JP3754241B2 - Telephone device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a speech device such as a telephone provided with a speech speed conversion device and an echo canceller.
[0002]
[Prior art]
Conventionally, for example, a telephone disclosed in Japanese Patent Application Laid-Open No. 6-311211 is known as a telephone having a speech speed conversion function. In the telephone disclosed in this publication, the “slow listening” mode is selected by operating the mode key, and during the mode selection, the time axis of the voice signal from the other party is extended by the speech speed conversion processing circuit. It is structured so that you can listen to the words with a slow voice.
[0003]
In this way, speaking speed conversion means that the time axis of the audio signal is compressed and its playback speed (speaking speed) is made faster than the original speed, or conversely, the time axis of the audio signal is extended and its playback speed is increased. To say that (speaking speed) is slower than the original speed.
[0004]
However, the conventional telephone has a problem in that it is difficult to talk because the side sound whose speech speed has been converted can be heard at the receiver on the transmitter side.
[0005]
In addition, in the case of a hands-free telephone, a video conference system, etc., an echo problem occurs in addition to the side sound. That is, the voice on the transmitting side (self) emitted from the speaker on the receiving side (the other party) returns to the transmitting side (self) as an echo through the microphone on the receiving side (the other party).
[0006]
In order to prevent such a problem, an echo canceller is conventionally known. By installing the echo canceller, not only the echo as described above but also the side sound can be removed. In the echo canceller, a pseudo echo is created by estimating the echo path of the speech path, and the echo and side sound as described above are removed by subtracting it from the echo signal sent to the speech path. The estimation of the echo path is performed by optimizing the filter coefficient of the adaptive filter that generates the pseudo echo signal by learning.
[0007]
Therefore, if a conventionally known echo canceller is installed in a telephone having the above-mentioned speech speed conversion function, side sounds and echoes are removed. It can basically solve the problem that it is difficult to hear from the side.
[0008]
[Problems to be solved by the invention]
FIG. 3 shows an example of the configuration of a telephone having a speech speed conversion function equipped with an echo canceller. In the figure, 1 is a microphone provided in the handset, and 2 is a speaker provided in the handset. Reference numeral 5 denotes a speech speed conversion unit that performs speech speed conversion processing on the received signal, and reference numeral 6 denotes an echo canceller unit 5 that removes side sounds and echoes. Reference numeral 3 denotes a two-wire / four-wire conversion circuit. A telephone line is connected via the two-wire / four-wire conversion circuit 3, and the voice signal of the transmitter input from the microphone 1 is converted into the two-wire / four-wire It is output to the telephone line via the circuit 3.
[0009]
Note that the microphone and speaker are plus / minus, two wires each, and a total of four wires, and the telephone line is also two wires, but only signal lines are shown in the figure, so the microphone and speaker are each One line at a time and one telephone line.
[0010]
In the above configuration, since the impedance mismatch between the telephone line and the telephone itself occurs, the transmitted signal is reflected by the 2-wire / 4-wire conversion circuit 3 and becomes a side sound signal. It is removed by the canceller unit 6. Also, the incoming signal received via the telephone line is converted into a slow voice signal by the speech speed conversion circuit 4 and emitted from the speaker 2 of the speaker's receiver.
[0011]
Furthermore, in the case of hands-free telephones, video conference systems, etc., as described above, in addition to the side sound, the voice on the sending side (self) emitted from the speaker on the receiving side (the other party) It returns as an echo to the transmitting side (self) through the (other party) microphone, but this is also removed by the echo canceller 6.
[0012]
However, since the echo canceller needs to learn the echo path and determine the filter coefficient of the adaptive filter every time a call is made, side echoes and echoes can be sufficiently removed during the learning period of the echo canceller immediately after the start of the call. Therefore, there was a problem that side sounds and echoes that could not be removed could be heard after the speech speed was converted.
[0013]
[Means for Solving the Problems]
In order to solve the above problems, in the present invention, an echo canceller means for learning an echo path, optimizing a filter coefficient of an adaptive filter, removing an echo signal and outputting a reception signal, and a reception output from the echo canceller means The speech speed conversion means of the speech speed conversion means is stopped during a predetermined period in which the filter coefficient of the adaptive filter is optimized in a speech device having a speech speed conversion means for converting the signal speed of speech. It is characterized by that.
