JPH10308816A - Voice switch for speaking equipment - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent frequent switching operation during a speech and to enable natural speech without interrupting the voice of a called party by discriminating the sound/silence of voice state based on the continuos comparison data. SOLUTION: A comparator part 51 of sound detection part 5 compares an inputted power value pi with a prescribed threshold value th and outputs the compared result ci . Based on the compared result ci at this comparator part 51, a continuos state discrimination part 52 discriminates a sound state si . The continuos state discrimination part 52 is provided with a sound state continuation counter Cnt1 and a silence state continuation counter Cnt0. When the value of silent state continuation counter Cnt0 exceeds a prescribed threshold value th0, silence is discriminated, but when the value of sound state continuation counter Cnt1 exceeds a prescribed threshold value th1, sound is discriminated, and when the value of silence state continuation counter Cnt0 is smaller than the threshold value th0 and the value of sounding state continuation counter Cnt1 is smaller than the prescribed threshold value th1, a sound state si-1 of the last discriminated result is maintained as it is.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice switch used for a hands-free telephone or the like. In a hands-free telephone using a voice switch system, it is necessary to perform continuation control of conversation when detecting a sound to perform switch switching control, thereby preventing frequent switch switching.

[0002]

2. Description of the Related Art To realize a hands-free function,
It is necessary to increase the volume of the speaker and the sensitivity of the microphone. However, in this case, as shown in FIG. 4, an acoustic echo occurs in which the received voice output from the voice output unit such as a speaker circulates to a voice input unit such as a microphone. This is a phenomenon in which one's voice sounds like an echo to the other party, and is very difficult to use. There are two methods for removing the acoustic echo: (1) an echo canceller method and (2) a voice switch method.

[0003] The echo canceller system removes an acoustic echo using an adaptive signal processing technique. For example, FIG.
As shown in (1), an acoustic echo r in which the output received voice wraps around the microphone is artificially generated inside the telephone, and is subtracted from the signal input to the microphone. The generation of the pseudo echo r 'is obtained by expressing a transfer function from the speaker to the microphone by using an FIR filter. Since this transfer function changes depending on the situation around the telephone, the filter is adaptively changed so that the error between the pseudo echo r 'and the acoustic echo r is minimized.

[0006] On the other hand, in the voice switch system, as shown in FIG. 6, the power of a speaker output voice is compared with the power of a microphone input voice, and one of them is suppressed to remove an acoustic echo. That is, while the signal is being output from the speaker, the signal input to the microphone is an acoustic echo. During this time, the microphone input signal is suppressed to prevent transmission of the acoustic echo to the other party.

As described above, there are two methods of removing an acoustic echo which is a problem in realizing the hands-free function, an echo canceller and a voice switch. A comparison of the advantages and disadvantages of both is shown in FIG. 7, which is a trade-off between the processing amount and the performance. To prioritize the cost, a voice switch method will be adopted. The present invention relates to this voice switch.

FIG. 8 shows a detailed conventional structure of a hands-free telephone having the voice switch. In FIG. 8, 1 is a receiving unit including a demodulator or the like that receives an audio signal from the other party, 2 is a power suppressing unit that can suppress and control the power of the received signal by changing the receiving gain gain-r,
Reference numeral 3 denotes an audio output unit which includes an amplifier, a speaker, and the like, and emits a received voice (R). Reference numeral 6 denotes a voice input unit which includes a microphone and an amplifier and inputs a transmission voice, and 7 denotes a transmission gain g.
A power suppression unit 8 that can suppress and control the power of the received signal by changing the ain-s, and a transmission unit 8 including a modulator that transmits the transmission voice signal to the other party.

Reference numeral 4 denotes a power calculator for calculating the power of the received signal received by the receiver 1. Reference numeral 5 'denotes whether the current received voice state sr is silence or sound based on the power calculated by the power calculator 4. , A power calculation unit 11 for calculating the power of the audio signal input to the audio input unit 6, 9 ′
Is a sound detection unit that detects whether the current transmission voice state s-s is silence or sound based on the power calculated by the power calculation unit 11, and 10 is a detection result of the sound detection units 5 'and 9'. The determination unit determines which side of the power suppression units 2 and 7 is to be in the suppression control state based on the power suppression units.

Here, the power calculators 4 and 11 calculate the power of the input voice data according to the following formula. That is, assuming that the input audio data is x _i , the output power p _i is obtained by p _i = 10 × log [Σ (x _ij x _ij )]. Here, Σ is an addition from j = 0 to J.

