JPH0262151A - Voice switch - Google Patents

Abstract

PURPOSE:To improve the naturalness of conversation and understandability by executing a prescribed operation, presuming the voice/amplitude ratio of a nearby speaker, and controlling an attenvator attenuating quantity based on a presumed value. CONSTITUTION:From the outputs of a reception amplitude calculating part 109 and a transmission amplitude calculating part 110, an exchange amplitude ratio calculating part 111 calculates the ratio between a reception input amplitude value and a transmission input amplitude value. Based on a calculation result and the respective attenuation quantity set values of a reception attenuator 102 and a transmission attenuator 107, a voice amplitude ratio presuming part 112 presumes the voice amplitude ratio of the amplitud values of a far edge speaker voice and a nearby speaker voice. From the presumed value, the new set attenuation quantity is calculated by an attenuation quantity control part 113, and the attenuation quantities of the attenuator 102 and 107 are controlled. Consequently, the attenuation quantity is steplessly controlled, the attenuation quantities of the two attenuators are equally set even when the voice amplitude values of the nearby speaker and the remote speaker are approximately the same like the case of a double talk when the nearby edge speaker and the remote edge speaker simultaneously speak, one voice is not extremely attenuated, and a voice switch which can improve the naturalness of the conversation and the understandability without the cut of a word top and a word end can be obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a voice switch for preventing howling used in loudspeaker telephones, teleconference systems, and the like.

2. Description of the Related Art In recent years, two-way communication devices such as loudspeaker telephones and teleconference systems have become widespread, and voice switches are used in these devices to prevent howling.

Hereinafter, a conventional audio switch will be explained with reference to the drawings.

FIG. 5 shows a block diagram of a conventional audio switch.

In FIG. 5, 501 is an input terminal into which a reception input signal is input from the line, 502 is a reception attenuator that attenuates the reception input signal and outputs it as a reception output signal to a speaker amplifier 503, 503 is a speaker amplifier, and 504 is a speaker. ,
505 is a microphone, 506 is a microphone amplifier, 507
is a transmitting attenuator that attenuates the transmitting input signal that is the output of the microamplifier 506 and outputs it as a transmitting output signal to the output terminal 508, 508 is an output terminal that outputs the transmitting output signal to the line, and 509 is a receiving input. A receiving amplitude calculation unit calculates a receiving input amplitude value which is an amplitude value of a signal, 501 is a transmitting amplitude calculation unit which calculates a transmitting input amplitude value which is an amplitude value of a transmitting input signal, and 511 is a unit that calculates a receiving input amplitude value and a transmitting input amplitude value. A transmitting/receiving determination unit compares the input amplitude values and determines that when the receiving input amplitude value is greater than or equal to the transmitting input amplitude value, it is in the receiving state, and when it is less than that, it is in the transmitting state. Attenuator 5
Attenuation amount control that sets the attenuation amount of 02 to be small and the attenuation amount of transmitting attenuator 507 to be large, and conversely, when in the sending state, the attenuation amount of receiving attenuator 502 is set to be small and the attenuation amount of transmitting attenuator 507 is set to be large. Department.

Regarding the conventional voice switch configured as described above,
The operation will be explained below.

6(a) to 6(i) show signal waveforms at various parts of the conventional example shown in FIG. 5, and the operation of FIG. 5 will be explained using these figures.

In FIG. 6, 601 is the waveform of the receiving input signal at point A in FIG. 5, 602 is the waveform of the receiving output signal at 8 points in FIG.
03 is the waveform of the transmitting input signal at point 0, 604 is the waveform of the transmitting output signal at point D, 605 is the receiving input amplitude value at point E, 60
6 indicates the transmitting input amplitude value at point F, 607 indicates the waveform of the transmitting/receiving determination signal at point G, 608 indicates the setting state of the attenuation amount of the receiving attenuator 502, and 609 indicates the setting state of the attenuation amount of the transmitting attenuator 507. It is something that

In a two-way call, the own terminal side is called the near end, the other party's terminal side is called the far end, the near end speaker is called the near end speaker, and the far end speaker is called the far end speaker. This term will be used in the following explanation.