[0014]
In addition, the present invention receives a transmission signal as a reference input signal, generates a pseudo echo signal by optimizing the filter coefficient of the adaptive filter based on the reference signal, and removes an incoming echo signal by the pseudo echo signal in communication apparatus comprising an echo canceller means, a speech speed converting means for outputting the speech speed converting a received signal input through the echo canceller means for outputting the received signal and the filter coefficients of the adaptive filter During the predetermined period to be optimized, the speech speed conversion process of the speech speed conversion means is stopped.
[0015]
The predetermined period is an echo path learning period in an echo canceller.
[0017]
The present invention also provides an echo canceller that learns an echo path, optimizes a filter coefficient of an adaptive filter, removes an echo signal, and outputs a received signal, and a speech speed conversion of the received signal output by the echo canceller The speech speed conversion means of the speech speed conversion means stops while the echo signal that cannot be removed is output at a predetermined level or higher from the echo canceller means. It is characterized by that.
[0018]
The present invention also provides an echo canceller that receives a transmission signal as a reference input signal, generates a pseudo echo signal based on the reference signal, removes an incoming echo signal by the pseudo echo signal, and outputs a reception signal. And a speech speed converting means for converting the speech signal input through the echo canceller means and outputting the speech speed, the echo signal that cannot be removed by the echo canceller means exceeds a predetermined level. While being output, the speech speed conversion process of the speech speed conversion means is stopped.
[0019]
The present invention also provides a reference input signal buffer means for inputting a transmission signal as a reference input signal, an adaptive filter, a pseudo echo signal based on a signal from the reference input signal buffer means and a signal from the adaptive filter. The product sum calculating means to be generated, the coefficient updating means for updating the filter coefficient of the adaptive filter, and the voice detecting means for detecting the signal level of the echo signal (removed error signal) that cannot be removed, Echo canceller means for outputting a signal; and speech speed conversion means for converting the speech speed of the received signal output from the echo canceller means to output the speech detection means, wherein the speech detection means has the removal error signal equal to or higher than a predetermined signal level. During the interval, a signal for prohibiting speech speed conversion is output to the speech speed converting means .
[0021]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, a call device provided with the speech speed conversion device of the present invention will be described in detail with reference to the drawings.
[0022]
First, FIG. 1 is a block diagram showing a configuration of a telephone provided with a speech speed converting apparatus of the present invention. In the figure, 1 is a microphone provided in the handset, and 2 is a speaker provided in the handset. Reference numeral 4 denotes a microcomputer (control unit) that controls the entire telephone. The microcomputer 4 includes a speech speed conversion unit 5 and an echo canceller unit 5 serving as a side-tone removing unit. Note that a DSP (digital signal processor) may be used instead of the microcomputer. Reference numeral 3 denotes a 2-line / 4-line conversion circuit as input / output means, and a telephone line is connected through the 2-line / 4-line conversion circuit 3. Further, 7 and 10 are A / D converters, and 8 and 9 are D / A converters.
[0023]
The speech speed conversion unit 5 has a time axis extension unit (not shown) for extending the time axis of the voice signal of the other party arriving via the telephone line, and the time axis extension used here. As a method, for example, a known method such as a pointer addition control (Pointer Interval Control Overlap and Add: PICOLA) or a TDHS (Time Domain Harmonic Scaling) method can be used. The present invention is not limited to the above, and it may be anything as long as the time axis of the audio signal can be extended and the reproduction speed can be changed slowly. For example, the voice of the other party may be temporarily stored in a semiconductor memory, read out at a desired speed, and the voice signal converted from the voice speed may be output.
[0024]
Next, FIG. 2 is a schematic block diagram showing the configuration of the echo canceller unit 6. In the figure, reference numeral 61 is a reference input signal buffer for inputting a speech signal (speech signal) of a speaker input from the microphone 1 as a reference input signal x, 62 is an adaptive filter, and 63 is a reference from the reference input signal buffer. A sum-of-products operation unit that generates a pseudo echo signal y ′ by multiplying the input signal signal and the reference input signal through the adaptive filter 62 to generate a pseudo echo signal y ′, and 64 optimizes the filter coefficient of the adaptive filter 62 by learning and echoes And a coefficient updating unit for better reducing side sounds.
