JPH06284046A - Echo canceller - Google Patents

Abstract

PURPOSE:To determine a tap coefficient at any echo continuing time and to execute proper echo cancel by measuing the echo continuation time from an impulse response under a certain acoustic environment and determining the tap coefficient based upon the measured echo continuation time. CONSTITUTION:An echo continuation time measuring instrument 6 generates impulse from a speaker 8 under a certain acoustic environment and an impulse response from an acoustic system is inputted to a microphone 7 to measure echo continuation time. The equipment 6 determines the number of taps necessary for canceling an acoustic echo under the acoustic environment based upon the echo continution time obtained by the measurement and outputs the number of taps, which is inputted to a false echo generator 1. The generator 1 generates and outputs a false echo by a tap coefficient by the convolution integration of a remote speaker voice input signal and the tap coefficient based upon the number of taps outputted from the equipment 6. The tap coefficient obtained from the number of taps is set up as the initial value of the tap coefficient. Thus an echo canceller in this example cancels an echo by a false echo.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an echo canceller for canceling acoustic echo of a mobile phone or the like.

[0002]

2. Description of the Related Art FIG. 6 is a block diagram showing the configuration of a conventional echo canceller. In the figure, reference numeral 61 is a pseudo echo generator using an adaptive digital filter (ADF), which is a far-end speaker voice input signal and tap coefficient. A pseudo echo is generated by convolving with. 62 and 63 are analog-
Digital (A / D) converters 64 and 65 are digital-analog (D / A) converters. Reference numeral 67 is a microphone that converts sound into an electric signal, 68 is a speaker that converts an electric signal into acoustic energy, and 69 is a computing unit that subtracts pseudo echo from acoustic echo. As the echo canceller shown in FIG. 6, the reference "Oki Electric Research and Development" Hands-free call by echo canceller ", January 1, 1988,
No. 41, VOL. 56, No. There is one.

Next, the operation will be described. In the echo canceller shown in FIG. 6, the far-end speaker voice input signal is A /
It is converted into a digital signal by the D converter 63 and input to the pseudo echo generator 61. On the other hand, the digital signal-converted far-end speaker voice input signal is converted into an analog signal (far-end speaker voice output signal) again by the D / A converter 64, and the speaker is output as the far-end speaker voice output. It is output at 68.

The pseudo echo generator 61 generates and outputs a pseudo echo by convolutional integration of the far-end speaker voice input signal and the tap coefficient. This tap coefficient is updated by the far-end speaker voice input signal and the residual echo. Microphone 67
The acoustic echo input to is converted into a digital signal by the A / D converter 62. The calculator 69 subtracts the pseudo echo from the acoustic echo, and the result is converted into an analog signal by the D / A converter 65. Further, the update of the tap coefficient performed in the pseudo echo generator 65 uses the following equation (1) for each temporal sample.

[0005]

[Equation 1]

In the above equation (1), h is a tap coefficient and n is a sampling frequency of 8k when t is time.
A value that takes n = 8000t at Hz, I is the number of taps, α is a correction coefficient, e is a residual echo, and x is a far-end speaker voice input signal.

[0007]

Since the conventional echo canceller is constructed as described above, the number of taps of the echo canceller is determined from the echo duration in the acoustic environment where the echo canceller is expected to be installed. ing. In this case, there is a problem that the echo canceller cannot be applied at all in an acoustic environment having an echo duration longer than the echo duration, and the echo canceling function cannot be utilized. Further, by using the A / D conversion and the D / A conversion, there is a problem that a quantization error appears in the far-end speaker's voice output and the reproducibility of the voice is deteriorated. Furthermore, since the calculation of updating the tap coefficient for each sample is added, there is a problem that the calculation amount becomes large.

The invention of claim 1 is to solve the above problems, and an object thereof is to obtain an echo canceller capable of performing appropriate echo cancellation under any acoustic environment. To do.

It is an object of the present invention to provide an echo canceller capable of improving reproducibility of far-end talker voice.

An object of the present invention is to provide an echo canceller capable of reducing the amount of calculation required to update the tap coefficient at a certain time without changing the amount of echo cancellation.

According to a fourth aspect of the present invention, it is possible to perform appropriate echo cancellation in any acoustic environment, and to change the amount of calculation required to update the tap coefficient at a certain time without changing the amount of echo cancellation. The aim is to obtain an echo canceller which can be reduced.