The receiving input signal 601 in FIG. 6 is the far end speaker's voice signal.
The received output signal 6 passes through the received attenuator 502 in FIG.
It becomes 02. This received output signal 602 is amplified by a speaker amplifier 503 and amplified by a speaker 504 on the near end side. The microphone 505 picks up the echo from the speaker 504 and the voice of the near-end speaker speaking into the microphone 505. The output is amplified by a microphone amplifier 506 and becomes a transmitting input signal 603. This transmitting input signal 603 passes through a transmitting attenuator 507 to a transmitting output signal 604.
and is output from the output terminal 508 to the far end. The reception amplitude calculation unit 509 calculates the amplitude value of the reception input signal 601 to obtain the reception input amplitude value 605. Transmission input amplitude calculation unit 5
10, the amplitude value of the transmitting input signal 603 is calculated to obtain the transmitting input amplitude value 606. In the transmitting/receiving determination unit 511, when the receiving input amplitude value 605 is equal to or higher than the transmitting input amplitude value 606,
In the following cases, it is determined that the call is being transmitted, and a transmission/reception determination signal 607 is sent to the attenuation amount control section 512. When the transmission/reception determination signal 607 is in the receiving state, the attenuation amount control unit 512 controls the receiving attenuator 502.
The attenuation amount of the transmitting attenuator 5 is reduced when in the transmitting state.
07 attenuation amount is controlled to be small. The attenuation amounts of these two attenuators are controlled reciprocally so that when the attenuation amount of one is large, the attenuation amount of the other is always small. As a result, the receiving attenuator 502 has an attenuation amount shown at 608. , the transmitting speech attenuator 507 is controlled to the attenuation amount shown in 609.

With the above configuration, it is determined whether the near-end speaker or the far-end speaker is speaking louder, and the attenuation amount of the receiving system and the transmitting system is controlled reciprocally based on the determination result. The voice of the speaker who is speaking louder between the near-end speaker and the far-end speaker is prioritized and amplified, without increasing the loop gain, that is, without causing howling. It enables loudspeaker calls at high volume.

Problems to be Solved by the Invention However, with the above configuration, when a near-end speaker and a far-end speaker speak at the same time, the voice of the speaker with a lower voice rapidly attenuates, resulting in the occurrence of truncations at the beginning and end of words. The problem was that the sense of conversation became louder and the naturalness and intelligibility of the conversation deteriorated.

To solve this problem, instead of controlling the two attenuators by switching them to two levels of attenuation, large and small, the attenuation of the two attenuators is controlled steplessly according to the amplitude level of both voices. There is a need. In the conventional configuration, two states, transmitting and receiving, are determined, and the attenuation amount is switched to two levels, large and small, depending on the result of this determination, and such control is not possible.

In view of the above-mentioned problems, the present invention provides a voice switch that allows natural and intelligible calls even during double talk.

Means for Solving the Problems To achieve this objective, the audio switch of the invention comprises:
a receiving attenuator that attenuates a receiving input signal and outputs it as a receiving output signal; a transmitting attenuator that attenuates a transmitting input signal and outputting it as a transmitting output signal; and a receiving input that is an amplitude value of the receiving input signal. a receiving amplitude calculation unit that calculates an amplitude value; a sending amplitude calculation unit that calculates a sending input amplitude value that is the amplitude value of the sending input signal; and a ratio of the receiving input amplitude value to the sending input amplitude value. A transmission/reception amplitude ratio calculation unit that calculates the transmission/reception amplitude ratio, and from the set values of the attenuation amounts of the two attenuators and the value of the transmission/reception amplitude ratio, calculate the amplitude value of the far-end speaker's voice and the amplitude value of the near-end speaker's voice. and a voice amplitude ratio estimator that estimates the voice amplitude ratio which is the ratio of the two attenuators. It is composed of an attenuation amount control section that controls the amount of attenuation.

Effect: With this configuration, it is possible to estimate the voice amplitude ratio, which is the amplitude ratio of the near-end speaker's voice and the far-end speaker's voice, and by controlling the attenuation amounts of the two attenuators according to this estimated voice amplitude ratio, double When the near-end speaker's voice and the far-end speaker's voice have approximately equal amplitude values, such as during talk, when the audio amplitude ratio is close to 1, the attenuation amounts of the two attenuators can be set approximately equal. Therefore, the conventional problem of excessively attenuating the voice of the speaker with a lower voice during double talk can be overcome, and it is possible to provide a voice switch that reduces the feeling of disconnection at the beginning and end of words and provides excellent naturalness and intelligibility of conversation. .

EXAMPLE An example of the present invention will be described below with reference to the drawings.

