JPS62120734A - Echo erasing equipment - Google Patents

Abstract

PURPOSE:To improve the erasing performance and the talking quality by providing a storage circuit storing a transfer function of plural echo paths and a means combining pseudo echo paths in interlocking with a gain of each microphone. CONSTITUTION:A training signal source 5 is operated at first prior to the start of talking, a gain alpha1 of a gain adjusting circuit 101 is set to '1', gains alpha2-alphan of gain adjusting circuits 102-10n are set to '0', an estimate circuit 6 obtains an estimated sample string of an impulse response of the echo path from a speaker 2 to a microphone 91 and the result is stored in a storage circuit 121. The gain alphaj of a gain adjusting circuit 10j varies in response to a level of an object signal Vj(t) entering each microphone 9j after the start of talking. The gain of the gain adjusting circuit 10j is interlocked equally with the gain of a gain circuit 13j. Thus, in giving an output of an adder 14 as a coefficient of an FIR filter of a pseudo echo path (FIR filter) 7, the pseudo echo path (FIR filter) 7 following momentarily the change in the gains alpha1-alphan of he microphone is obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an echo canceling device that eliminates echoes that cause howling and impair hearing, mainly in conference loudspeaker systems.

(Prior Art) With the spread of voice conferences, it is desired to provide a loudspeaker telephone device that can ensure simultaneous conversation without busy schedules and has less reverberation. There is an echo canceling device that satisfies this requirement.

FIG. 2 is a block diagram showing an example of a conventional echo canceling device, which includes a receiving system from the receiving terminal 1 that receives the receiving signal (receiving signal) to the speaker 2, and a transmitting system extending from the microphone 3 to the transmitting output terminal 4. In the communication system, the training signal source 5 sends a training signal to the receiving system before the start of the call! '(t
), and the echo signal y(t) which is input from the speaker 2 to the microphone 3 via the echo path is sent to the estimation circuit 6 to estimate the in-resonance response h(t) of the echo path and create a pseudo echo path. (FIR filter) 7 is generated. After the call starts, the receiving signal x(t) is applied to the receiving end 1
passes around the microphone 3K via the echo path and becomes the echo signal )'(t), and a transmission signal Z(t) is generated to which the target signal v(t) added to the echo path is added. On the other hand, when the reception signal "(t) is input, the pseudo echo path (FI) is
A pseudo echo signal 7 (t
) from the transmitted signal z(t) by the subtractor 8, the echo signal y(t) is canceled and the target signal v(t
) is sent to the transmitter output terminal 4. False echo path (FI
R filter) 7 needs to follow the change in the echo path over time, and the in i4 pulse response is the estimated error signal z(t) = y(t) − when the target signal v(t) is not added.
” (t) is successively estimated by the estimation circuit 6, and by correcting the pseudo echo path (FIR filter) 7, optimal echo cancellation is always maintained. ) 7 and the subtractor 8 use digital signal processing, and therefore require an appropriate D/A converter (not shown in the figure).The transfer function estimated by the estimation circuit 6 is based on the impulse response of The sample value sequence h(it) (i''1 p2..., τ: sample time), and the pseudo echo path (FIR filter) 7 is inno? Are there a number of response responses +7)? It is composed of a convolution calculator using an FIR filter using a sample value sequence SC17) (i=1 g...N, N: filter length) as a coefficient. Further, h(t) in the figure indicates a true inno-fourth response.

(Problem to be solved by the invention) However, in a conference call device, it is necessary to use multiple microphones depending on the number of speakers, but in a conventional echo canceling device, as shown in FIG. microphones 3, 3. , . . . , 3n outputs can only be used as they are added together. By the way, the advantage of using multiple microphones is that they shorten the distance between the microphones and the speaker present in the meeting, thereby securing the volume and suppressing quality deterioration due to reverberation, but all the microphones are in operation at all times. In this case, the quality deteriorates due to a decrease in SA and interference of sound waves.

Therefore, each output of the microphone is provided with a gain adjustment circuit so that the overall output level is equivalent to one microphone, and the gain of each microphone is adjusted according to the transmission level of each microphone. It is desirable to have a design that adjusts the sensitivity of microphones that do not have the power of the transmitter and lowers the sensitivity of the microphone. However, when such a design is applied to a conventional echo canceler, changes in the gain adjustment circuit are treated as fluctuations in the echo path, and correction by adaptive operation is required, making it difficult to follow frequent gain adjustments. was there.

Therefore, an object of the present invention is to solve the problem that, in an echo canceling device in which a plurality of microphones are used, the pseudo echo path cannot be corrected without adaptive operation in response to changes in the echo path caused by gain adjustment of each microphone. It is an object of the present invention to provide a device that can instantly respond to gain adjustment and allow an echo canceling device to operate normally.

(Means for Solving the Problems) The present invention includes a storage circuit that stores a transfer function of an echo path for each of a plurality of microphones, and adjusts the gain of a gain adjustment circuit added to the output of each microphone. The apparatus has means for synthesizing a pseudo echo path by changing the gain of the output of the storage circuit accordingly, and the transfer function of each storage circuit is adapted to be adaptively corrected according to fluctuations in the echo path.