According to the invention of claim 5, the reproducibility of the far-end speaker's voice can be improved, and the amount of calculation required to update the tap coefficient at a certain time can be reduced without changing the amount of echo cancellation. The purpose is to obtain an echo canceller capable of

[0013]

An echo canceller according to the invention of claim 1 comprises a structure for measuring an echo duration from an impulse response under a certain acoustic environment and determining the number of taps from the echo duration. It is a thing.

An echo canceller according to a second aspect of the present invention has a configuration for outputting a far-end speaker voice as an analog signal of a far-end speaker voice input signal.

An echo canceller according to a third aspect of the present invention has a configuration in which the tap coefficient is divided into a plurality of blocks and a part of all blocks is updated.

An echo canceller according to a fourth aspect of the invention has, in addition to the configuration of the first aspect of the invention, a configuration in which the tap coefficient is divided into a plurality of blocks and a part of all blocks is updated. is there.

An echo canceller according to a fifth aspect of the present invention has, in addition to the configuration of the second aspect of the invention, a configuration in which a tap coefficient is divided into a plurality of blocks and a part of all blocks is updated. is there.

[0018]

The echo canceller according to the invention of claim 1 is
Since the echo duration is measured from the impulse response under a certain acoustic environment and the tap coefficient is determined from the echo duration, the tap coefficient is determined for any echo duration.

The echo canceller according to the second aspect of the present invention has a configuration in which the far-end speaker voice input signal is output as an analog signal from the far-end speaker voice output. It is separately and independently executed based on the far-end speaker voice input signal in the state of an analog signal.

The echo canceller according to the third aspect of the present invention has a configuration in which the tap coefficient is divided into a plurality of blocks and a part of all blocks is updated. The scale will be reduced.

In the echo canceller according to the invention of claim 4, in addition to the configuration of the invention of claim 1, the tap coefficient is divided into a plurality of blocks and a part of all blocks is updated. In the invention of claim 1, the scale of software or hardware for updating the tap coefficient is reduced.

According to the echo canceller of the invention of claim 5, in addition to the configuration of the invention of claim 2, the tap coefficient is divided into a plurality of blocks and a part of all blocks is updated. In the invention of claim 2, the scale of software or hardware for updating the tap coefficient is reduced.

[0023]

EXAMPLES Example 1. An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of an echo canceller according to an embodiment of the present invention. In the figure, for example, 1 is an adaptive digital filter (AD).
The pseudo echo generator according to F) generates a pseudo echo by convolving the far-end speaker voice input signal and the tap coefficient. 2 and 3 are analog-digital (A / D) converters,
Reference numerals 4 and 5 are digital-analog (D / A) converters. Reference numeral 6 denotes an echo duration measuring device (impulse response measuring means, echo measuring means, determining means), which determines the number of taps used in the pseudo echo generator 1. This echo duration measuring device 6 is a CP operated by a program.
U, a ROM storing a program, and a RAM used as a work area of the CPU. 7 is a microphone (impulse response measuring means) for converting sound into an electric signal,
Reference numeral 8 is a speaker (impulse response measuring means) for converting an electric signal into acoustic energy, and 9 is a calculator for subtracting pseudo echo from acoustic echo. In addition, in the pseudo echo generator 1,
References Oki Electric Research and Development "Hands-free call with echo canceller", January 1988, No. 141, VO
L. 56, No. The pseudo echo generation technique disclosed in No. 1 is applied.

Next, the operation will be described. FIG. 2 is a flowchart illustrating the flow of tap number determination processing according to the first embodiment. The flow shown in FIG. 2 is mainly shown by the operation of the echo duration measuring device 6. Figure 1
In the echo canceller shown in FIG. 1, first, the echo duration measuring device 6 generates an impulse from the speaker 8 under a certain acoustic environment (step ST1), and inputs the impulse response of the acoustic system into the microphone 7 (step ST2). , Echo duration is measured (step ST3). Further, the echo duration measuring device 6 determines the number of taps required to eliminate the acoustic echo under the acoustic environment based on the echo duration obtained by the measurement (step ST
4) The tap number is output and input to the pseudo echo generator 1 (step ST5). In the pseudo echo generator 1, the far-end speaker voice input signal and the echo duration measuring device 6
A pseudo echo is generated and output by convolution integration with the tap coefficient based on the number of taps from. Here, with respect to the relationship between the number of taps and the tap coefficient, refer to the above-mentioned formula (1). The tap coefficient obtained from the number of taps is positioned as the initial value of the tap coefficient. In this way, the echo canceller of the first embodiment performs echo cancellation by pseudo echo. Note that the operation of generating a pseudo echo from the number of taps and performing echo cancellation is the same as that of the conventional technique, and therefore description thereof is omitted.

