JPS62116025A - Echo canceler - Google Patents

Abstract

PURPOSE:To obtain an echo canceler which normally operates, by detecting the first rise time of a sampled value string of the impulse response of an echo line connecting a delay detecting section which determines a delaying time in accordance with said detected result with the output of an estimation circuit, and providing a delay circuit at the preceding stage of an artificial echo line (FIR filter). CONSTITUTION:A delayed sample number (d) (d=d'-1) is determined from the sample number of the first rise of the sampled value row of an impulse response by a delay detecting section 12 and the delayed sample number of a delaying circuit 3 is set to the (d). When the initialization is completed, receiving speech signals x(t) arrive after a call is started, and an objective signal v(t) to be given to a microphone 3 is interrupted, correction to the fluctuation of an echo line of the impulse response is performed by means of an estimation circuit 7 by regarding transmitting speech signals z(t) as echo signals y(t) and output z(t) of a subtractor 10 as an estimation error signal. IN certain circumstances, correction to the delayed sample number (d) is also performed by the delay detecting circuit 12 and the delaying time of the delaying circuit 13 and the filter coefficient of an artificial echo line (FIR filter) 8 are also corrected in accordance with the correction.

Description

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

(Prior Art) With the spread of audio conferences, it is desired to provide a loudspeaker communication device that can ensure simultaneous communication 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 a receiving terminal 1 receiving a receiving signal X(1) to a speaker 2, and a transmitting system extending from a microphone 3 to a transmitting output terminal 4. In a communication system consisting of two systems, training signal #5 is sent to the receiving system before the start of a call, and training signal X' is sent to the receiving system.
(t), and the echo signal y (t) input from the speaker 2 through the echo path to the microphone 3 is converted into a zero converter 6.
The sample value is converted into a sample value and sent to the estimating circuit 7 to estimate the in-gurus response h (t) of the echo path and generate a pseudo echo path (FIR filter) 8. After the call starts, the reception signal x (t) applied to the receiver terminal 1 passes through the echo path to the microphone 3, and the echo signal y (
t), and the transmitted signal (1) to which the target signal v(t) added to the echo path is added is generated. On the other hand, a pseudo echo signal z(1) obtained from the receiving system via the A/D converter 9 and a pseudo echo path (F'IR filter) 8 using the receiving signal x(1) as a transmitting signal, ( 1) to subtractor 10
By subtracting by , the echo signal y (t) is canceled and only the target signal v(1) is sent to the transmitter output terminal 4. The pseudo echo path (FIR filter) 8 needs to follow the change in the echo path over time, and the in-gurus response h (
t) is successively estimated by the estimation circuit 7 using the estimation error signal (ct) when the target signal v(1) is not added, and the pseudo echo path (FIR filter) 8 is corrected to always achieve optimal echo cancellation. is supported. Note that digital signal processing is used for the estimating circuit 7, the pseudo echo path 8, and the subtracter 10, and the transfer function estimated by the estimating circuit 7 is the sample value sequence h(iτ
) (i = 1.2..., (τ: sample time), and the pseudo echo path 8 is a finite sample value sequence h (jτ) (l = 11..., N), ( N:
It is composed of a convolution calculator using an FIR filter using the filter length) as a coefficient. Further, in the figure, h (t) indicates an X impulse response, and 11 indicates a D/A converter.

(Problems to be Solved by the Invention) However, in the conventional echo canceller described above, the filter length of the FIR filter is limited due to limitations in real-time processing. Therefore, if the echo path is a sound field and the transmission delay of the echo is long and the echo time is long, the in-group response h (t) used as the coefficient of the FIR filter
The finite number of sample values of is a truncated version of the true in-force response h(t), and the echo signal y
There was a problem that (t) could not be erased sufficiently.

When there are no restrictions on the placement of speaker 2 and microphone 3, if speaker 2 and microphone 3 are close to each other, sufficient impulses can be generated within N filter lengths as shown in Figure 3(a). It is possible to capture the response sample value sequence and eliminate the echo signal, but if speaker 2 and microphone 3 are far apart, or if the distance between speaker 2 and microphone 3 becomes long during use. In this case, as shown in Fig. 3(b), it becomes impossible to capture the necessary sequence of sample values of the innoculus response within the N filter lengths, and the echo signal cannot be sufficiently erased, resulting in unerased traces. There was also a problem.

Therefore, an object of the present invention is to provide a device which solves the problem of deterioration of cancellation characteristics when the propagation delay of the echo path becomes severe and allows the echo cancellation device to operate normally with respect to the freedom of speaker and microphone installation. be.

