be for aforementioned the 4th analog signal is become; From the signal that aforementioned the 1st analog signal and aforementioned the 3rd analog signal combine, eliminate or aforementioned the 1st analog signal that decays after and signal; And to the filter factor of aforementioned 2FIR filter configuration, and to the filter factor of aforementioned 1FIR filter configuration according to aforementioned the 2nd response signal gained according to aforementioned the 1st response signal gained;
When aforementioned filter factor initial setting section is started at aforementioned echo prevention device; As in response in the aforementioned the 1st and the 1st and the 2nd response signal data of the 2nd response signal when being stored in memory; For aforementioned the 4th analog signal is become; From the signal that aforementioned the 1st analog signal and aforementioned the 3rd analog signal combine, eliminate or aforementioned the 1st analog signal that decays after and signal; And aforementioned the 2nd finite impulse response filter set basis there is the filter factor of aforementioned the 1st response signal gained; And aforementioned the 1st finite impulse response filter set basis had the filter factor of aforementioned the 2nd response signal gained, when not having the aforementioned the 1st and the 2nd response signal data to be stored in aforementioned memory, the aforementioned the 1st and the 2nd finite impulse response filter is set aforementioned predetermined filter factor;
Aforementioned response signal obtains portion when not having the aforementioned the 1st and the 2nd response signal data to be stored in aforementioned memory, obtains the aforementioned the 1st and the 2nd response signal.
Description of drawings
About further feature of the present invention, can understand by the record of accompanying drawing and this specification.
In order more fully to understand the present invention and advantage thereof, please together with reference to description and accompanying drawing, wherein:
Fig. 1 is the calcspar of an example of the demonstration echo prevention device that the present invention was suitable for.
Fig. 2 is the part of the storage part that has of display-memory.
Fig. 3 shows constituting of the function square realized by the DSP kernel executive.
Fig. 4 is the flow chart that shows an example of filter coefficient setting processing.
Fig. 5 is the calcspar of formation that shows the echo prevention device of this example.
Fig. 6 is the routine calcspar of formation that shows the Echo Canceller of this example.
Fig. 7 is the flow chart that shows an example of the processing of calculating noise power in the Echo Canceller of this example.
Fig. 8 is the flow chart that shows an example of the processing of the open and close of decision echo cancellation process in the Echo Canceller of this example.
Fig. 9 is the routine calcspar of other formation that shows the Echo Canceller of this example.
Figure 10 is the calcspar of formation that shows the echo prevention device of this example.
Figure 11 is the calcspar of the example that shows that the periphery of the noise eliminator of this example constitutes.
Figure 12 is the flow chart of an example of processing of open and close that shows the noise eliminator of this example of control.
Figure 13 is an example that shows the echo prevention device that uses DSP.
Embodiment
According to the record of this specification and accompanying drawing, at least can clear and definite following item.
" the 1st example "
==all formations==
Fig. 1 is the calcspar of an example of the demonstration echo prevention device that the present invention was suitable for.Echo prevention device 1 is to possess digital signal processing circuit (DSP; Digital Signal Processor) 3, AD converter 4,5, DA transducer 6 to 8, amplifying circuit 9 to 11, differential amplifier circuit 12 and output/input terminal 13.And, in the outside of echo prevention device 1 be provided with ear microphone 20, CPU (Central Processing Unit) 25, memory 26, study button 27, with emergency warning lamp 28.Also have, DA transducer 6 is to be equivalent to 1DA transducer of the present invention, and DA transducer 7 is to be equivalent to 2DA transducer of the present invention.In addition, differential amplifier circuit 12 is to be equivalent to subtracting section of the present invention.
CPU25 be stored in memory 26 by execution program with the anti-locking apparatus 1 of integration control echo.For example, CPU25 is when detecting the electric power starting of the echo prevention device 1 action usefulness of sening as an envoy to, will be with so that the index signal of the filter coefficient setting processing execution of stating after treating exports DSP3 to.In addition, in the time of for example also can working as input in order to the reset signal of (reset) echo prevention device 1 of resetting, CPU25 just exports aforesaid index signal to DSP3.
DSP3 contains input terminal 30,31, lead-out terminal 32 to 34, DSP core (core) 40, RAM (Random Access Memory) 41 and ROM (Read Only Memory) 42 and constitutes.In addition, DSP3 possesses FIR filter 50,51.Those FIR filters the 50, the 51st are carried out the program that is stored in RAM41 or ROM42 by DSP core 40 and are realized.In addition, the filter factor of FIR filter 50,51 is to be stored in RAM41.Here, FIR filter 50 is to be equivalent to 1FIR filter of the present invention, and FIR filter 51 is to be equivalent to 2FIR filter of the present invention.Also have, also can realize FIR filter 50,51 by hardware.
In AD converter 4 are input audio signals for example.And AD converter 4 is with voice signal is simulated. digital translation handle and digital signal (the 1st digital signal) input to DSP3 through input terminal 30.
The digital signal that inputs to DSP3 is to input to FIR filter 50,51 respectively.FIR filter 50 is to handle digital signal (the 2nd digital signal: the 1st input signal) export lead-out terminal 32 to that gets with the digital signal (the 3rd input signal) of input is implemented convolution operation according to the filter factor of this FIR filter 50.Simultaneously, FIR filter 51 is to handle digital signal (the 3rd digital signal: the 1st input signal) export lead-out terminal 33 to that gets with the digital signal (the 3rd input signal) of input is implemented convolution operation according to the filter factor of this FIR filter 51.
In DA transducer 6 is to import the output signal that comes from FIR filter 50 through lead-out terminal 32.And DA transducer 6 is with the output signal that comes from FIR filter 50 is implemented numeral. analog-converted and analog signal (the 1st analog signal) export amplifying circuit 9 to.Amplifying circuit 9 is with predetermined magnification ratio analog signal to be amplified and output.
In DA transducer 7 is to import the output signal that comes from FIR filter 51 through lead-out terminal 33.And, DA transducer 7 be with the output signal that comes from FIR filter 51 is implemented the digital-to-analog conversion and analog signal (the 2nd analog signal) export amplifying circuit 11 to.Amplifying circuit 11 is analog signal is amplified and to export differential amplifier circuit 12-(bearing) input terminal to predetermined magnification ratio.
DSP core 40 (processor) is to carry out the program that is stored in RAM41 or ROM42, can carry out the various processing of DSP3 therefrom.Fig. 3 shows constituting of the function square realized by DSP core 40 executive programs.DSP3 be possess filter factor initial setting section 65, warning efferent 66, impulse response obtain portion 67 and filter coefficient setting portion 68.Also have, obtaining the processing unit that portion 67 and filter coefficient setting portion 68 constitute by filter factor initial setting section 65, warning efferent 66, impulse response is to be equivalent to filter coefficient setting device of the present invention.
Filter factor initial setting section 65 is when echo prevention device 1 starting, and the initial value (predetermined filter factor) of FIR filter 50,51 being set filter factor is with the digital signal decay output again with input.At the 1st example, be that the initial value of setting filter factor all is all zero in the position.Also have, when whole positions of filter factor were all zero, the digital signal of input was exported from FIR filter 50,51 with the mode of noise reduction just by decay fully.In addition, filter factor initial setting section 65 is when storing in impulse response storage part 62 in response in the data of impulse response the time, in response to the data in impulse response FIR filter 50,51 being set filter factors according to this.