As described above, according to the first embodiment, the echo duration measuring device 6 does not determine the number of taps based on the echo duration under the expected acoustic environment, but determines what acoustic environment Even at the bottom, the tap number is determined based on the echo duration obtained therefrom, so that the tap number can be freely determined according to the varying echo duration. In other words, the echo canceller according to the first embodiment can perform appropriate echo cancellation under any acoustic environment.

Example 2. FIG. 3 is a block diagram showing the configuration of an echo canceller according to an embodiment of the invention of claim 2. In the figure, reference numeral 31 is a pseudo echo generator (echo cancellation processing means). This pseudo echo generator 31 is also referred to Oki Electric R & D “Hands-free call by echo canceller”, January 14, 1988, in Reference.
No. 1, VOL. 56, No. The pseudo echo generation technique disclosed in No. 1 is applied. 32 and 33 are A / D converters, 34 is a far-end speaker voice input signal,
A signal line (transmission means) input to the D converter 33 and the speaker 38, 35 is a D / A converter for converting a digital signal into an analog signal, and 36 is a far-end talk converted into a digital signal by the A / D converter 33. This is a signal line for transmitting a human voice input signal to the pseudo echo generator 31, which is the main configuration of the echo cancellation processing. Reference numeral 37 is a microphone, and 38 is a speaker (voice generating means). In the second embodiment, the far-end speaker voice input signal on the signal line 34 is directly transmitted to the speaker 38 in the state of an analog signal. Therefore, the same configuration as the D / A converter 4 shown in FIG. 1 is used. Is unnecessary.

Next, the operation will be described. In the echo canceller shown in FIG. 3, the far-end speaker voice input signal passes through the signal line 34 and is input to the A / D converter 33 and the speaker 38, respectively. The far-end talker voice input signal input to the A / D converter 33 is converted into a digital signal, and is input to the pseudo echo generator 31 through the signal line 36. Then, the same pseudo echo generation (echo cancellation processing) as in the conventional case is performed. The far-end speaker voice input signal input to the speaker 38 is in an analog signal state on the signal line 34 without performing analog-digital conversion and digital-analog conversion. In this way, the speaker 38
Produces a far-end speaker audio output from.

As described above, according to the second embodiment, the input far-end talker voice input signal is subjected to analog-digital conversion,
Further, since the signal can be directly transmitted to the speaker 38 without performing the digital-analog conversion, it is possible to remove an error occurrence such as a quantization error generated when the analog-digital conversion and the digital-analog conversion are performed. In other words, the echo canceller according to the second embodiment can improve the reproducibility of the far-end talker voice.

Example 3. FIG. 4 is a block diagram showing a configuration relating to tap coefficient updating of an echo canceller according to an embodiment of the invention of claim 3, wherein, for example, 41 is a tap coefficient block dividing unit (dividing means) at a certain time n. The tap coefficient in is divided into M (M is a natural number) blocks. B _{1 to} B _M are tap coefficient block division units 41
It is a block divided by. A block extracting unit (extracting means) 42 extracts one block from the blocks B _{1 to} B _M. Reference numeral 43 denotes a tap coefficient calculation unit (updating unit), which calculates only one block B _i (1 ≦ i ≦ M) extracted by the block extraction unit 42 using the far-end speaker voice input signal and the residual echo. . i is the number of the block to be calculated. The above-mentioned equation (1) is applied to this calculation.

Next, the operation will be described. FIG. 5 is a flowchart illustrating the flow of tap coefficient update processing according to the third embodiment. In the echo canceller shown in FIG. 4,
The tap coefficient at time n in the tap coefficient block division unit 41 is input and divided into M blocks (step ST51). Then, the block extracting unit 42 extracts one block for updating the tap coefficient (step S
(T52), the tap coefficient calculation unit 43 executes the calculation for the one block (step ST53). In this case, mod (n, M) (remainder of dividing n by M) is used. When mod (n, M) = 0, tap coefficient h
For (n, M), the calculation for each tap coefficient is executed for i = M, 2M, 3M ... That is, in FIG. 4, in the block B ₁ .about.B _M obtained by dividing one tap coefficients, block B _M which initially with number M divided is extracted in block extraction section 42 and sent to the tap coefficient calculation unit 43. With respect to the tap coefficient to be input next, a block having a division number of 2M is a target of calculation, and thereafter, every time a tap coefficient is input, block extraction and calculation are similarly executed.