[0025]
In the above configuration, the reference input signal x stored in the reference input signal buffer 61 is supplied to the coefficient updating unit 64 and the product-sum operation unit 63 at a predetermined timing. In the product-sum operation unit 63, the reference input signal buffer 61 supplies the coefficient update unit 64 to the reference input signal via the adaptive filter 62 and the reference input signal supplied from the reference input signal buffer 61. Based on this, a pseudo echo signal y ′ is generated.
[0026]
Then, an echo signal y (the side sound and echo, hereinafter referred to as the echo signal y without being distinguished) coming through the A / D converter 10 is generated by the product-sum operation unit. Although most of the echo signal is removed by the sum of the pseudo echo signal y ′ (≈−y), the echo signal that cannot be removed is output as a removal error signal e.
[0027]
The removal error signal e is guided to the coefficient update unit 64 and the subsequent speech speed conversion unit 5, and the coefficient update unit 64 learns the filter coefficient of the adaptive filter 62 so that the removal error signal e is minimized. By updating to the optimum value, the echo signal can be better removed.
[0028]
During the learning period in the coefficient update unit 64, the filter coefficient of the adaptive filter 62 is not optimized, and an echo signal (removal error signal e) that cannot be removed is converted into a speech speed by the speech speed conversion unit 5 in the subsequent stage. Furthermore, there is a problem that sound is emitted from the speaker 8 via the D / A converter 8 and heard.
[0029]
Therefore, in the present invention, as shown in FIG. 2, a voice detector 65 for detecting the level of the removal error signal e is provided, and the voice detector 65 detects a removal error signal e of a predetermined level or higher. During the interval, the prohibition signal p is output to the speech speed conversion unit 5. Then, the speech speed conversion unit 5 stops the speech speed conversion process while receiving the prohibition signal p, and outputs a speech signal not subjected to the speech speed conversion process to the D / A converter 8 at the subsequent stage. It is configured.
[0030]
The voice detector 65 may detect the signal level of the removal error signal e by calculating the background noise power using an adaptive threshold, for example.
[0031]
Further, the echo canceller 6 is not limited to completely removing the echo signal, and can reduce the level of the echo signal to such an extent that the user's call is not hindered (so that it is hardly audible during a call). If it is, it can be considered substantially removed. Accordingly, the echo signal removing means in this specification is not limited to means for completely removing the echo signal, but includes means for reducing the echo signal.
[0032]
Furthermore, the telephone line is not limited to a wired line, but may be a wireless (radio) line such as a mobile phone.
[0033]
Next, the operation of the telephone according to the present embodiment will be described with reference to FIGS.
[0034]
First, when the user lifts the handset to start a call and enters an off-hook state, the echo canceller unit 6 outputs a prohibition signal p to the speech rate conversion unit 5. As shown in FIG. 2, the prohibition signal p is output from the voice detector 65 in the echo canceller unit 6, and is initially sent to the speech speed conversion unit 5 regardless of the presence or absence of the removal error signal e. In response to this, the prohibition signal p is output. The speech rate conversion unit 5 stops the speech rate conversion process while receiving the prohibition signal p, and the received signal or the like input via the echo canceller unit 6 is used as a speech signal not subjected to the speech rate conversion process. To the D / A converter 8.
[0035]
When the telephone is connected to the other party by the dial operation and the conversation is started, the echo canceller unit 6 optimizes the filter coefficient of the adaptive filter 62 by learning the echo path. As shown in FIG. 2, the echo canceller unit 6 uses the transmission signal from the microphone 1 as a reference input signal x and generates a pseudo echo signal y ′ based on the signal, but immediately after the start of the call, the adaptive filter Since the filter coefficient 62 is not optimized, a removal error signal e of a predetermined level or higher is output. Therefore, the coefficient updating unit 64 learns the echo path and optimizes the filter coefficient of the adaptive filter 62 so that the removal error signal e is minimized.