Impulse response obtains portion 67 and obtains flag storage part 61 through the CPU25 inquiry, when not obtaining impulse response as yet, obtains impulse response.Particularly, it is from lead-out terminal 32 output pulses (the 1st signal) that impulse response obtains portion 67, obtains the impulse response IR ' 1 (Z) (the 1st response signal) of the represented path A of the solid line of Fig. 1 therefrom from input terminal 31.In addition, it is from lead-out terminal 33 output pulses (the 2nd signal) that impulse response obtains portion 67, obtains the impulse response IR ' 2 (Z) (the 2nd response signal) of the represented path B of the solid line of Fig. 1 therefrom from input terminal 31.In addition, impulse response obtains portion 67 impulse response that obtains is stored in impulse response storage part 62 through CPU25.Also have, impulse response obtains portion 67 and only the impulse response that obtains is stored in impulse response storage part 62, also can with in response in the data storing of impulse response in impulse response storage part 62.For example, also can the filter factor according to impulse response that filter 50,51 is set be stored in impulse response storage part 62.
Filter coefficient setting portion 68 is according to being obtained impulse response IR1 ' that portion 67 obtains by impulse response (Z), setting the filter factor of FIR filter 51.In addition, filter coefficient setting portion 68 is according to being obtained impulse response IR2 ' that portion 67 obtains by impulse response (Z), setting the filter factor of FIR filter 50.
The principle that==echo is eliminated==
Then, eliminating principle to the echo of echo prevention device 1 describes.Here, make the dotted line of Fig. 1 represented from lead-out terminal 32 to differential amplifier circuit 12+impulse response (transfer function) of input terminal is IR1 (Z).In addition, make the dotted line of Fig. 1 represented from lead-out terminal 33 to differential amplifier circuit 12-impulse response (transfer function) of input terminal is IR2 (Z).In addition, make the dotted line of Fig. 1 represented from differential amplifying circuit 12 ± the back level of input terminal is W (Z) to the impulse response (transfer function) of input terminal.
At this moment, impulse response (transfer function) IR1 ' Z of the represented path A of the solid line of Fig. 1 just be IR1 ' (Z)=IR1 (Z) W (Z).In addition, impulse response (transfer function) IR2 ' Z of the represented path B of the solid line of Fig. 1 just be IR2 ' (Z)=-IR2 (Z) W (Z).Also have, IR2 (Z) phasing back is owing to inputing to the event of differential amplifier circuit 12-input terminal.
Now, the filter factor that makes FIR filter 50 for IR2 ' (Z) after the phasing back-IR2 ' (Z), then just become from the characteristic IRall_1 that is input to input terminal 31 of FIR filter 50 (Z):
IRall_1(Z)＝-IR2’(Z)·IR1’(Z)
＝(-(-IR2(Z)·W(Z)))·(IR1(Z)·W(Z))
＝IR2(Z)·W(Z)·IR1(Z)·W(Z)。
In addition, the filter factor that makes FIR filter 51 be IR1 ' (Z), then just become from the characteristic IRall_2 that is input to input terminal 31 of FIR filter 51 (Z):
IRall_2(Z)＝IR1’(Z).IR2’(Z)
＝IR1(Z)·W(Z)·(-IR2(Z)·W(Z))
＝IR1(Z)·W(Z)·(-IR2(Z))·W(Z)
＝-IRall_1(Z)。
That is, can learn become for from the characteristic IRall_1 that is input to input terminal 31 of FIR filter 50 (Z), with the characteristic of cancelling each other from the characteristic IRall_2 that is input to input terminal 31 of FIR filter 51 (Z).Therefore, can know if with the filter coefficient setting of FIR filter 50 for IR2 ' (Z) phasing back-IR2 ' (Z), be that IR1 ' (Z) just can with the filter coefficient setting of FIR filter 51.
Perhaps, the filter factor that makes FIR filter 50 be IR2 ' (Z), then just become from the IRall_1 that is input to input terminal 31 of FIR filter 50 (Z):
IRall_1(Z)＝IR2’(Z)·IR1’(Z)
＝(-IR2(Z)·W(Z))·(IR1(Z)·W(Z))
＝-IR2(Z)·W(Z)·IR1(Z)·W(Z)。
In addition, the filter factor that makes FIR filter 51 for IR1 ' (Z) after the phasing back-IR1 ' (Z), then just become from the characteristic IRall_2 that is input to input terminal 31 of FIR filter 51 (Z):
IRall_2(Z)＝-IR1’(Z)·IR2’(Z)
＝(-(IR1(Z)·W(Z)))·(-IR2(Z)·W(Z))
＝IR1(Z)·W(Z)·IR2(Z)·W(Z)
＝-IRall_1(Z)。
That is, can learn become for from the characteristic IRall_1 that is input to input terminal 31 of FIR filter 50 (Z), with the characteristic of cancelling each other from the characteristic IRall_2 that is input to input terminal 31 of FIR filter 51 (Z).Therefore, can know if with the filter coefficient setting of FIR filter 50 be IR2 ' (Z), with the filter coefficient setting of FIR filter 51 for IR1 ' (Z) phasing back-IR1 ' (Z) just can.
Then, by so according to impulse response IR1 ' (Z) and IR2 ' (Z) set the filter factor of FIR filter 50,51, just can be passed to the signal of path A at differential amplifier circuit 12 with the signal cancellation that is passed to path B.Therefore, the echo in the time of can preventing that digital signal from inputing to input terminal 30.
In addition, as shown in Figure 1, under the state that is connected with ear microphone 20, obtain impulse response IR1 ' (Z), and (Z) set the filter factor of FIR filter 51 according to this IR1 ', thus, can prevent echo effectively in response to the transmission characteristic of ear microphone 20.In addition; The modes such as earholes or covering auricle of inserting at the ear microphone 20 with connection are worn and under the state on the ear, are obtained impulse response IR1 ' (Z); And (Z) set the filter factor of FIR filter 51 according to this IR1 '; Thus, can be in response to the transmission characteristic in the transmission characteristic of ear microphone 20 and user's the ear and prevent echo effectively.Also have, also can under the state that does not connect ear microphone 20, obtain impulse response IR1 ' (Z), and (Z) set the filter factor of FIR filter 51 according to this IR1 '.
The processing of==filter coefficient setting==
Then, the filter coefficient setting processing to echo prevention device 1 describes.Fig. 4 is the flow chart that shows an example of filter coefficient setting processing.At first, when echo prevention device 1 starting (S401), filter factor initial setting section 65 be through CPU25 read be stored in obtain flag storage part 61 obtain flag (S402).
When from obtaining that flag storage part 61 reads and obtain flag for " 1 " time; That is when impulse response be (S402:1) when being stored in impulse response storage part 62, filter factor initial setting section 65 be through CPU25 read be stored in impulse response storage part 62 impulse response IR1 ' (Z) and IR2 ' (Z) (S403).Then, filter factor initial setting section 65 be according to read from impulse response storage part 62 the impulse response IR1 ' that gets (Z) and IR2 ' (Z) set the filter factor (S404) of FIR filter 50,51.Particularly, the filter factor of for example setting FIR filter 50 for IR2 ' (Z) phasing back-IR2 ' (Z), the filter factor of setting FIR filter 51 be IR1 ' (Z).