When mod (n, M) = 1,
The tap coefficient h (n, i) is i = 1, M + 1, 2M + 1 ...
When mod (n, M) = 2, the tap coefficient h (n, i) is calculated for each tap coefficient for i = 2, M + 2, 2M + 2. mod (n,
M) = 3 and thereafter are similarly executed, and for example, mod (n,
When M) = M-1, the tap coefficient h (n, i) is i
= M−1, 2M−1, ..., Calculation for each tap coefficient is executed.

As described above, according to the third embodiment, the tap coefficient for one temporal sample is divided into M blocks, and only one of the blocks is updated to update the entire tap coefficient. Therefore, the amount of calculation required to update the tap coefficient can be reduced without changing the echo cancellation amount. The tap coefficient update can be processed from both sides of software and hardware, and this reduction in the amount of calculation is effective in reducing the scale of software or hardware.

In the third embodiment described above, only one block of tap coefficients divided into a plurality of blocks is updated, but even if two or more blocks are updated, the echo cancellation amount does not change, The amount of calculation required to update the tap coefficient can be reduced. Also in this case, two or more blocks among the 1 to M blocks may be regularly extracted like the relationship between mod (n, M) and i described above.

Example 4. Now, it is also possible to apply the above-described third embodiment to the first embodiment. The configuration of the echo canceller in this case is a combination of FIG. 1 and FIG. Specifically, the configuration of the fourth embodiment can be obtained by incorporating the configuration of FIG. 4 into the pseudo echo generator 1 of the echo canceller shown in FIG.

Next, the operation will be described. In the echo canceller of the fourth embodiment, after the number of taps is determined according to the flow shown in FIG. 2, the above equation (1) is calculated from the number of taps.
Is used to obtain the tap coefficient, and the tap coefficient is M (M
One block to be updated is extracted by dividing the block into ≧ 2) blocks and extracting one block from the M blocks. Further, the extracted one block is updated to update the entire tap coefficient.

As described above, according to the fourth embodiment, appropriate echo cancellation can be performed under any acoustic environment, and the tap coefficient can be updated at a certain time without changing the echo cancellation amount. The amount of calculation can be reduced.

Example 5. Now, it is also possible to apply the third embodiment described above to the second embodiment. The configuration of the echo canceller in this case is a combination of FIG. 3 and FIG. Specifically, the configuration of the fifth embodiment can be obtained by incorporating the configuration of FIG. 4 into the pseudo echo generator 31 of the echo canceller shown in FIG.

Next, the operation will be described. In the echo canceller of the fifth embodiment, as is apparent from the configuration of the echo canceller shown in FIG. 3, the far-end speaker voice input signal on the signal line 34 is bidirectional between the speaker 38 and the A / D converter 32. And processed by each. That is, in one of the speakers 38, the far-end speaker voice input signal is converted from an electric signal into acoustic energy and generated as a far-end speaker voice output. The far-end speaker voice input signal transmitted to the other A / D converter 33 is converted into a digital signal and subjected to echo cancellation processing by the pseudo echo generator 1. In this echo cancellation processing, the tap coefficient and the far-end speaker voice input signal are used for pseudo echo generation. The tap coefficient is divided into M (M ≧ 2) blocks and the M
By extracting one block from this block, one block to be updated is extracted. Further, the extracted one block is updated to update the entire tap coefficient.

As described above, according to the fifth embodiment, the reproducibility of the far-end speaker's voice is improved by not performing the analog-digital conversion and the digital-analog conversion on the far-end speaker's voice input signal. It is possible to reduce the amount of calculation required to update the tap coefficient at a certain time without changing the amount of echo cancellation.

[0040]

As described above, according to the invention of claim 1, the echo duration is measured from the impulse response under a certain acoustic environment, and the tap coefficient is determined from the echo duration. The tap coefficient is determined for any echo duration. Therefore, there is an effect that appropriate echo cancellation can be performed under any acoustic environment.

According to the second aspect of the invention, since the far-end speaker voice input signal is an analog signal and the far-end speaker voice is output as it is, the far-end speaker voice output and the echo canceling process are analog signals. It is executed separately and independently based on the far-end speaker voice input signal of the state. Therefore, the reproducibility of the far-end speaker's voice can be improved.