In FIG. 1, reference numeral 101 is an input terminal into which a reception input signal is input from the line, 102 is a reception attenuator that attenuates the reception input signal and outputs it as a reception output signal to a speaker amplifier 103, 103 is a speaker amplifier, 104 is a speaker, 105 is a microphone, 106 is a microphone amplifier, 107 is a transmitting attenuator that attenuates the transmitting input signal that is the output of the microphone amplifier 106 and outputting the same as a transmitting output signal to the output terminal 108; An output terminal for outputting a speech output signal to the line; 109 is a reception amplitude calculation unit for calculating a reception input amplitude value, which is an amplitude value of the reception input signal; 11;
0 is a sending/receiving amplitude calculation unit that calculates a sending input amplitude value that is the amplitude value of the sending input signal, and 111 is a sending/receiving amplitude calculating unit that calculates a sending/receiving amplitude ratio that is a ratio of the receiving input amplitude value and the sending input amplitude value. A ratio calculation unit 112 is a voice amplitude ratio that estimates the amplitude value of the far-end speaker's voice and the amplitude value of the near-end speaker's voice from the set values of the attenuation amounts of the two attenuators and the value of the transmission/reception amplitude ratio. The estimation section 113 is an attenuation amount control section that calculates an attenuation amount to be newly set from the estimated audio amplitude ratio estimated by the audio amplitude ratio estimation section and controls the attenuation amount of the two attenuators.

The operation of the audio switch according to the embodiment configured as described above will be explained below.

Prior to explaining the operation of the embodiment shown in FIG. 1, the principle of estimating the audio amplitude ratio and the principle of calculating the amount of attenuation in this embodiment will be briefly explained using the analytical model of FIG.

In FIG. 2, the far-end speaker is speaking at an amplitude level of Vr, and the near-end speaker is speaking at an amplitude level of Vs. At this time, the receiving input signal is R.
-i, the receiving output signal is Ro, the transmitting input signal is Si, the transmitting output signal is So, and the gain of the receiving attenuator is G.
r, the gain of the transmit attenuator is set to Gs. In this model, the following relational expression holds.

Ri=Vr+β-3°5o=Gs-3t Transmission/reception amplitude ratio; L is determined by equation (2). This (2
) The normalized amplitude ratio: N shown in equation (3) can be obtained from equation (3). In this equation, the voice amplitude ratio (Vs/Vr), which is the amplitude ratio of the near-end speaker's voice and the far-end speaker's voice, is expressed as K, and the difference between the transmitting input amplitude value: Si and the receiving input amplitude value: R1 is expressed as K. The transmission/reception amplitude ratio: (St/R1) is represented by L, and the normalized amplitude ratio is represented by N.

FIG. 3 illustrates equation (3).

Curves 31 to 39 in FIG. 3 are obtained by plotting the N value at nine types of Gr values on the horizontal axis with α and β as constants. At this time, G' and Gs are controlled reciprocally as shown in (4) 2.

G r = Hm/G s ”(
4) In equation (4), )(m is a constant value that is sufficiently smaller than l.

A curve 30 in FIG. 3 shows the locus drawn by N when Gr and GS are controlled by the method according to the present invention, and FIG. 4 shows a method for setting Gr and Gs.

The value in formula (3) is the value determined by the amplitude ratio calculation section.

The audio amplitude ratio calculation section first divides this L by Gr and α to obtain N
Then, using Fig. 3, estimate K from Gr and GS (!:N).In the attenuation control section, estimate G from Fig. 4 and K.
Determine r and Gs.

Next, the operation of the voice switch of the present invention will be explained in more detail using FIGS. 3 and 4 as described above.

First, it is assumed that an N value of N1 is obtained when the Gr value is set to Grl and the Gs value is set to Gsl. That is, it is assumed that the state is at point A shown in FIG. 3 (Grl, Gsl, Kl, N1). At this time, in the audio amplitude ratio estimator,
The value of Kl is estimated from Grl, Gsl, N1, and FIG. 3. From FIG. 3, it can be seen that an estimated value Kl can be obtained in this state. Next, the attenuation amount control section is
The Gr value Gr2 and the Gs value (N2) are determined from the figure and Gr and Gs are updated. As a result, the curve 30 in Fig.
It can be seen that it is stable at point B indicated by (Gr2.GS2.Kl, N2) above. In FIG. 3, it is assumed that K1 is a state in which only the far end speaker is speaking, and it can be seen that in a state in which only the far end speaker is speaking, K1 is set to Gr)Gs.

Next, the near-end speaker interrupts, such that the K value is 2,
That is, it is assumed that there is a transmission interruption. At this time,
Since the Gr value is set to Gr2 and the Gs value is set to GsZ value, immediately after the interrupt there is a transition to 6 points (Gr2°Gs 2. to 2.N3), and an N value of N3 is obtained. The audio amplitude ratio estimating section consists of Gr2, GS2, N3, and FIG.
The attenuation amount control section determines the values Gr3 and GS3 from 2 and FIG. 4, and corrects Gr and Gs. As a result, the third cause (2°K4 to Gr3.GS3.)
It becomes stable at point D. It can be seen that K2 is assumed to be in a double talk state and is set to GrζGs.