When from obtaining that flag storage part 61 reads and obtain flag for " 0 " time; That is when impulse response is not stored in impulse response storage part 62 (S402:0), filter factor initial setting section 65 is that the whole position to the filter factor of FIR filter 50,51 is set at zero as initial value (S405).In addition, export warning signal to CPU25, make the bright lamp of emergency warning lamp (S406) by warning efferent 66.
Afterwards, when CPU25 notice study button 27 had been pressed, impulse response obtained portion 67 and just obtains impulse response IR1 ' (Z) and IR2 ' (Z) (S407).Then, filter coefficient setting portion 68 be according to obtain by impulse response impulse response IR1 ' that portion 67 obtains (Z) and IR2 ' (Z) set the filter factor (S408) of FIR filter 50,51.Particularly, the filter factor of for example setting FIR filter 50 for IR2 ' (Z) phasing back-IR2 ' (Z), the filter factor of setting FIR filter 51 be IR1 ' (Z).Then, will represent that by warning efferent 66 caution release signal of having obtained impulse response exports CPU25 to, make emergency warning lamp light-off (S409).
In addition; Impulse response obtain portion 67 be with the impulse response IR1 ' that obtains (Z) and IR2 ' (Z) be stored in impulse response storage part 62 (S410) through CPU25, and will be set at that to represent impulse response be that the flag that obtains of acquired " 1 " is stored in through CPU25 and obtains flag storage part 61 (S411).
Also have; At the 1st example; Though be to the filter coefficient setting of FIR filter 50 with impulse response IR2 ' (Z) phasing back-IR2 ' (Z), (Z) to the filter coefficient setting impulse response IR1 ' of FIR filter 51, also can to the filter coefficient setting impulse response IR2 ' of FIR filter 50 (Z), to the filter coefficient setting of FIR filter 51 with impulse response IR1 ' (Z) phasing back-IR1 ' (Z).
More than, describe to the 1st example.As previously mentioned; When echo prevention device 1 starting; The filter factor (predetermined filter factor) that utilizes 50,51 settings of 65 pairs of FIR filters of filter factor initial setting section that the digital signal decay of input is exported again thus, can make the signal level of the echo that is sent to the other side's side reduce.That is, the uncomfortable echo of making us that can suppress to be sent to the other side's side.In addition, by at FIR filter 50 with signal attenuation, also diminish from the sound of ear microphone 20 output.Therefore, the user of ear microphone 20 will find not obtain as yet impulse response easily, and can impel obtaining of impulse response.
In addition, shown in the 1st example, filter factor initial setting section 65 can be set at zero as initial value to whole positions of the filter factor of FIR filter 50,51.Under this situation, the digital signal that inputs to FIR filter 50,51 is exported with the mode of noise reduction by decay fully.Therefore, can not transmit echo to the other side's side.In addition, owing to becoming the state that the sound that transmits from adversary's side is not exported from ear microphone 20 fully, so become the state that easy discovery does not obtain impulse response as yet.
In addition, filter factor initial setting section 65 is when echo prevention device 1 starting, when when impulse response storage part 62 stores impulse response, sets the filter factor of FIR filter 50,51 according to the impulse response that reads out from impulse response storage part 62.That is after echo prevention device 1 starting, according to acquired impulse response FIR filter 50,51 is set and can be suppressed the suitable filter factor of echo, make us uncomfortable echo and can suppress to the other side's side transmission.
In addition; Impulse response obtains portion 67 impulse response that obtains is stored in impulse response storage part 62; Thus, just can when starting echo prevention device next time, can suppress the suitable filter factor of echo to the FIR filter configuration; Can not need to obtain once again impulse response, and suppress to the other side's side transmission and make us uncomfortable echo.
In addition, when when the impulse response storage part does not store impulse response, warning efferent 66 is output warning signals and carry out the bright lamp etc. of emergency warning lamp 28, thus, can let the user of ear microphone 20 be easier to understand impulse response and not obtain as yet.
Also have, at the 1st example, though be the filter factor that the impulse response that obtains by pulse is produced is set at the FIR filter, the signal that uses for the setting filter factor is not to be defined as pulse.The response signal that for example, also can obtain when step (step) signal is produced is set the filter factor of FIR filter.
" the 2nd example "
==all formations==
At first, the formation to the echo prevention device of the present invention's the 2nd example describes.Fig. 5 is the calcspar of formation of the echo prevention device of the 2nd example.Also have, the formation identical with the 1st example is mark same components symbol and omits explanation.
CPU25 is when detecting the electric power starting of the echo prevention device 1 action usefulness of sening as an envoy to, will be with so that the index signal of stating after treating according to the setting processing execution of the filter factor of impulse response exports DSP3 to.In addition, in the time of for example also can working as input in order to the reset signal of replacement echo prevention device 1, CPU25 just exports aforesaid index signal to DSP3.
DSP3 contains input terminal 30,31, lead-out terminal 32 to 34, DSP core 40, RAM (Random Access Memory) 41 and ROM (Read Only Memory) 42 and constitutes.In addition, DSP3 be possess FIR filter 50,51, impulse response obtains portion 52, filter coefficient setting portion 53 and Echo Canceller (EC) 54.It is to carry out the program that is stored in RAM41 or ROM42 by DSP core 40 to realize that those FIR filters 50,51, impulse response obtain portion 52, filter coefficient setting portion 53 and Echo Canceller 54.In addition, the filter factor of FIR filter 50,51 is to be stored in RAM41.Also have, FIR filter 50 is to be equivalent to 1FIR filter of the present invention, and FIR filter 51 is to be equivalent to 2FIR filter of the present invention.
Impulse response obtain portion 52 (the signal response characteristic obtains portion) be obtain the impulse response from input terminal 31 (the 1st signal response characteristic) when making pulse result from the output of FIR filter 50, with the impulse response from input terminal 31 (the 2nd signal response characteristic) when making pulse result from the output of FIR filter 51.Filter coefficient setting portion 53 is according to being obtained filter factor that impulse response that portion 52 obtains sets FIR filter 50,51 by impulse response with the output signal that utilizes amplifying circuit 11 at differential amplifier circuit 12 will combine the output signal of amplifying circuit 9, the signal that reflects through ear microphone 20 with the output signal of amplifying circuit 9 get signal (echo) elimination or decay.
==Echo Canceller==
Then, the Echo Canceller 54 to the 2nd example is elaborated.Fig. 6 is the routine calcspar of formation that shows the Echo Canceller 54 of the 2nd example.Echo Canceller 54 is to contain conversation test section 70,71, echo level test section 72, noise level storage part 73, subtracting section 74, adaptive filter 75, switch 76 and control part 77 and constitute.Also have, subtracting section 74 and adaptive filter 75 are to be equivalent to echo elimination of the present invention portion.
In conversation test section 70 are input signals (the 1st input signal) from 50 outputs of FIR filter.Whether conversation test section 70 is to detect according to the signal from 50 outputs of FIR filter to have voice signal to send from the other side's side.For example; Conversation test section 70 is if be the words more than the predetermined level from the signal level of the signal of FIR filter 50 output; Just be judged to be voice signal and sent,, just be judged to be the voiceless sound signal and send from the other side's side if less than predetermined level from the other side's side.In addition, the signal of exporting from FIR filter 50 is to detect processing at conversation test section 70, and exports lead-out terminal 32 to.