[0036]
Eventually, when the optimization of the filter coefficient of the adaptive filter 62 is completed, the echo canceller unit 6 can sufficiently remove side sounds and echoes, and the removal error signal e falls below a predetermined level. The voice detector 65 detects this and stops sending the prohibition signal p to the speech speed converter 5, so the speech speed converter 5 starts the speech speed conversion process and goes to the D / A converter 8 at the subsequent stage. The voice signal that has been subjected to speech speed conversion processing is output. As a result, the received signal is subjected to speech speed conversion processing at a predetermined speed, and sound is emitted from the speaker 2.
[0037]
【The invention's effect】
As described above in detail, according to the present invention, while the echo removal means is learning the echo path so as to optimize the filter coefficient of the adaptive filter, the operation of the speech speed conversion means is stopped, and the speech signal is subjected to the speech speed conversion process. Therefore, it is possible to prevent side sounds and echoes that cannot be removed during learning by the echo removing means from being converted by the speech speed converting means and output.
[Brief description of the drawings]
FIG. 1 is a schematic block diagram showing a configuration of a telephone according to the present invention.
FIG. 2 is a schematic block diagram showing a main configuration of a telephone according to the present invention.
FIG. 3 is a schematic block diagram showing a configuration of a conventional telephone.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 Microphone 2 Speaker 3 2-line / 3-line conversion circuit 4 Microcomputer 5 Speech rate conversion part 6 Echo canceller part 7 A / D converter 8 D / A converter 9 D / A converter 10 A / D converter 61 Reference input Signal buffer 62 Adaptive filter 63 Product-sum operation unit 64 Coefficient update unit 65 Voice detector

Claims (6)

An echo canceller that learns an echo path, optimizes a filter coefficient of an adaptive filter, removes an echo signal, and outputs a received signal; and a speech speed converting unit that converts the received signal output by the echo canceller In a communication device equipped with
A speech device characterized in that the speech speed conversion processing of the speech speed conversion means is stopped during a predetermined period in which the filter coefficient of the adaptive filter is optimized. A transmission signal is input as a reference input signal, a filter coefficient of an adaptive filter is optimized based on the reference signal, a pseudo echo signal is generated, an incoming echo signal is removed by the pseudo echo signal, and a reception signal is output. A speech apparatus comprising: echo canceller means for performing speech speed conversion means for converting the speech speed of a received signal input via the echo canceller means and outputting the speech speed;
A speech apparatus characterized in that the speech speed conversion processing of the speech speed conversion means is not performed during a predetermined period in which the filter coefficient of the adaptive filter is optimized.  3. The communication apparatus according to claim 1, wherein the predetermined period is an echo path learning period in an echo canceller. An echo canceller that learns an echo path, optimizes the filter coefficient of the adaptive filter, removes the echo signal, and outputs a received signal; and a speech speed converting unit that converts the received signal output by the echo canceller In a communication device equipped with
A speech device characterized in that the speech speed conversion processing of the speech speed conversion means is not performed while an echo signal that cannot be removed is output from the echo canceller means at a predetermined level or higher. Echo canceller means for receiving a transmission signal as a reference input signal, generating a pseudo echo signal based on the reference signal, removing an echo signal arriving by the pseudo echo signal and outputting a reception signal; and the echo canceller In a speech device comprising speech speed conversion means for converting the speech speed of the received signal input via the means and outputting the speech signal,
A speech device characterized in that the speech speed conversion processing of the speech speed conversion means is not performed while an echo signal that cannot be removed is output from the echo canceller means at a predetermined level or higher. Reference input signal buffer means for inputting a transmission signal as a reference input signal, an adaptive filter, and a product-sum operation means for generating a pseudo echo signal based on the signal from the reference input signal buffer means and the signal from the adaptive filter And an echo canceller for outputting a received signal, and a coefficient updating means for updating a filter coefficient of the adaptive filter and a voice detecting means for detecting a signal level of an echo signal (removal error signal) that cannot be removed. Means and speech speed converting means for converting the speech signal output by the echo canceller means and outputting the speech speed,
The voice detecting means outputs a signal for prohibiting speech speed conversion to the speech speed converting means while the removal error signal is equal to or higher than a predetermined signal level.

Images