(Means for Solving All Problems) In order to achieve all of the above objects, the present invention detects the initial rise time of the sample value sequence of the impulse response in the echo path at the output of the estimation circuit, and delays the delay based on this. By connecting all the delay detectors that determine the time and immersing all the delay circuits in the front stage of the pseudo echo path (FIR filter), the received signal is delayed by the delay time, and the delay time is used as the coefficient of the FIR filter. By using a sample value sequence of an impulse response whose time has been advanced by 1, a finite number of sample value sequences close to the true innoculous response are secured as coefficients of all FIR filters.

(Function) As described above, in this invention, the number of delay samples is corrected by the delay detecting section, and the delay time of the delay circuit and the filter coefficient of the FIR filter in the pseudo echo path are also corrected accordingly. The cancellation characteristics of echo signals are improved compared to the conventional technique.

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, 3 is a microphone, 4 is a transmitting output terminal, 5 is a training signal, and 6 is a line. Converter, 7 is an estimation circuit, 8 is a pseudo echo path (FIR filter), 9 is a fitted converter, 10 is a subtracter, and 1 is a D/A converter, which are the same as the conventional ones. A delay detection unit 12 is connected to the output of the estimation circuit 7, which detects the initial rise time of the sample value sequence of the impulse response of the echo path and determines the delay time based on this. )8
A delay circuit 13 is connected in front of the FIR filter to delay the received signal by the delay time, advance the time by the delay time as a coefficient of the FIR filter, and use the sample value sequence of the impulse response to generate the true impulse. A finite number of sample value sequences close to the response are secured as coefficients of the FIR filter.

Next, its operation will be explained.

First, before starting a call, activate the training signal source 5,
The estimation circuit 7 calculates a sample value sequence π (i'' The size of the pseudo echo path (FIR filter) 8 is also set larger than N. Next, the delay detection section 1
2, determine the number of delayed samples d, where d = d'-1, from the number of samples d' at the first rising edge of the impulse response sample value sequence, and set the number of delayed samples of the delay circuit 7 to d.
Set to . In the pseudo echo path (FIR filter) 8, a finite sample value sequence h((i+d)τ)(i=1...
..., N)? : Set as filter coefficient.

With the above steps, the initial settings are completed, and after starting a call, the receiving signal χ(
1) is input, the target signal v(1) applied to the microphone 3
When the transmission signal Z (t) e echo signal y (regarded as the output of the subtractor 1 o Z (t) k is regarded as the estimation error signal, the estimation circuit 7 calculates the echo path of the impulse response. In some cases, the delay detection unit 12 also corrects the number of delay samples d, and changes the delay time of the delay circuit 13 and the pseudo echo path (FIR).
The filter coefficients of filter) 8 are also modified accordingly. Note that the number of delayed samples d was set to d'-1, but
It is good also as d≦d'-1.

Because of this effect, if the delay until the first rise of the impulse response is large in an echo path with a long impulse response length, the pseudo echo path (FIR filter) 8 is advanced by the delay circuit 13 and the number of delay samples d. sample value sequence h((i+d
) τ) (i=1,...,N) as a filter coefficient, it is possible to include a large number of effective components of the innogruous response within the limited filter length N. Therefore, the cancellation performance of the echo signal y(t) can be improved. Further, when the microphone 3 and the speaker 2f are freely arranged, even if they move during use, the delay time dt can be detected and corrected to the optimal pseudo-echo path. Therefore, the echo signal cancellation characteristics can be improved compared to the conventional technology.

(Effects of the Invention) As explained in detail above, since the delay detection section and the delay circuit are configured in the echo cancellation device mainly consisting of the estimation circuit and the FIR filter, the cancellation performance of the echo cancellation device is improved and the sound amplification device is improved. There is an advantage that the quality of telephone calls is improved. Further, in a public address communication device in which a microphone and a speaker can be freely arranged independently, the arrangement may change during use, and the delay in the echo path may change considerably, and this is particularly effective in such cases.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a block diagram showing a conventional echo canceling device, and FIG. 3 is a diagram showing a sample value sequence of an estimated impulse response.
(,) is () if the delay time until the first startup is short.
b) shows a large case. 1...Receiver end, 2...Speaker, 3...My/
. Phone, 4... Sending output end, 5... Training signal source, 6... Mechanical converter, 7... Estimating circuit, 8... Pseudo echo path (FIR filter), 9... Meme conversion Vessel, 10...
- Subtractor, 1... D/A converter, 12... Delay detection section, 13... Delay circuit. Block diagram according to the present invention FIG.

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. In an echo canceling device for a communication system consisting of a transmission system and a transmission system, an impulse response is estimated as a transfer characteristic of the echo path, a delay detection section detects a rise time of the impulse response to determine a delay time, and the pseudo By using a delay circuit provided before the echo path (FIR filter) and adjusting the delay time, the FI of the pseudo echo path can be adjusted.
An echo canceling device characterized by having a function of reducing the rise time of an impulse response used as a coefficient of an R filter.