In conversation test section 71 are input signals from subtracting section 74 outputs.Conversation test section 71 is the telephone conditions that detect ear microphone 20 wearers according to the signal from subtracting section 74 outputs.For example; Conversation test section 71 is if be the words more than the predetermined level from the signal level of the signal of subtracting section 74 output; Just the wearer who is judged to be ear microphone 20 gives orders or instructions, if less than predetermined level, the wearer who just is judged to be ear microphone 20 does not give orders or instructions.In addition, the signal of exporting from subtracting section 74 is to detect processing at conversation test section 71, and exports lead-out terminal 34 to.
Echo level test section 72 is the testing results according to conversation test section 70,71, and is not transmitting the signal level that detects under the state that receives voice signal from the signal of input terminal 31 inputs, and is stored in noise level storage part 73.In detected signal level under this state is the level that is presented at the noise that produces under the state of conversing.In addition; Echo level test section 72 is the testing results according to conversation test section 70,71, and detects from the signal (the 2nd input signal) of input terminal 31 inputs under the state that the wearer of ear microphone 20 does not give orders or instructions there being voice signal to send from the other side's side.Detected signal is that representative contains the level of not eliminating the signal of most echo at differential amplifier circuit 12 under this state.And echo level test section 72 is that the comparative result of level with the noise level that is stored in noise level storage part 73 and the signal that contains echo exports control part 77 to.
Also have, in the 2nd example, echo level test section 72 is to detect from the signal of input terminal 31 inputs at the power of scheduled period as signal level.But in the signal level that echo level test section 72 detects not is to be defined as power, for example as long as just can at the peak swing of scheduled period or the level person that the absolute value sum total of the amplitude of scheduled period etc. can shows signal.
Subtracting section 74 is to deduct signal and output from adaptive filter 75 outputs with the signal from input terminal 31 inputs.In adaptive filter 75 are inputs from the signal of FIR filter 50 output, with the output signal of subtracting section 74.And adaptive filter 75 is to have voice signal to send from the other side's side, and under the state that the wearer of ear microphone 20 does not give orders or instructions, changes filter factor adaptively so that the signal of exporting from subtracting section 74 becomes below the predetermined level.Switch 76 is whether to export the output signal of adaptive filter 75 to subtracting section 74 in order to control.Also having, is and to move identical with the formation of the adaptive filter that for example is disclosed in TOHKEMY 2006-304260 communique about the setting action of the formation of adaptive filter 75 and filter factor.
Control part 77 (output control part) is the testing result according to conversation test section 70,71; There is being voice signal to send from the other side's side; And under the state that the wearer of ear microphone 20 does not give orders or instructions, make adaptive filter 75 carry out the adaptability update processing of filter factor.In addition, control part 77 is whether will carry out echo elimination at subtracting section 74 according to controlling from the comparative result of echo level test section 72 outputs.Particularly; There is being voice signal to send from the other side's side; And under the state that the wearer of ear microphone 20 does not give orders or instructions; When the echo that signal contained from input terminal 31 input is predetermined level when above, control part 77 makes the elimination (unlatching echo cancellation process) of carrying out echo at subtracting section 74 therefrom just with switch 76 conductings.In addition, under identical state, when from the echo that signal contained of input terminal 31 input during less than predetermined level, control part 77 just turn-offs switch 76, does not make the elimination (closing echo cancellation process) of carrying out echo at subtracting section 74 therefrom.Also have, when the echo that does not make subtracting section 74 elimination is carried out, also can switch 76 not turn-offed and the filter factor of adaptive filter 75 all is made as zero.In addition, control part 77 is according to the comparative result from echo level test section 72 output, when from the echo that signal contained of input terminal 31 inputs when very big, and will be with so that the warning signal of emergency warning lamp 28 bright lamps exports CPU25 to.
Then, the open and close control to the echo cancellation process of Echo Canceller 54 is elaborated.Fig. 7 is the flow chart of an example of the processing of calculating noise power that shows the Echo Canceller 54 of the 2nd example.Echo level test section 72 is when the power initiation of echo prevention device etc.; Testing result according to conversation test section 70,71; And do not have voice signal to send from the other side's side and state that the wearer of ear microphone 20 does not give orders or instructions under, calculate from the power Np (S301) of the noise of input terminal 31 inputs.Then, echo level test section 72 is that power Np with the noise of calculating is stored in noise level storage part 73 (S302).
Fig. 8 is the flow chart of an example of processing of open and close of decision echo cancellation process that shows the Echo Canceller 54 of the 2nd example.Echo level test section 72 is the testing results according to conversation test section 70,71; There is being voice signal to send from the other side's side; And under the state that the wearer of ear microphone 20 does not give orders or instructions, calculate from input terminal 31 and the power Ip (S401) of the input signal that comes.Then, echo level test section 72 is that the power Ip of input signal and the level Np that is stored in the noise of noise level storage part 73 are compared, and will represent the signal of comparative result to export control part 77 (S402) to.At the 2nd example, be carry out with alpha multiply by noise power Np gained value and input signal power Ip size relatively.Also have, alpha be than 1 big numerical value, and according to the suitable value of measured result setting of the power Ip of the power Np of noise and input signal.
When the comparative result of echo level test section 72 was Ip＞α Np, control part 77 was to judge to remain in from input terminal 31 and the amount of the echo of the input signal that comes is many, and opens the echo cancellation process (S403) of subtracting section 74.In addition; When the comparative result of echo level test section 72 is Ip≤α Np; Control part 77 be judge remain in from input terminal 31 and the amount of the echo of the input signal that comes for few, differential amplifier circuit 12 is to have eliminated echo fully, and closes the echo cancellation process (S404) of subtracting section 74.And control part 77 is when all being set at zero and when closing echo cancellation process by the filter factor with adaptive filter 75, because the treating capacity of DSP3 alleviates, so output makes signal that the operating frequency of DSP3 reduces to CPU25.Also have, when the renewal of the filter factor of adaptive filter 75 is when proceeding under the state to close echo cancellation process by switch 76 is turn-offed, do not export with so that the signal that the operating frequency of DSP3 reduces.In addition; Control part 77 is all not to be set at zero mode with the filter factor with adaptive filter 75; And turn-off and when closing echo cancellation process by the renewal of ending filter factor and with switch 76, output makes signal that the operating frequency of DSP3 reduces to CPU25.
In addition; When the comparative result of echo level test section 72 be Ip＞＞during α Np; Control part 77 is to judge to remain in from input terminal 31 and the amount of the echo of the input signal that comes is many for very, and output indicates the warning signal of bright lamp of emergency warning lamp 28 to CPU25 (S406).By emergency warning lamp 28 bright lamps, the wearer of ear microphone 20 just can discover is just having very large echo to be sent to the other side's side.