(Function) As described above, this invention is provided with a memory circuit for the transfer functions of a plurality of echo paths, so that changes in the gain of the gain adjustment circuit can be instantaneously dealt with, distinguishing them from fluctuations in the echo path, thus eliminating echoes. The erasure performance of the device is improved, thereby improving call quality. Therefore, the above-mentioned problem can be eliminated.

(Embodiment) FIG. 1 is a block diagram showing an embodiment of the present invention, in which 1 is a receiver terminal, 2 is a speaker, 4 is a transmitter output terminal, 5 is a training signal source, 6 is an estimation circuit, and 7 is a The pseudo-echo path 8 is a subtracter as before. 9□ to 9n are n microphones, 1-01 to 10n are the plurality of microphones 9□ to 9
11 is an adder for these gains, 121 to 12n are storage circuits for storing the transfer functions of the echo paths, and 131 to 13n are interlocked with the gain adjustment circuits 91 to 9n. gain circuit, 14
indicates an adder for these gains.

Next, clouds will explain how it works.

First, before starting a call, the training signal source 5 is activated, the gain α of the gain adjustment circuit 10□ is set to 1, and the gains α2 to α of the gain adjustment circuits 10□ to 10n are set to 1.

is set to 0, and the estimation circuit 6 generates a sample value sequence h (if) (1m1 e 2 ) of the estimated value of the impulse response of the echo path from the speaker 2 to the microphone 9□.

...N) is determined and stored in the memory circuit 12□. Similarly, the gain of the gain adjustment circuit 101 is set to 1 and the others are set to 0, and the estimated value h(ir) (i=1.2
．． ...N) and stores it in the memory circuit 121, j=1,
Estimated impulse responses of all echo paths 2...n (i=1, 2...N) are stored in respective storage circuits.

After the call starts, the input adjustment circuit 10j is adjusted according to the level of the target signal vj(t) input to each microphone 9j.
O1'in αj varies. At this time, gain adjustment circuit 1
The gain of 0.0 and the gain of the gain circuit 13j are equally linked. That is, although not shown in the figure, they are connected in parallel, for example. Therefore, the output of the adder 14 is Σαj
πj(if) (i=1, 2,...N), and by giving this as the coefficient of the FIR filter of the m1 pseudo echo path (FIR filter) 7, the microphone gain α,
A pseudo echo path (FIR filter) 7 that instantaneously follows changes in ~α0 is generated.

Furthermore, fluctuations in the echo path after the start of the call are detected by the estimation circuit 6 and stored in the memory circuit 12j (j=1...n
) is successively corrected.

Detection of this echo path variation is possible with all microphones 9.
This is done using a reverberation signal when the target signal vj(t) is not added to j (J = 1 p...n), but in this case, one of the microphones (9j) has a large gain, and the others The estimated value hj (i) is corrected by maintaining the state so that the gain of By maintaining the gain of the microphone 9j that has been receiving a high target signal until another microphone receives a new target signal, each echo path is corrected according to the frequency of microphone operation.

Therefore, it is possible to maintain optimal echo signal cancellation characteristics compared to the prior art.

(Effects of the Invention) As explained in detail above, in an echo canceling device using multiple individual microphones, a memory circuit that stores transfer functions of multiple echo paths and a gain of each microphone are used. Since a means for synthesizing a pseudo echo path is provided, the cancellation performance of the echo cancellation device is improved, and the speech quality in the loudspeaker communication device is improved. Therefore, it can be used for audio conference calls where multiple microphones are used.
Its effects are fully demonstrated.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention, and FIG. 2 is a block diagram showing a conventional echo canceling device when the echo path is a sound field. 1... Receiving human power end, 2... Speaker, 3... Fist/microphone, 4... Sending output end, 5... Training signal source, 6... Estimating circuit, 2... Pseudo echo Road (FI
R filter), 8...Subtractor, 9□~9n...Microphone, 10□~10n...Gain adjustment circuit, 11
°°°Adder, 121~12n...Memory circuit, 13□
~13n...gain circuit, 14...adder.

Claims (1)

[Scope of Claims] A pseudo echo path is generated by estimating the transfer characteristics of the echo path from the transmitted signal to the echo path and the echo signal after the transmitted signal passes through the echo path, and the transmitted signal is converted into the pseudo echo path. The receiver generates an estimated echo signal obtained by inputting the echo path, and subtracts the estimated echo signal from the echo signal to eliminate the echo signal and extract a target signal added to the echo path. An echo canceling device for a communication system consisting of a transmission system and a transmission system, in the case where there are a plurality of echo paths due to the use of a plurality of microphones, the apparatus comprises a storage circuit for storing transfer functions for each of the plurality of echo paths, means for synthesizing a pseudo echo path from each of the transfer functions according to the gain of a gain adjustment circuit connected to a subsequent stage of each of the plurality of microphones, and the transfer function is configured to adjust the variation of the echo path corresponding to each of the microphones. An echo canceling device characterized by sequentially rewriting the contents of a memory circuit.