According to the invention of claim 3, since the tap coefficient is divided into a plurality of blocks and a part of all the blocks is updated, the scale of software or hardware for updating the tap coefficient is reduced. To be done. Therefore, there is an effect that the amount of calculation required to update the tap coefficient at a certain time can be reduced without changing the amount of echo cancellation.

According to the invention of claim 4, in addition to the configuration of the invention of claim 1, the tap coefficient is divided into a plurality of blocks and a part of all blocks is updated. In the invention, the scale of software or hardware for updating the tap coefficient is reduced. Therefore, it is possible to perform appropriate echo cancellation under any acoustic environment, and it is possible to reduce the calculation amount required to update the tap coefficient at a certain time without changing the echo cancellation amount. There is.

According to the invention of claim 5, in addition to the configuration of the invention of claim 2, the tap coefficient is divided into a plurality of blocks and a part of all blocks is updated. In the invention, the scale of software or hardware for updating the tap coefficient is reduced. Therefore, the reproducibility of the far-end talker voice can be improved, and the amount of calculation required to update the tap coefficient at a certain time can be reduced without changing the amount of echo cancellation.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of an echo canceller according to an embodiment of the present invention.

FIG. 2 is a flowchart illustrating a flow of tap number determination processing according to the first embodiment.

FIG. 3 is a block diagram showing a configuration of an echo canceller according to an embodiment of the invention of claim 2;

FIG. 4 is a block diagram showing a configuration related to tap coefficient updating of an echo canceller according to an embodiment of the invention of claim 3;

FIG. 5 is a flowchart illustrating a flow of tap coefficient update processing according to the third embodiment.

FIG. 6 is a block diagram showing a configuration of a conventional echo canceller.

[Explanation of symbols]

6 Echo duration measuring device (impulse response measuring means, echo measuring means, determining means) 7 Microphone (impulse response measuring means) 8 Speaker (impulse response measuring means) 31 Pseudo echo generator (echo cancellation processing means) 34 Signal line ( Transmitting means) 38 Speaker (sound generating means) 41 Tap coefficient block dividing section (dividing means) 42 Block extracting section (extracting means) 43 Tap coefficient calculating section (update means)

Claims (5)

[Claims] 1. An echo canceller for canceling an acoustic echo based on the number of taps, an impulse response measuring means for measuring an impulse response under a certain acoustic environment, and an echo measurement for measuring an echo duration from the measured impulse response. An echo canceller comprising: a means and a determining means for determining the number of taps based on the measured echo duration. 2. Transmission means for transmitting a far-end speaker voice input signal which is an analog signal, and far-end talk based on the far-end speaker voice input signal connected to said transmission means and transmitted by said transmission means. Voice generating means for generating a person's voice output, and the far end transmitted by the transmitting means when performing echo cancellation processing based on the far-end talker voice input signal transmitted by the transmitting means. An echo canceller having an echo cancellation processing means for converting an end speaker voice input signal into a digital signal. 3. An echo canceller for canceling an acoustic echo based on the number of taps, a dividing means for dividing a tap coefficient at a certain time into M (M ≧ 2) blocks, and the divided M blocks. An echo canceller, comprising: an extracting unit that extracts a part of the extracted block; and an updating unit that updates the extracted partial block to update the entire input tap coefficient. 4. An echo canceller for canceling an acoustic echo based on the number of taps, an impulse response measuring means for measuring an impulse response under a certain acoustic environment, and an echo measurement for measuring an echo duration from the measured impulse response. Means, determining means for determining the number of taps based on the measured echo duration, calculation means for calculating the tap coefficient based on the determined number of taps,
A dividing unit that divides the calculated tap coefficient into M (M ≧ 2) blocks, an extracting unit that extracts a part of the divided M blocks, and a part of the extracted block. An echo canceller comprising: an updating unit that updates the entire input tap coefficient. 5. An echo canceller for canceling an acoustic echo based on the number of taps, a dividing means for dividing a tap coefficient at a certain time into M (M ≧ 2) blocks, and the divided M blocks. Extracting means for extracting a part, updating means for updating the extracted part of the blocks to update the entire input tap coefficient,
Transmitting means for transmitting a far-end speaker voice input signal which is an analog signal, and generating a far-end speaker voice output based on the far-end speaker voice input signal connected to the transmitting means and transmitted by the transmitting means. And a far-end speaker voice input transmitted by the transmitting means when performing echo cancellation processing based on the far-end speaker voice input signal transmitted by the transmitting means and connected to the transmitting means. An echo canceller comprising: an echo cancellation processing means for converting a signal into a digital signal.