Behind emergency warning lamp 28 bright lamps; When the wearer of ear microphone 20 presses study button 27 (S407: be); Impulse response obtains portion 52 and just obtains impulse response (S408) again, and filter coefficient setting portion just sets the filter factor (S409) of FIR filter 50,51 again according to the impulse response that obtains again.Behind the filter factor of setting FIR filter 50,51 again, carry out once again from input terminal 31 and the calculating of the power Ip of the input signal that comes (S401) and with the comparison (S402) of the power Np of noise.Then, carry out and aforementioned identical determination step, if the words of Ip＞α Np are just opened echo cancellation process (S403), if the words of Ip≤α Np are just closed echo cancellation process (S404) at control part 77.Afterwards, control part 77 is that the signal of light-off with indication emergency warning lamp 28 exports CPU25 (S410) to.Also have, control part 77 makes impulse response obtain portion 52 with also can not making emergency warning lamp 28 bright lamps and obtains impulse response again, make filter coefficient setting portion 53 set the filter factor of FIR filter 50,51 again.
In addition, behind emergency warning lamp 28 bright lamps, (S407: not), control part 77 is just opened the echo cancellation process (S411) of subtracting section 74, and indication emergency warning lamp 28 turn off the light (S410) when learning button 27 when having passed through the scheduled time and not being pressed yet.
Fig. 9 is the routine calcspar of other formation that shows the Echo Canceller 54 of the 2nd example.Echo Canceller 54 is to possess the switch 76 of switch 78 to replace formation shown in Figure 6.
Echo level test section 72 is identical ground during with formation shown in Figure 6; Under the state that the wearer of ear microphone 20 does not give orders or instructions there being voice signal to send from the other side's side, detect from the signal level Ip of the signal (the 2nd input signal) of input terminal 31 inputs.In addition, echo level test section 72 is the testing results according to conversation test section 70,71, and under the state that the wearer of ear microphone 20 does not give orders or instructions there being voice signal to send from the other side's side, detects from the signal level Sp of the signal of subtracting section 74 outputs.And echo level test section 72 is the comparative results with signal level with the signal level of the signal of exporting from subtracting section 74 of the signal of importing from input terminal 31, that is the signal of the degree of the echo elimination of demonstration subtracting section 74 exports control part 77 to.
Also have, identical ground during with formation shown in Figure 6, echo level test section 72 are to detect power in scheduled period of signal as signal level.But in the signal level that echo level test section 72 detects not is to be defined as power, as long as just for example can show that the level person of signals such as absolute value sum total of peak swing or the amplitude of scheduled period of scheduled period can.
In addition, when switch 78 switched to the B side, control part 77 also can be ended the adaptability of the filter factor of adaptive filter 75 and upgrade.In addition, control part 77 also can all be set at zero at the filter factor that makes switch 78 switch under the state of A side adaptive filter 75, does not eliminate or attenuate echo at subtracting section therefrom.And control part 77 is when the adaptability of the filter factor of ending adaptive filter 75 is upgraded when all being set at zero (also comprise filter factor), because the treating capacity of DSP3 alleviates, therefore output makes signal that the operating frequency of DSP3 reduces to CPU25.
Also have, when switch 78 switches to the B side, also can continue the adaptability of the filter factor of adaptive filter 75 and upgrade.Under this situation, are detections of proceeding the degree that echo eliminates at echo level test section 72, the degree of eliminating when echo becomes predetermined level when above, just switch 78 is switched to the A side.That is when the degree of eliminating when the echo of subtracting section 74 was improved, just output was eliminated or the signal of attenuate echo at subtracting section 74, and can eliminate echo effectively.
More than, be illustrated to the 2nd example.As previously mentioned; Echo prevention device at the 2nd example; Be signal level,, just do not carry out the interior echo cancellation process of DSP3 when being judged to be differential amplifier circuit 12 when eliminating echo fully (Ip≤α Np) according to the input signal that comes from input terminal 31; When the elimination that is judged to be echo is also insufficient (Ip＞α Np), just carry out the echo cancellation process in the DSP3.In view of the above, when differential amplifying circuit 12 is eliminated echo fully, just be suppressed at the increase of carrying out the echo that echo cancellation process caused in the DSP3, and can eliminate echo effectively.In addition, when differential amplifying circuit 12 can't be eliminated echo fully, just, promptly can eliminate echo effectively by in DSP3, carrying out echo cancellation process.
In addition, in the echo prevention device of the 2nd example, upgrade, and will export lead-out terminal 34 to, can not carry out the echo cancellation process in the DSP3 from the signal of input terminal 31 inputs by the adaptability of the filter factor of ending adaptive filter 75.Under this situation, just no longer need upgrade filter factor, can alleviate the treating capacity of Echo Canceller 54, and suppress consumed power in response to the output signal adaptive ground of subtracting section 74.
And, in the echo prevention device of the 2nd example, all be made as zero by filter factor with adaptive filter 75, can not carry out the echo cancellation process in the DSP3.Under this situation, can under state, not carry out the echo cancellation process in DSP3 with switch 76 conductings.
In addition, when the adaptability of the filter factor of ending adaptive filter 75 is upgraded, cooperate alleviating of treating capacity and the operating frequency of DSP3 is reduced, can further suppress consumes electric power therefrom.
In addition, in the echo prevention device of the 2nd example, the update processing of the filter factor of adaptive filter 75 also can make under its state of proceeding, by switch 76 is turn-offed, and the echo cancellation process in DSP3 is closed.Under this situation because the update processing of the filter factor of adaptive filter 75 is to proceed, so when switch 76 when shutoff becomes conducting, can eliminate echo effectively by horse back.
In addition; In the echo prevention device of the 2nd example; Be according to the signal level of the input signal that comes from input terminal 31 and signal level from the signal of subtracting section 74 outputs, when the echo of subtracting section 74 is eliminated degree when high than predetermined level (Ip＞α Sp), the just output signal at subtracting section 74 echo having been eliminated or decayed; When echo is eliminated degree when low than predetermined level (Ip≤α Sp), just will be from input terminal 31 and input signal directly export.That is, being only limited to the interior echo of DSP3 and eliminating the effect that a certain degree is arranged, the echo processing in DSP3 is effectively, and can eliminate echo effectively.
In addition; In the echo prevention device of the 2nd example, when the echo processing degree of subtracting section 74 is low than predetermined level, just ends the adaptability of the filter factor of adaptive filter 75 and upgrade; And switch 78 is switched to the B side, can not carry out the echo cancellation process in DSP3 thus.Under this situation, just no longer need upgrade filter factor, can alleviate the treating capacity of Echo Canceller 54, and suppress consumed power in response to the output signal adaptive ground of subtracting section 74.
In addition, in the echo prevention device of the 2nd example, when the echo processing degree of subtracting section 74 is low than predetermined level, all be made as zero, can not carry out the echo cancellation process in DSP3 by filter factor with adaptive filter 75.Under this situation, can under state, not carry out the echo cancellation process in DSP3 with switch 76 conductings.
In addition, when the echo processing degree of subtracting section 74 was low than predetermined level, when the adaptability of the filter factor of ending adaptive filter 75 was upgraded, the operating frequency that cooperates alleviating of treating capacity to make DSP3 reduced, and can further suppress consumed power therefrom.
In addition; In the echo prevention device of the 2nd example; When the echo processing degree of subtracting section 74 is low than predetermined level; Update processing at the filter factor that makes adaptive filter 75 also can make under its state of proceeding, by switch 78 is switched to the B side, and the echo cancellation process in DSP3 is closed.Under this situation; Because the update processing of the filter factor of adaptive filter 75 is to proceed; So when the degree of eliminating when the echo of subtracting section 74 was improved, just output was eliminated echo or the signal of decay at subtracting section 74, and can eliminate echo effectively.