Next, when moving to a state in which only the near-end speaker is speaking with a K value of 3, the state immediately after the transition is at point E where (Gr3°Gs 3. to 3.K5) is shown in FIG. The attenuation amount estimation section estimates 3 from the values of Gr3, GS3, and K5, and corrects Gr and Gs to the values of Gr4 and GS4. As a result, (Gr
4. GS4. It becomes stable at point F, which is 3°K6). Therefore, when only the near-end speaker is speaking, Gr(Gs
You can see that it is set to .

The above is the basic operation of the voice switch of the embodiment shown in FIG.

As described above, according to this embodiment, by providing the transmitting/receiving amplitude ratio calculating section, the audio amplitude ratio estimating section, and the attenuation amount controlling section, it is possible to estimate the audio amplitude ratio of the far-end speaker's voice and the near-end speaker's voice. , it becomes possible to set the optimal attenuation amount commensurate with this audio amplitude ratio, and when the audio amplitude ratio is close to 1, such as during double talk, the attenuation amount of the two attenuators is set to the same degree,
To completely overcome the conventional problem that the voice of a speaker with a lower voice is excessively attenuated during double talk, and to realize an excellent voice switch that does not give a feeling of disconnection at the beginning or end of a word. Furthermore, since switching between sending and receiving is performed smoothly, the naturalness of conversations can be dramatically improved.

Effects of the Invention The present invention provides a receiving attenuator that attenuates a receiving input signal and outputs it as a receiving output signal, a transmitting attenuator that attenuates a transmitting input signal and outputs it as a transmitting output signal, and a transmitting attenuator that attenuates a receiving input signal and outputs it as a receiving output signal. a receiving amplitude calculation unit that calculates a receiving input amplitude value that is an amplitude value of the transmitting input signal; a transmitting amplitude calculating unit that calculates a transmitting input amplitude value that is the amplitude value of the transmitting input signal; A transmission/reception amplitude ratio calculator calculates a transmission/reception amplitude ratio, which is a ratio of input amplitude values, and calculates an amplitude value close to the amplitude value of the far-end speaker's voice from the setting values of the attenuation amounts of the two attenuators and the value of the transmission/reception amplitude ratio. A voice amplitude ratio estimating unit that estimates a voice amplitude ratio that is a ratio to the amplitude value of the voice of an end talker; and calculating a new attenuation amount to be set from the value of the estimated voice amplitude ratio estimated by the voice amplitude ratio estimating unit. By providing an attenuation control unit that controls the attenuation of the two attenuators, it is possible to estimate the amplitude ratio of the far-end speaker's voice and the near-end speaker's voice, and to determine the optimal attenuation amount commensurate with this voice amplitude ratio. This allows for the realization of a voice switch that has excellent naturalness in conversation, with little sense of truncated beginnings and endings.

[Brief explanation of the drawing]

Fig. 1 is a block diagram of an audio switch in an embodiment of the present invention, Fig. 2 is a model diagram for explaining the operation, Fig. 3 is an illustration for explaining the operation for estimating the audio amplitude ratio, and Fig. 4 is an illustration for determining the amount of attenuation. FIG. 5 is a block diagram of a conventional audio switch, and FIG. 6 is a waveform diagram of signals in each part of FIG. 5. 102... Receiving voice attenuator, 107... Transmitting voice attenuator, 109... Receiving voice input amplitude calculation unit, 1
10... Sending input amplitude calculation unit, 111...
. . . Transmission/reception amplitude ratio calculation unit, 112 . . . Audio amplitude ratio estimation unit, 113 . . . Attenuation amount control unit. Agent's name Patent attorney Shigetaka Awano Haga 12, Figure 2, Figure 3, σ5

Claims (1)

[Claims] a receiving attenuator that attenuates a receiving input signal and outputs it as a receiving output signal; a transmitting attenuator that attenuates a transmitting input signal and outputting it as a transmitting output signal; and a receiving input that is an amplitude value of the receiving input signal. a receiving amplitude calculation unit that calculates an amplitude value; a sending amplitude calculation unit that calculates a sending input amplitude value that is the amplitude value of the sending input signal; and a ratio of the receiving input amplitude value to the sending input amplitude value. A transmission/reception amplitude ratio calculation unit that calculates the transmission/reception amplitude ratio, and from the set values of the attenuation amounts of the two attenuators and the value of the transmission/reception amplitude ratio, calculate the amplitude value of the far-end speaker's voice and the amplitude value of the near-end speaker's voice. and a voice amplitude ratio estimator that estimates the voice amplitude ratio which is the ratio of the two attenuators. An audio switch comprising: an attenuation amount control section that controls an attenuation amount.