As shown in FIG. 9, the sound detectors 5 'and 9' comprise a comparator for comparing the input power p _i with a predetermined threshold value th. _i
The compared with the threshold th, current audio state S _i is determined whether voiced or silence. Here, s _i = 0 is silence, s _i
i = 1 means a sound. The judgment formula is if (p _i <th) s _i = 0 if (p _i > th) s _i = 1. This is because the input power p _i is smaller than the threshold th, the speech state s _i to "0", if greater More threshold th, in which the speech state s _i to "1". As a result, it is possible to prevent erroneous determination of existence of background noise having a threshold value th or less.

[0010] The determination unit 10 receives the reception gain gai of the reception power suppression unit 2 according to the determination logic table shown in FIG.
nr and the transmission gain gain-s of the transmission power suppressing unit 7 are controlled. Here, the reception gain gain-r and the transmission gain gain-s are in the range of 0.0 ≦ gain ≦ 1.0. In the determination logic table of FIG. 10, when the transmission voice state s−s = 0 and the reception voice state s−r = 0, both the transmission gain gain-s and the reception gain gain−r are “0.0”. I do. When the transmitted voice state s−s = 1 and the received voice state s−r = 0, the transmitted gain “gain-s” is set to “1.0” and the received gain “gain-r” is set to “0.0”. When the transmitted voice state s−s = 0 and the received voice state sr = 1, the transmitted gain “gain-s” is set to “0.0” and the received gain “gain-r” is set to “1.0”. When the transmission voice state s-s = 1 and the reception voice state sr = 1, the reception is prioritized, and the transmission gain gain-s is set to "0.
0 ", and the reception gain gain-r is set to" 1.0 ". Control.

According to the determination result of the determination unit 10, the power
-The suppression units 2 and 7 are input audio data x _iFor
And output audio data x_iOutput as x_i= X_i× gain

As described above, this voice switch system suppresses one of the received voice and the transmitted voice depending on the state of the voice.
If the other is the reception voice, the speaker output is performed, and if the transmission voice is the transmission voice, the transmission is performed. When both of them are voices, various criteria can be considered, such as a case where a received voice is prioritized, and a case where a higher voice power is prioritized.

[0013]

The voice switch operates at a high speed in order to improve the simultaneous callability as much as possible. As a side effect, switching may occur in response to an instantaneous silent part during a call. There is. This means that the conversation partner hears the voice of the other party intermittently because the switch is frequently switched, and the conversation becomes very difficult to hear, and the user feels discomfort.

The present invention has been made in view of the above problems, and has as its object to prevent frequent switching of a switch during a call and to enable a natural call without interrupting the voice of the other party.

[0015]

In order to solve the above-mentioned problems, a voice switch of a telephone according to the present invention comprises: a receiver-side power calculator for calculating a power of a received voice; and a power calculator of a transmitted voice. Transmitting-side power calculating means, and receiving-side voice detecting means for determining the voiced / non-voiced voice state of the received voice from the power of the received voice of the receiving power calculating means;
A transmitting-side voice detecting means for determining a voiced / unvoiced voice state of the transmitting voice from the power of the transmitting voice of the transmitting-side power calculating means; a receiving-side suppressing means for suppressing the power of the receiving voice; A transmitting-side suppressing means for suppressing the power of the transmitting voice; a receiving state based on a voice state of the receiving voice of the receiving side voice detecting means and a voice state of the transmitting voice of the transmitting side voice detecting means. Determining means for determining which one of the received voice of the side-side suppressing means and the transmitted voice of the transmitting-side suppressing means is to be suppressed, wherein at least one of the voice receiving detecting means on the receiving side and the transmitting side is provided with an input. A continuation state determining means for determining whether or not the sound / silence comparison result obtained by comparing the power of the voice with a predetermined threshold value is the same for a predetermined time; Voice / silence based on the result Configured to perform the judgment. According to the present invention, in order to prevent frequent switch switching during a conversation, a continuation state determination unit that determines how much sound and no sound continue based on a simple comparison result between a threshold and input voice power. Is provided. This is to determine whether or not the comparison result continues with no sound (or sound). If the same result continues for a predetermined time or more, it is finally determined to be no sound (or sound).

The sound detection means having the continuation state determination means determines that the sound state is silent when the silent comparison result continues for a first predetermined time or more, and determines that the sound comparison result is a second predetermined time or more. If the sound state is determined to be sound if continued for a while, the duration of the silent comparison result is smaller than the first predetermined time and the duration of the sound comparison result is shorter than the second predetermined time. It can be configured to maintain the previous voice state determination result.