In addition, in the echo prevention device of the 2nd example, when the echo that input signal contained of coming from input terminal 31 is very big (Ip＞＞α Np), carry out obtaining again and the setting again of the filter factor of FIR filter 50,51 of impulse response.That is, when for example The noise etc. causes the low precision of the impulse response that obtains, can't carry out echo effectively when eliminating, just carry out obtaining again and the setting again of filter factor of impulse response, and can eliminate echo effectively.
In addition, in the echo prevention device of the 2nd example, when the echo that input signal contained of coming from input terminal 31 is very big (Ip＞＞α Np), can make emergency warning lamp 28 bright lamps.In view of the above, can let the wearer of ear microphone 20 can find just having very large echo to be sent to the other side's side, and impel and press obtaining again and the setting again of the filter factor of FIR filter 50,51 of impulse response that study button 27 etc. carries out.Also have, being used for making the wearer of ear microphone 20 to discover the means that just have very large echo to produce not is to be defined as to make emergency warning lamp 28 bright lamps.For example, also can make emergency warning lamp 28 be changed to the light-off state, also can make the bright lamp change color of emergency warning lamp 28 from lighting lamp state.In addition, for example, also exportable warning.
" the 3rd example "
==all formations==
At first, describe to the formation that is the echo prevention device of an example of the present invention.Figure 10 is the calcspar of formation of the echo prevention device of the 3rd example.Also having, is mark same components symbol and omission explanation with the identical formation of the 1st or the 2nd example.
DSP3 contains input terminal 30,31, lead-out terminal 32 to 34, DSP core 40, RAM (Random Access Memory) 41 and ROM (Read Only Memory) 42 and constitutes.In addition, DSP3 be possess FIR filter 50,51, impulse response obtains portion 52, filter coefficient setting portion 53, Echo Canceller (EC) 54, noise eliminator (NC) 55 and noise eliminator control part (NC control part) 56.It is to carry out the program that is stored in RAM41 or ROM42 by DSP core 40 to realize that those FIR filters 50,51, impulse response obtain portion 52, filter coefficient setting portion 53, Echo Canceller 54, noise eliminator 55 and noise eliminator control part 56.In addition, the filter factor of FIR filter 50,51 is to be stored in RAM41.
Noise eliminator 55 (noise removing portion) is the signal from Echo Canceller 54 outputs to be imposed noise removing handle output again.Noise eliminator 55 is the formations that for example can use the gimmick that is disclosed in " Y.Ephraim; D.Malah, " Speech enhancement using a minimum mean-square error short-time spectral amplitude estimator ", IEEE Trans.on ASSP; vol.ASSP-32; No.6, pp.1109-1121,1984 ".In the noise removing of noise eliminator 55 is handled; It is the input signal that for example accumulates scheduled volume; And the data of accumulating are calculated the signal noise ratio of each frequency band; According to this signal noise ratio decision noise suppressed coefficient, this coefficient multiply by the amplitude composition of input signal spectrum on frequency domain, carry out the elimination of noise therefrom.Also have, it is can open and close according to the control of NC control part 56 that the noise removing of noise eliminator 55 is handled, and when noise removing is handled when closing, is to export lead-out terminal 34 without delay to from the signal of Echo Canceller 54 outputs.
NC control part 56 (output control part) is to control the open and close that the noise removing of noise eliminator 55 is handled in response to remaining in from the signal level of the echo of the signal of Echo Canceller 54 output.For example, be predetermined level when above when remaining in from the signal level of the echo of the signal of Echo Canceller 54 output, if be sent to the other side's side, then there is the first speaker that makes the other side's side to feel the anxiety of inharmonious sense because of noise eliminator 55 makes echo delay.Therefore, under situation so, noise eliminator control part 56 is the noise removing of noise eliminator 55 to be handled close, and thus, makes from the signal of Echo Canceller 54 outputs and exports lead-out terminal 34 without delay to.On the contrary, when remaining in signal level from the echo of the signal of Echo Canceller 54 output less than the prearranged signals level, even be sent to the other side's side because of noise eliminator 55 makes echo delay, the first speaker of the other side's side has not probably thought inharmonious sense yet.Therefore, under this situation, noise eliminator control part 56 is the noise removing of noise eliminator 55 to be handled open, thus, will impose echo cancellation process to signal from Echo Canceller 54 outputs and signal export lead-out terminal 34 to.
The open and close control of==noise eliminator==
Then, the open and close control to the noise eliminator 55 of the 3rd example is elaborated.Figure 11 is the calcspar of the example that shows that the periphery of the noise eliminator 55 of the 3rd example constitutes.Echo Canceller 54 is to contain subtracting section 72, adaptive filter (ADF; Adaptive Digital Filter) 73 and adaptive filter control part (ADF control part) 74 and constituting.In addition, DSP3 possesses conversation test section 70,71 again and constitutes.
In conversation test section 70 are input signals from 50 outputs of FIR filter.Whether conversation test section 70 is to detect according to the signal from 50 outputs of FIR filter to have voice signal to send from the other side's side.For example; Conversation test section 70 is if be the words more than the predetermined level from the signal level of the signal of FIR filter 50 output; Just be judged to be voice signal and sent,, just be judged to be the voiceless sound signal and send from the other side's side if less than predetermined level from the other side's side.In addition, the signal of exporting from FIR filter 50 is to detect processing at conversation test section 70, and exports lead-out terminal 32 to.
In conversation detecting element 71 are input signals from subtracting section 72 outputs.Conversation test section 71 is the telephone conditions that detect ear microphone 20 wearers according to the signal from subtracting section 72 outputs.For example; Conversation test section 71 is if be the words more than the predetermined level from the signal level of the signal of subtracting section 72 output; Just the wearer who is judged to be ear microphone 20 gives orders or instructions, if less than predetermined level, the wearer who just is judged to be ear microphone 20 does not give orders or instructions.In addition, the signal of exporting from subtracting section 72 is to detect processing at conversation test section 71, and exports noise eliminator 55 to.
Subtracting section 72 is that output is with the signal z (n) that the contains echo (echo signal: the 2nd input signal) deduct from the signal e (n) (elimination error signal) of signal y (n) (virtual echo signal) gained of adaptive filter 73 outputs from input terminal 31 inputs.
In adaptive filter 73 are inputs from the signal x (n) of FIR filter 50 output (reference signal: the 1st input signal), with the output signal e (n) of subtracting section 72.And adaptive filter 73 is under the state that the wearer of ear microphone 20 does not give orders or instructions there being voice signal to send from the other side's side, changes filter factor adaptively so that the signal e (n) that exports from subtracting section 72 becomes below the predetermined level.Also have, about the setting action of the formation of adaptive filter 73 and filter factor be can be for example with the formation of the adaptive filter that is disclosed in TOHKEMY 2006-304260 communique and move identical.
NC control part 56 is the testing results according to conversation test section 70,71, and under the state that the wearer of ear microphone 20 does not give orders or instructions there being voice signal to send from the other side's side, detects from the signal level of the signal e (n) of subtracting section 72 outputs.The signal level of detected signal e (n) is not eliminated the level of the remaining echo of getting off (echo that has been attenuated) to the greatest extent for representative Echo Canceller 54 under this state.And; NC control part 56 is that the signal level as signal e (n) is that predetermined level is when above; The noise removing of noise eliminator 55 handled close,, the noise removing of noise eliminator 55 handled open when the signal level of signal e (n) during less than the prearranged signals level.Here, when also comprising, the signal level of signal e (n) situation during less than the prearranged signals level fully echo is eliminated and the situation of the signal level of signal e (n) when being zero.Also have, the signal level of the signal e (n) that detects be for example the maximum power of scheduled period or the peak swing of scheduled period, the absolute value sum total of the amplitude of scheduled period etc. can shows signal e (n) level person just can.
Figure 12 is the flow chart of an example of the processing that shows that the startup of the noise eliminator 55 of control the 3rd example is closed.At first, when electric power starting and echo prevention device when starting (S301), NC control part 56 is just handled the noise removing of noise eliminator 55 and is closed.Afterwards, the wearer of beginning ear microphone 20 and the conversation (S303) between the other side's side.After the echo prevention device starting, the adaptability of the filter factor of adaptive filter 73 is upgraded and still can't be finished at once, and can become mostly at the state that from the signal e (n) of Echo Canceller 54 outputs, contains the echo of a certain degree.Therefore, after echo prevention device starting, immediately noise eliminator 55 cuts out, thus, just do not have and handle the echo of taking place to postpone because of noise removing and be sent to the facts of the other side's side, and can suppress the inharmonious sense that causes because of the echo of delay.
After the conversation beginning; NC control part 56 is the testing results according to conversation test section 70,71; And under the state that the wearer of ear microphone 20 does not give orders or instructions there being voice signal to send from the other side's side, detect from the signal level E (S304) of the signal e (n) of Echo Canceller 54 outputs.Then, NC controller 56 is more detected signal level E and the threshold value Sh (S305) that representes predetermined level.
When the signal level E of signal e (n) is that (S305:E >=Sh), just handle the noise removing of noise eliminator 55 close (S306) when above by NC control part 56 for threshold value Sh.That is; When becoming when the signal e (n) from Echo Canceller 54 output contains the state of echo of a certain degree; Just the noise removing of noise eliminator 55 is handled and closed; Thus, just do not have and handle the echo postpones because of noise removing and be sent to the facts of the other side's side, and can suppress the inharmonious sense that causes because of the echo of delay.
(S305:E＜Sh=, NC control part 56 just handle the noise removing of noise eliminator 55 and open (S307) during less than threshold value Sh as the signal level E of signal e (n).That is, when the echo of containing in signal e (n) from Echo Canceller 54 output very hour, just the noise removing of noise eliminator 55 is handled and is opened, by this, transmit eliminated noise signal clearly to the other side's side.Under this situation, because the echo that signal e (n) is contained is very little, so even there is the echo of delaying because of the noise removing processing of noise eliminator 55 to be sent to the other side's side, the other side's side feels that the possibility of inharmonious sense also can step-down.
More than, be illustrated to the 3rd example.As previously mentioned; In the 3rd example; The signal level of the echo of being contained as the signal e (n) from Echo Canceller 54 output is predetermined level when above, is not impose the noise removing processing of noise eliminator 55 and export from the signal of Echo Canceller 55 outputs.On the other hand, the signal level of the echo of being contained as the signal e (n) from Echo Canceller 54 output is during less than predetermined level, is output again after the noise removing that imposes noise eliminator 55 is handled from the signal of Echo Canceller 55 outputs.That is; First speaker to the other side's side when echo delay transmits is felt the degree of inharmonious sense; Be the signal level of the echo contained as signal e (n) when big, handle the delay that causes because of noise removing and produce by not making, and can reduce the inharmonious sense that causes because of echo.On the other hand; Can not feel the degree of inharmonious sense even transmit the first speaker of the other side's side to echo delay yet; Be the echo contained as signal e (n) signal level hour, handle by implementing noise removing, and can the signal processing that be sent to the other side's side be become signal more clearly.
In addition, in the echo prevention device of the 3rd example, Echo Canceller 54 is to use by making the filter factor adaptations make the adaptive filter 73 that echo reduces gradually and constitutes.Therefore, start from echo prevention device that to be updated to suitable value to the filter factor of adaptive filter 73 be to take the time of a certain degree.So it is especially effective that the signal level of the echo of being contained in response to the signal e (n) from Echo Canceller 54 output comes the open and close noise removing to handle.
In addition, in the echo prevention device of the 3rd example, when starting, be that the noise removing of noise eliminator 54 is closed at power connection.That is, owing to be not the state that has been updated to suitable value for the filter factor of adaptive filter 73 when echo prevention device starts, so by being handled, noise removing closes, and can suppress the first speaker of the other side's side is transmitted the signal that inharmonious sense is arranged.
More than, aforesaid invention example is to explain in order to understand the present invention easily, is not in order to limited interpretation the present invention.Under the prerequisite that does not exceed purport of the present invention, can change, improve, and the present invention also comprises its equivalent.
For example, in the 1st to the 3rd example, though be the filter factor that the impulse response that obtains by pulse is produced is set at the FIR filter, the signal that uses for the setting filter factor is not to be defined as pulse.The response signal that for example, also can obtain when step (step) signal is produced is set the filter factor of FIR filter.
In addition; In the 1st to the 3rd example; Though be to describe to the echo prevention device that together uses with ear microphone 20, be not to be defined as ear microphone 20, the device that needs only to complying with voice signal foldback echo to the other side's side that the other side's side sends just can be suitable for.For example, when using phone, even transmit the constituting of echo to the other side's side by microphone input, also can be suitable for echo prevention device from the sound of the loud speaker output of phone from phone with hand-free mode.
Claims (4)
1. filter coefficient setting device is a filter coefficient setting device of setting the filter factor of echo prevention device, and aforementioned echo prevention device possesses and has:
The 1st finite impulse response filter receives the input of the 1st digital signal, and exports the 2nd digital signal;
The 2nd finite impulse response filter receives the input of aforementioned the 1st digital signal simultaneously with aforementioned the 1st finite impulse response filter, and exports the 3rd digital signal;
The 1st digital/analog converter converts aforementioned the 2nd digital signal to the 1st analog signal and output;
The 2nd digital/analog converter converts aforementioned the 3rd digital signal to the 2nd analog signal and output;
Output/input terminal is exported aforementioned the 1st analog signal or is imported the 3rd analog signal;
Subtracting section, output will aforementioned the 1st analog signal and the signal that combines of aforementioned the 3rd analog signal deduct aforementioned the 2nd analog signal and must the 4th analog signal; And
Analog/digital converter will become digital signal and output from the analog signal conversion of aforementioned subtracting section output; And
This filter coefficient setting device is characterized in that, possesses:
The filter factor initial setting section when aforementioned echo prevention device starts, is set predetermined filter factor to the aforementioned the 1st and the 2nd finite impulse response filter;
Response signal obtains portion;, the 1st signal obtains output by being produced from aforementioned the 1st finite impulse response filter; Arrive the 1st response signal of the output of aforementioned analog/digital converter; And obtain output from aforementioned the 2nd finite impulse response filter by the 2nd signal is produced, to the 2nd response signal of the output of aforementioned analog/digital converter; And
Filter coefficient setting portion; For aforementioned the 4th analog signal is become; From the signal that aforementioned the 1st analog signal and aforementioned the 3rd analog signal combine, eliminate or aforementioned the 1st analog signal that decays after and signal; And to the filter factor of aforementioned the 1st response signal gained of aforementioned the 2nd finite impulse response filter set basis, and to the filter factor of aforementioned the 2nd response signal gained of aforementioned the 1st finite impulse response filter set basis;
When aforementioned filter factor initial setting section is started at aforementioned echo prevention device; As in response in the aforementioned the 1st and the 1st and the 2nd response signal data of the 2nd response signal when being stored in memory; For aforementioned the 4th analog signal is become; From the signal that aforementioned the 1st analog signal and aforementioned the 3rd analog signal combine, eliminate or aforementioned the 1st analog signal that decays after and signal; And aforementioned the 2nd finite impulse response filter set basis there is the filter factor of aforementioned the 1st response signal gained; And, when not having the aforementioned the 1st and the 2nd response signal data to be stored in aforementioned memory, the aforementioned the 1st and the 2nd finite impulse response filter is set aforementioned predetermined filter factor to the filter factor of aforementioned the 1st finite impulse response filter set basis by aforementioned the 2nd response signal gained;
Aforementioned response signal obtains portion when not having the aforementioned the 1st and the 2nd response signal data to be stored in aforementioned memory, obtains the aforementioned the 1st and the 2nd response signal.