Further, in the above-mentioned voice switch, it is possible to further comprise a duration changing means for changing a predetermined time used for the continuation state judgment by the continuation state judgment means to an arbitrary value. With this duration changing means, the user can freely set a duration suitable for the environment in which the telephone is used, so that a natural telephone conversation can be performed.

In the above-mentioned voice switch, when the voice detection unit on the receiving side or the transmission side, whichever has priority, determines that the voice state is voice, the power calculation of the voice detection unit on the other side and the voice calculation are performed. At least one of the sound detection processes can be stopped. Thereby, unnecessary calculation processing can be reduced.

[0019]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 shows a hands-free telephone having a voice switch according to an embodiment of the present invention. In the figure, a receiving unit 1, power suppressing units 2 and 7, an audio output unit 3, power calculating units 4 and 11, an audio input unit 6, a transmitting unit 8, and a determining unit 10 are circuit elements described in the conventional device of FIG. Therefore, detailed description is omitted here.

On the other hand, the sound detection units 5 and 9 are different from those of the conventional circuit. That is, the sound detection unit 5 includes the comparison unit 51
The sound detection unit 9 includes a comparison unit 91 and a continuation state determination unit 92. Since the sound detection units 5 and 9 have the same configuration, only the function and operation of the sound detection unit 5 will be described below.

The comparing section 51 is a circuit for comparing the input power signal p _i with a predetermined threshold value th, thereby outputting a comparison result c _i by the following judgment formula.
Here, c _i = 0 means no sound, and c _i = 1 means sound.
Judgment formula _{is, if (p i <th)} c i = 0 if (p i> th) is c _i = 1. This means that if the input power p _i is smaller than the threshold th, the comparison result c _{i is set} to “0”, and if the input power p _i is larger than the threshold th, the comparison result c _{i is set} to “1”. Incidentally, the comparison result c _i corresponds to the voice state s _i in the conventional apparatus.

Based on the comparison result c _i of the comparison section 51,
Continuation state determination unit 52 determines the speech state s _i. Continuation state determination unit 52 includes a talk-spurt state continuation counter Cnt1 and silent state continuation counter Cnt0 to the comparison result c _i are input. This sound state continuation counter Cn
t1 and the silent state continuation counter Cnt0 operate according to the following equation. Input comparison result c _i = 0 → Cnt0 = Cnt0 + 1,
Cnt1 = 0 input comparison result _{c i = 1 → Cnt1 = Cnt1} + 1,
Cnt0 = 0 That is, the sounding state continuation counter Cnt1 is equal to the comparison result c.
_{When i} = 1 is input, the count value is incremented by one,
When the comparison result c _i = 0 is input, the count value is reset to zero. The silent state continuation counter Cnt0 indicates that the comparison result c _i =
When 0 is input, the count value is incremented by one, and when the comparison result c _i = 1 is input, the count value is reset to 0.

The voice state s _i is determined based on the values of the silent state continuation counter Cnt0 and the sound state continuation counter Cnt1 as follows: voice state s _i = 0 if (Cnt0> th0) voice state s _i = 1 if (Cnt1> th1) Voice state s _i = s _i-1 if (Cnt0 <th0 and Cnt1 <th1). Here, th0 is the threshold for determining that a voice state s _i is made of sound in silence, th1 is the threshold value for determining that the voice state s _i is made from a silence In the case of this embodiment, "2"
Is set to

The above-mentioned determination equation is based on the following equation.
If it exceeds 0, the voice state s _i = 0, that is, no sound, and the sound state continuation counter Cnt1 is set to a predetermined threshold th.
If the value exceeds 1, the voice state s _i = 1, that is, sound is set. If the silent state continuation counter Cnt0 is smaller than the threshold th0 and the sound state continuation counter Cnt1 is smaller than the threshold th1, the previous state is determined. That is, the voice state s _i-1 as the determination result is maintained as it is.

The operation example of determination of the speech state s _i according to this embodiment apparatus is shown in FIG. Comparison result c _i of the comparator 51 in the sound detecting unit is, for example, that shown in FIG. 2 (a). In this example, the time t at which the comparison result c _i = 0 continues
From time 1 to time t9, a spontaneous sound detection (c _i =
1) is performed, and the comparison result c _i = 1 continues at time t10.
At 24, single silence detection (c _i = 0) is performed at times t15 and t19, respectively, and the comparison result c _i = 0 continues after time t25. In the conventional device, the time t
The switch is switched even for one-off changes at 6, t15, and t19, so that conversation is difficult to hear.

FIG. 2B shows a silent state continuation counter Cn.
The count value of t0 is shown. Silence state continuation counter Cnt
0 until the time t4 continuously incremented by counting consecutively each comparison result c _i = 0 (count value in the example of FIG CNT0 =
Increased to 8), the time reset has been counted value by the comparison result c _i = 1 at t6 CNT0 = 0, and the continued again in count of each comparison result c _i = 0, after the time t10 the comparison result c _i = 1 count CNT0 continue to be reset will continue to "0", time t15, which is only incremented by count CNT0 = 1 when sporadic comparison c _i = 0 at time t19. Then, after time t25, the comparison result c _i = 0 is continuously input again, so that the count value Cnt0 is sequentially incremented.

On the other hand, FIG. 2C shows the count value of the sounding state continuation counter Cnt1. The voiced state continuation counter Cnt1 is continuously reset by each comparison result c _i = 1 until time t9, and the count value is “0”.
Is incremented only when the one-time comparison result c _i = 0 in step (1), and the count value Cnt1 = 1. After time t10, each comparison result c _i = 1 is sequentially counted and incremented, but is reset only when the comparison result c _i = 0 at times t15 and t19, and the count value Cnt
1 = 0, and the counting is restarted from 0. Time t2
From 5 onward, the count value Cnt1 = 0 is continued since the reset is continued by the comparison result c _i = 0.

The continuation state determination unit 52 based on the count value of the silent state continuation counter Cnt0 and talk-spurt state continuation counter Cnt1 makes a determination of the speech state s _i as shown in Figure 2 (d). That is, until the time t4, since the count value of the silent state continuation counter Cnt0 is larger than the threshold value th0 = 2, it is determined that the sound state s _i = 0 (that is, no sound). At time t5, the count value of the silence state continuation counter Cnt0 is reset to “0”. However, since the count value of the sound state continuation counter Cnt1 is also smaller than the threshold th1 = 2, the voice state s _i is determined according to the above-described determination formula. The previous voice state s _i-1 = 0. At time t9, the silent state continuation counter Cnt0 = 3, so that the audio state s _i = 0 according to the determination formula. On the other hand, from time t10, the comparison result c
_i = 1, whereby the voiced state continuation counter Cnt
1 is incremented, and the count value Cnt1 is incremented.
Is There threshold th1 (= 2) between a smaller, speech state s _i according to the determination expression continues taking the last speech state s _i-1 = 0, the time the talk-spurt state continuation counter Cnt1 = 3 t
For the first time at 12, the voice state s _i = 1. Time t15, also the comparison result c _i in time t19 is changed momentarily "0", to keep the previous speech state s _i-1 = 1 according to the same manner as above speech state s _i determination expression. Is reset at time t25 to a sound state continuation counter Cnt1 = 0, and is silent state continuation counter Cnt0 begins to be incremented, the previous speech state s _i-1 = 1 according to the voice state s _i is at this time still discriminants At the time t27 when the silent state continuation counter Cnt0 = 3 is maintained, the voice state s _i = 0 is determined according to the determination formula.

As described above, in the present embodiment, by providing the continuity determination unit in the sound detection unit, it is possible to prevent a switch from being switched in a short sound part in a silent part or a short sound part in a sound part. Realize a natural call.

In the above-described embodiment, the thresholds th0 and th1 for determining the duration of the sound / non-speech determination are fixed, but the thresholds th0 and th1 can be set arbitrarily. If the means is provided, the predetermined duration used for the determination of the continuation state can be arbitrarily changed, whereby the user can freely set the duration suitable for the environment in which the telephone is used, thereby providing a natural and comfortable environment. Make calls possible.

FIG. 3 shows another embodiment of the present invention.
Here, the sound detection units 5 and 9 are the same as those described in the above embodiment. The voice state sr is notified from the voice detector 5 on the receiving side to the voice detector 9 and the power calculator 11 on the transmitting side. Accordingly, when it is determined that the voice state sr of the received voice is voiced, the power calculation unit 11 does not calculate the power of the transmitted voice, and the voice detection unit 9 outputs the voice status of the transmitted voice. No determination is made and it is uniquely determined that there is no sound. As a result, the amount of calculation can be reduced, and the load of software processing can be reduced when these processes are performed by software using a microprocessor or DSP.
In this example, it is assumed that the criterion at the time of the simultaneous call is the reception priority. Therefore, when the determination criterion for simultaneous communication is transmission priority, the voice detection unit 9 on the transmission side notifies the voice detection unit 5 on the reception side, etc., and the voice state s of the transmission voice
When it is determined that -s is a sound, the operations of the power calculator 4 and the sound detector 5 on the receiving side may be stopped.

[0032]

As described above, according to the present invention,
The switch can be prevented from being switched by a short voice in a silent section or a short silent in a voice section, and the other party's voice can be prevented from being interrupted, so that a natural call can be made.

[Brief description of the drawings]

FIG. 1 is a diagram showing a hands-free telephone having a voice switch according to an embodiment of the present invention.

FIG. 2 is a diagram illustrating an operation example of the embodiment apparatus.

FIG. 3 is a diagram showing another embodiment of the present invention.

FIG. 4 is a diagram illustrating acoustic echo in a hands-free communication device or the like.

FIG. 5 is a diagram illustrating an echo canceller method.

FIG. 6 is a diagram illustrating an audio switch system.

FIG. 7 is a diagram comparing an echo canceller system and a voice switch system.

FIG. 8 is a diagram showing a conventional hands-free telephone having a voice switch.

FIG. 9 is a diagram showing a configuration of a sound detection unit in a conventional device.

FIG. 10 is a diagram illustrating an example of a sound / silence determination table.

[Explanation of Symbols] 1 receiving unit 2, 7 power suppressing unit 3 sound output unit 4, 11 power calculating unit 5, 5 ', 9, 9' sound existence detecting unit 6 sound input unit 8 transmitting unit 51, 91 comparing unit 52 , 92 continuation state determination unit

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Hitoshi Matsuzawa 4-1-1, Kamiodanaka, Nakahara-ku, Kawasaki-shi, Kanagawa Prefecture Inside Fujitsu Limited (72) Inventor Hiroshi Katayama 4-1-1, Kamiodanaka, Nakahara-ku, Kawasaki-shi, Kanagawa Prefecture No. 1 Inside Fujitsu Limited (72) Inventor Tomonari Kimura 3-22-8 Hakata Ekimae, Hakata-ku, Fukuoka City, Fukuoka Prefecture Inside Fujitsu Kyushu Digital Technology Co., Ltd.

Claims (4)

[Claims] A receiving power calculator for calculating a power of the received voice; a transmitting power calculator for calculating a power of the transmitted voice; and a power of the received voice of the receiving power calculator. A receiving-side voice detecting means for determining a voiced / non-voiced voice state; and a transmitting-side voice detecting means for determining a voiced / non-voiced voice state of the transmitted voice from the power of the transmitted voice of the transmitting side power calculating means. Sound detecting means, a receiving-side suppressing means for suppressing the power of the received voice, a transmitting-side suppressing means for suppressing the power of the transmitted voice, a voice state of the received voice of the receiving-side voice detecting means, and Judgment means for determining which of the received voice of the receiving side suppressing means and the transmitted voice of the transmitting side suppressing means is to be suppressed, based on the voice state of the transmitted voice of the transmitting side voice detection means, Sound detection of the receiving side and the transmitting side At least one of the outputting means includes a continuation state determining means for determining whether or not the sound / silence comparison result obtained by comparing the power of the input voice with a predetermined threshold value is the same for a predetermined time. Voiced sound based on the ongoing comparison result
A voice switch for a telephone set configured to determine silence. 2. A sound detecting means having the continuation state judging means judges that the sound state is silence if the comparison result of silence continues for more than a first predetermined time, and the sound comparison result is determined by a second predetermined time. If the sound state continues for more than the time, it is determined that the voice state is sound, and the duration of the silent comparison result is smaller than the first predetermined time, and the duration of the sound comparison result is longer than the second predetermined time. 2. The system according to claim 1, wherein when the sound level is small, the previous voice state determination result is maintained.
Voice switch of the described handset. 3. The voice switch according to claim 1, further comprising a duration changing means for changing a predetermined time used for the continuous state determination by the continuous state determination means to an arbitrary value. 4. When the sound detector on the receiving side or the transmitting side, whichever has priority, determines that the voice state is sound, at least the power calculation of the sound detector on the other side and the sound detection processing are performed. 4. The voice switch for a telephone according to claim 1, wherein one of said voice switches is stopped.