2. filter coefficient setting device as claimed in claim 1; It is characterized in that; When not having the aforementioned the 1st and the 2nd response signal data to be stored in aforementioned memory, aforementioned filter factor initial setting section with the aforementioned the 1st and the filter coefficient setting of the 2nd finite impulse response filter be zero.
3. filter coefficient setting device as claimed in claim 1; It is characterized in that; Wherein aforementioned response signal obtains portion when obtaining the aforementioned the 1st and the 2nd response signal, will be stored in aforementioned memory in response to the aforementioned the 1st and the 2nd response signal data in the aforementioned the 1st and the 2nd response signal.
4. filter coefficient setting device as claimed in claim 1 is characterized in that wherein also possessing:
The warning efferent, when not having the aforementioned the 1st and the 2nd response signal data to be stored in aforementioned memory, the output warning signal.
Applications Claiming Priority (9)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2007-004814 | 2007-01-12 | ||
| JP2007004814 | 2007-01-12 | ||
| JP2007004814A JP2008172612A (en) | 2007-01-12 | 2007-01-12 | Filter coefficient setting device and program |
| JP2007-050337 | 2007-02-28 | ||
| JP2007050337A JP4887181B2 (en) | 2007-02-28 | 2007-02-28 | Echo prevention device and program |
| JP2007050337 | 2007-02-28 | ||
| JP2007058962A JP4829816B2 (en) | 2007-03-08 | 2007-03-08 | Echo prevention device and program |
| JP2007-058962 | 2007-03-08 | ||
| JP2007058962 | 2007-03-08 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101222243A CN101222243A (en) | 2008-07-16 |
| CN101222243B true CN101222243B (en) | 2012-10-03 |
Family
ID=39631859
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2008100028834A Expired - Fee Related CN101222243B (en) | 2007-01-12 | 2008-01-11 | Filter coefficient setting device and echo prevention device |
Country Status (2)
| Country | Link |
|---|---|
| JP (1) | JP2008172612A (en) |
| CN (1) | CN101222243B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109739199B (en) * | 2019-01-17 | 2021-02-19 | 玖龙纸业（太仓）有限公司 | Automatic change control system filter equipment and automatic control system |
| WO2023272631A1 (en) * | 2021-06-30 | 2023-01-05 | 深圳市汇顶科技股份有限公司 | Noise cancellation method for audio signal, audio signal processing apparatus, and electronic device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6434110B1 (en) * | 1998-03-20 | 2002-08-13 | Cirrus Logic, Inc. | Full-duplex speakerphone circuit including a double-talk detector |
| JP2004282434A (en) * | 2003-03-17 | 2004-10-07 | Matsushita Electric Ind Co Ltd | Device and method for canceling echo, and program and recording medium |
| US7003099B1 (en) * | 2002-11-15 | 2006-02-21 | Fortmedia, Inc. | Small array microphone for acoustic echo cancellation and noise suppression |
-
2007
- 2007-01-12 JP JP2007004814A patent/JP2008172612A/en active Pending
-
2008
- 2008-01-11 CN CN2008100028834A patent/CN101222243B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6434110B1 (en) * | 1998-03-20 | 2002-08-13 | Cirrus Logic, Inc. | Full-duplex speakerphone circuit including a double-talk detector |
| US7003099B1 (en) * | 2002-11-15 | 2006-02-21 | Fortmedia, Inc. | Small array microphone for acoustic echo cancellation and noise suppression |
| JP2004282434A (en) * | 2003-03-17 | 2004-10-07 | Matsushita Electric Ind Co Ltd | Device and method for canceling echo, and program and recording medium |
Also Published As
| Publication number | Publication date |
|---|---|
| JP2008172612A (en) | 2008-07-24 |
| CN101222243A (en) | 2008-07-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104396275B (en) | Use the wire and wireless earphone of insert type microphone in ear | |
| JP4975277B2 (en) | Noise reduction headset | |
| CN111836147B (en) | Noise reduction device and method | |
| US10382092B2 (en) | Method and system for full duplex enhanced audio | |
| KR102150844B1 (en) | A persnal audio device, and a method of canceling ambient audio sounds in the proximity of a transducer of a personal audio device | |
| US5099472A (en) | Hands free telecommunication apparatus and method | |
| US20090323925A1 (en) | System and Method for Telephone Based Noise Cancellation | |
| US20100166199A1 (en) | Acoustic echo reduction circuit for a "hands-free" device usable with a cell phone | |
| EP1944956B1 (en) | Filter coefficient setting device and echo prevention device | |
| TWI416506B (en) | Voice signal processing device | |
| US5058153A (en) | Noise mitigation and mode switching in communications terminals such as telephones | |
| JP2013529427A (en) | Speech enhancement method, apparatus and noise reduction communication headphones | |
| JP2009089397A (en) | Two-way communication device having single transducer, and method thereof | |
| ATE464738T1 (en) | ECHO CANCELLATION | |
| CN104126297A (en) | Audio signal processing in communication system | |
| JP5824404B2 (en) | headset | |
| KR20060130067A (en) | Echo canceller having a series arrangement of adaptive filters with individual update control strategy | |
| ATE476056T1 (en) | INTEGRATED NOISE CANCELLATION AND RESTECHNO SUPPRESSION | |
| JP2009219040A (en) | Echo canceller, method for cancelling echo, and program | |
| CN101222243B (en) | Filter coefficient setting device and echo prevention device | |
| JP4541159B2 (en) | Nonlinear acoustic echo canceller | |
| US8705758B2 (en) | Audio processing device and method for reducing echo from a second signal in a first signal | |
| US7035396B1 (en) | Configurable echo canceller | |
| JP4887181B2 (en) | Echo prevention device and program | |
| US20100029345A1 (en) | Acoustic echo reduction circuit for a "hands-free" device usable with a cell phone |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20121003 Termination date: 20220111 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |