JP2008032933A - Voice communication apparatus and voice communication system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a voice communication apparatus and a voice communication system, capable of converting a speaking speed to a suitable speaking speed without depending on speaker's individuality (speaking speed), without the need of advanced processing such as voice recognition. <P>SOLUTION: Each voice conference device 111 collects voice of a speaker and detects the speaking speed from a collected voice signal. Each voice conference device 111 transmits information of the detected speaking speed to a network server 101. The network server 101 calculates the average value of the information of the speaking speed received from each device, and returns it to each device. On the basis of the information of the speaking speed detected by own device and the average value of the information of the speaking speed received from the network server 101, each voice conference device converts the speaking speed so that the collected voice signal become an average speaking speed. The voice signal converted to the average speaking speed is transmitted to other devices. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an audio communication system that performs a remote conference by communicating audio signals with each other via a network, and an audio communication apparatus used in the audio communication system.

  When an audio conference is performed by connecting a plurality of points via a network, the conversation speed varies depending on each speaker, and it may be difficult to hear a specific speaker. Therefore, it is conceivable to convert the speech speed to make it easy to grasp the content of the speech, but in this case, the problem is how much the speech speed should be converted.

  In order to solve this problem, there has been proposed a speech speed conversion device that performs speech recognition processing on an input speech signal and calculates speech speed from comparison of time length with speech data of a standard pattern (see, for example, Patent Document 1).

In addition, there has been proposed a telephone having a speech speed conversion function in which a speech speed conversion is performed at a speech speed preset for each outgoing telephone number by using a number display service in the telephone (for example, Patent Documents). 2).
JP 2001-134299 A JP 2001-268175 A

  However, in the apparatus of Patent Document 1, it is necessary to perform voice recognition processing, and the speech speed cannot be calculated unless a voice that matches the standard pattern is input.

  In addition, the apparatus of Patent Document 2 has a problem in that it cannot cope with a change in the destination speaker because the speaking speed is set for each outgoing telephone number.

  The present invention relates to a voice communication apparatus and a voice communication system capable of converting a speech speed to an appropriate speech speed without depending on a speaker's individuality (speech speed) without performing sophisticated processing such as speech recognition. The purpose is to provide.

  The voice communication device according to the present invention includes a microphone that picks up the voice of a speaker and outputs a voice signal, a speech speed converter that converts a voice speed of the voice signal output from the microphone, and the speech speed converter The target speech speed is obtained from the speech signal output unit that outputs the speech signal subjected to the speed conversion to another device, the speech speed detection unit that detects the speech speed of the speaker from the speech signal output from the microphone, and the server device. A speech speed information communication unit that receives target speech speed information, which is information to indicate, a speech speed of a speaker detected by the speech speed detection unit, and target speech speed information received from the server device by the speech speed information communication unit. And a speech speed conversion amount determination unit that determines a speech speed conversion amount of the speech speed conversion unit based on the indicated speech speed.

  In the present invention, the speech speed is detected from the collected voice signal. A speech speed conversion amount is determined based on the speech speed detected by the own device and the speech speed received from another device, and the collected voice signal is converted and output. In other words, if the speech speed detected by the own device is slow compared to the speech speed received from the other device, the speech signal is compressed and transmitted, and if the speech speed detected by the own device is fast, the speech signal is expanded. Send from.

  Further, in the voice communication device of the present invention, the speech speed information communication unit transmits information indicating the speech speed of the speaker detected by the speech speed detection unit to the server device as speech speed information of the own device. Features.

  In the present invention, information indicating the detected speech speed is transmitted to and received from the server device. Thereby, in the server apparatus, the received speech speed information can be used for calculation of the target speech speed information.

  Further, in the voice communication device according to the present invention, the speech speed detection unit cuts out the consonant voice section and the vowel voice section from the voice signal output from the microphone, and sets the time length between the consonant voice section and the vowel voice section. Based on the ratio, the speaking speed of the speaker is detected.

  In the present invention, a consonant section and a vowel section are cut out from the speech signal. For example, the periodicity of the audio signal may be detected, and among the audio signals having a predetermined level or higher, an audio segment with a high periodicity may be set as a vowel interval, and other intervals may be set as a consonant interval. The speech speed is detected from the ratio between the time lengths of the consonant section and the vowel section. The speech speed may be determined to be slow (slow) when the time length of the vowel section with respect to the consonant section is long, and the speech speed may be determined to be fast (fast mouth) when the time length of the vowel section with respect to the consonant section is short.

  The voice communication system of the present invention is a voice communication system in which a server device and a plurality of voice communication devices are connected to each other, and each voice communication device picks up the voice of a speaker and outputs a voice signal. A speech speed conversion unit that converts the speech speed output from the microphone, a speech signal output unit that outputs the speech signal converted by the speech speed conversion unit to another device, and the microphone outputs A speech speed detection unit for detecting a speech speed of a speaker from a speech signal; and a speech speed information communication unit for transmitting information indicating the speech speed of the speaker detected by the speech speed detection unit to the server device as speech speed information; The server device includes target speech speed determining means for determining a target speech speed based on the speech speed information transmitted from each voice communication device and distributing the target speech rate to each voice communication device. The speech speed information communication unit of the communication device further supports The speech communication device receives the target speech speed amount from the communication device, and the voice communication device, based on the speech speed of the speaker detected by the speech speed detection unit, and the speech speed included in the target speech speed amount, A speech speed conversion amount determining unit for determining a speech speed conversion amount of the speech speed converting unit is further provided.

  In the present invention, the speech speed is detected from the collected voice signal. Information indicating the detected speech speed is transmitted to and received from the server device. The server device calculates a target speech speed amount (for example, an average value) from the speech speed information received from each voice communication device, and distributes the target speech speed amount to each voice communication device. In each voice communication device, the speech speed conversion amount is determined based on the speech speed detected by the own device and the target speech speed received from the server device, and the collected voice signal is converted and output. . In other words, if the speech speed detected by the device is slow compared to the target speech speed (average value), the speech signal is compressed and transmitted, and if the speech speed detected by the device is fast, the speech signal is expanded. Then send.

  Further, in the speech communication system of the present invention, the speech speed detection unit cuts out the consonant speech section and the vowel speech section from the speech signal output from the microphone, and sets the time length between the consonant speech section and the vowel speech section. Based on the ratio, the speaking speed of the speaker is detected.

  According to the present invention, by detecting the speech speed from the collected speech signal and receiving the target speech speed from the server device, it is not necessary to perform advanced processing such as speech recognition, so that the individuality of the speaker (speech speed) It is possible to convert the speech speed to an appropriate speech speed without depending on.

In the following embodiments, an audio conference system will be described as an example of a specific audio communication system with reference to the drawings. FIG. 1 is a configuration diagram of an audio conference system according to the present embodiment.
The audio conference system of this embodiment includes audio conference apparatuses 111 </ b> A to 111 </ b> C connected to a network 100 and a network server 101.

  The audio conference apparatuses 111 </ b> A to 111 </ b> C are respectively arranged at points a to c that are separated from each other. The audio conference device 111A is arranged at the point a, the audio conference device 111B is arranged at the point b, and the audio conference device 111C is arranged at the point c.

  FIG. 2 is a block diagram showing a configuration of the audio conference apparatus 111A. Since the audio conference apparatus 1B and the audio conference apparatus 1C have the same configuration and function as the audio conference apparatus 111A, the audio conference apparatus 111A will be described as a representative in FIG. The audio conference apparatus 111A includes a microphone 11, a sound collecting amplifier (AMP) 12, an A / D converter 13, a voice analyzing unit 14, a speech speed converting unit 15, a communication I / F 16, a D / A converter 17, a sound emitting amplifier (AMP). ) 18 and a speaker 19.

  The microphone 11 collects ambient sounds including voices from speakers existing around the device, and provides a collected signal to the sound collecting amplifier 12. The sound collection amplifier 12 amplifies the sound collection signal and supplies it to the A / D converter 13. The A / D converter 13 digitally converts the collected sound signal in the analog format and outputs it to the voice analysis unit 14 and the speech speed conversion unit 15.

  The speech speed conversion unit 15 performs speech speed conversion processing on the collected sound signal input from the A / D converter 13 as necessary. The necessity of the speech speed conversion process and the speech speed conversion amount are instructed by the speech speed conversion rate calculation unit 144 of the voice analysis unit 14. The speech speed conversion unit 15 outputs the speech signal after the speech speed conversion to the communication I / F 16.

  The communication I / F 16 converts the audio signal from the speech speed conversion unit 15 into a network communication format and transmits it to another device via the network 100. The communication I / F 16 receives audio data from other devices via the network 100. The communication I / F 16 converts the received audio data from network format data into a general audio signal, and outputs it to the D / A converter 17. The D / A converter 17 converts the input audio signal into an analog signal and supplies it to the sound output amplifier 18. The sound output amplifier 18 amplifies the analog sound signal and supplies it to the speaker 19. The speaker 19 emits the input audio signal.

  Further, the communication I / F 16 acquires the talk speed information (details will be described later) of the speaker from the time comparison unit 143 of the voice analysis unit 14, and transmits this to the network server 101. In addition, the communication I / F 16 receives the average speech speed information of the speakers of each device from the network server 101, and gives this to the time comparison unit 143.

  The speech analysis unit 14 includes a speech segment extraction unit 141, a consonant / vowel segment detection unit 142, a time comparison unit 143, and a speech speed conversion rate calculation unit 144, and the speech speed of the speech signal input from the A / D converter 13. Is detected.

  The voice segment cutout unit 141 cuts out a voice segment from the voice signal input from the A / D converter 13. The distinction between the voice section and the noise section may be determined from the level of the voice signal and the periodicity of the voice signal, for example. An audio signal having a predetermined level or higher is divided into predetermined lengths, and a correlation value is calculated by multiplying or subtracting corresponding sample data. When this correlation value is lower than a predetermined threshold, it is determined as a noise interval, and when it is higher, it is determined as a speech interval. In the case of an audio signal having a high periodicity such as speech, the correlation value is high, and in the case of an audio signal having a low periodicity such as noise, the correlation value is low.

  The consonant / vowel section detection unit 142 further detects a consonant section and a vowel section from the speech sections cut out by the voice section cutout unit 141. That is, the speech segment cut out by the speech segment extraction unit 141 is further divided into predetermined lengths, and the correlation value (correlation value for consonant / vowel segment determination) is calculated by multiplying or subtracting corresponding sample data. When this correlation value is lower than a predetermined threshold (consonant / vowel section determination threshold), it is determined as a consonant section, and when it is higher, it is determined as a vowel section. Further, the consonant / vowel section may be determined from the spectrum of the voice section cut out by the voice section cutout unit 141. A vowel section may be determined when the spectrum has a high degree of coincidence with a standard vowel spectrum.

  The time comparison unit 143 calculates the ratio of the time lengths of the consonant section and the vowel section detected by the consonant / vowel section detection unit 142. The time comparison unit 143 outputs the ratio of the time lengths of the consonant sections and the vowel sections to the communication I / F 16 and the speech speed conversion rate calculation unit 144 as speech speed information. FIG. 3 is a diagram showing the relationship between the time length of the consonant section and the vowel section and the speech speed. In the case of FIG. 5A, the time length of the consonant section is 10 ms and the time length of the vowel section is 150 ms, so consonant section: vowel section = 1: 15. In FIG. 5B, the time length of the consonant section is 10 ms, and the time length of the vowel section is 100 ms, so consonant section: vowel section = 1: 10. In FIG. 6C, the time length of the consonant section is 10 ms and the time length of the vowel section is 50 ms, so consonant section: vowel section = 1: 5. The time comparison unit 143 outputs information of this consonant section: vowel section as speech speed information. When the time length of the vowel section is long with respect to the time length of the consonant section as shown in FIG. 9A, the speech speed is slow (slow), and the time of the vowel section with respect to the time length of the consonant section as shown in FIG. If the length is long, the speech speed is fast (quick).

  The speech speed information output to the communication I / F 16 is transmitted to the network server 101 via the network 100. The network server 101 aggregates the speech speed information transmitted from each device, calculates the average speech speed information in the network, and returns this to each device.

  FIG. 4 is a block diagram showing the main configuration of the network server 101 of this embodiment. As illustrated in FIG. 4, the network server 101 includes a network control unit 102 and a conference information storage unit 103.

  The network control unit 102 controls the entire network 100. Further, the speech speed information transmitted from each device is recorded in the conference information storage unit 103 as a speech speed information history. The conference information storage unit 103 stores the number of conference participants currently participating in the conference, and records the average value of speech speed information in the speech speed information average value DB. In addition, voices transmitted and received between voice conference apparatuses are stored in the voice DB as minutes. The network control unit 102 transmits the average value of the speech speed information stored in the conference information storage unit 103 to each audio conference device. The average speech speed information is calculated by moving average.

The time comparison unit 143 of each voice communication device receives the average speech speed information given from the network server 101 and gives this to the speech speed conversion rate calculation unit 144.
The speech speed conversion rate calculation unit 144 receives the average speech speed information and the speech speed information of the own device from the time comparison unit 143, and calculates the speech speed conversion rate based on these values. For example, information indicating consonant interval: vowel interval = 1: 10 as shown in FIG. 3 (B) is received as the average speech speed information, and as shown in FIG. When information indicating consonant section: vowel section = 1: 15 is received, it is determined that the speech speed of the speaker of the device is slower than the average, and the speech speed conversion unit 15 is set to compress the speech signal. To do. On the other hand, when information indicating consonant section: vowel section = 1: 5 is received as the speech speed information of the own device as shown in FIG. 3C, the speech speed of the speaker of the own device is faster than the average. It is determined that the voice signal is present, and the speech speed conversion unit 15 is set to expand the voice signal.

  The speech speed conversion unit 15 performs speech speed conversion processing based on the speech speed conversion amount (compression rate, expansion rate) set by the speech speed conversion rate calculation unit 144 as described above. The speech speed conversion processing is performed as follows, instead of simply outputting the audio signal at high speed and low speed. That is, when expanding the audio signal, the audio signal is cut into a waveform of one cycle, a new periodic waveform is generated by synthesizing one section before and after each periodic waveform, and the newly synthesized periodic waveform between each periodic waveform Is inserted to increase the number of periodic waveforms of the signal, and the signal is time-extended while maintaining the pitch. When compressing an audio signal, the new periodic waveform is replaced with each periodic waveform, and the number of periodic waveforms of the signal is reduced and output.

  FIG. 5A is a flowchart showing the procedure of decompression processing. FIG. 2B is a diagram for explaining the expansion method. In FIG. 9A, first, the number of samples in one cycle (sampling frequency × 1 / signal frequency) of the head portion of the input audio signal is detected (S91). Two periodic waveforms, which are sample data for one period, are taken out and, as shown in FIG. 5B, an attenuation wave is created by multiplying the first periodic waveform A by an attenuation gain coefficient, An increasing wave is created by multiplying the second periodic waveform B by an increasing gain coefficient (S92). Then, by adding and synthesizing these, a periodic waveform having an intermediate shape between A and B is synthesized (S93). As shown in FIG. 6A, the synthesized waveform is inserted between the periodic waveform A and the periodic waveform B and output (S94), thereby performing acoustically natural time base expansion.

  In the case of compressing audio data, as shown in FIG. 6B, by synthesizing a synthesized waveform having an intermediate shape between A and B synthesized in S93, instead of the periodic waveforms A and B, Audio data can be compressed 1/2 times in the time axis direction.

  By defining the cycle for performing the speech speed conversion process, the conversion speed can be made variable. For example, as shown in FIG. 6C, by synthesizing two periodic waveforms for each period and inserting them between each periodic waveform, the audio data can be expanded twice in the time axis direction, As shown in FIG. 4D, by synthesizing two periodic waveforms every two periods, it can be expanded to 3/2 times.

  The voice signal subjected to the speech speed conversion by the speech speed conversion unit 15 as described above is input to the communication I / F 16 and output to other devices via the network 100. As a result, each voice conference apparatus receives the voice signal converted to the average speech speed on the sound collection side, and can listen to the voices of all the conference participants at the average speech speed.

  In the above example, the voice signal after the speech speed conversion is transmitted on the sound collection side, and the sound signal is emitted as it is on the sound emission side. However, the following application examples are also conceivable. .

  FIG. 7 is a block diagram illustrating a configuration of an audio conference apparatus according to an application example. In this audio conference apparatus, components common to the audio conference apparatus shown in FIG. 2 are denoted by the same reference numerals, and description thereof is omitted. This voice conference apparatus includes a speech speed conversion unit 50 connected to the output side of the communication I / F 16. A voice signal input from another device via the network 100 and the communication I / F 16 is input to the speech speed conversion unit 50 and input to the D / A converter 17 after the speech speed conversion. The necessity of speech speed conversion of the speech speed conversion unit 50 and the amount of speech speed conversion are set by the speech speed conversion rate calculation unit 144.

  In this example, the speech speed conversion rate calculation unit 144 calculates the speech speed conversion rate of the speech speed conversion unit 50 based on the speech speed information of the own device and the average speech speed information received from the network server 101. That is, the speech speed conversion rate calculation unit 144 converts the speech speed of the speech speed conversion unit 50 so that the speech speed of the audio signal received from the other device is close (substantially the same) as the speech speed of the speaker of the own device. Calculate the rate. For example, information indicating consonant interval: vowel interval = 1: 10 as shown in FIG. 3 (B) is received as the average speech speed information, and as shown in FIG. When information indicating consonant section: vowel section = 1: 15 is received, the speech speed conversion unit 50 is set to extend the voice signal from another device. On the other hand, when information indicating consonant section: vowel section = 1: 5 is received as the speech speed information of the own device as shown in FIG. 3C, the speech speed conversion unit 50 is set to compress the speech signal. To do.

  In other words, in a voice conference device in which a person who speaks quickly is present, the voice from the other device is emitted at a fast speech speed, and in a voice conference device in which a person who speaks slowly exists, the other device at a slow speech speed. The sound from is emitted. Thereby, each voice conference participant can listen to other uttered voices at a voice speed close to (similar to) his / her own voice speed, so that the voice conference can be performed without stress.

  In the present embodiment, an example has been described in which the network server 101 receives speech speed information from each audio conference device and determines the average speech speed information. However, for example, one-to-one using two audio conference devices. When conducting a teleconference, any voice conference device may function as a server device. For example, the time comparison unit 143 of one of the audio conference apparatuses transmits / receives the speech speed information in each of the audio conference apparatuses, and calculates the average speech speed information. The speech speed conversion rate calculation unit 144 is configured to determine the speech speed conversion amount based on the speech speed information of the own device and the average speech speed received from another device (or the time comparison unit 143 of the own device). That's fine.

The block diagram which shows the structure of the audio | voice communication system of embodiment of this invention. Block diagram showing the configuration of the main part of the voice communication device The figure which shows the relation between the time length of the consonant section and the vowel section and the speech speed Block diagram showing the configuration of the network server Diagram showing speech speed conversion processing The figure which shows the speech speed conversion processing when changing the expansion rate The block diagram which shows the structure of the principal part of the voice communication apparatus which concerns on an application example

Explanation of symbols

100-network 111-voice conference apparatus 101-network server

Claims (5)

A microphone that picks up the voice of the speaker and outputs a voice signal;
A speech speed conversion unit for converting the speech speed of the audio signal output from the microphone;
A voice signal output unit that outputs the voice signal converted by the voice speed conversion unit to another device;
A speech speed detection unit for detecting a speech speed of a speaker from an audio signal output from the microphone;
A speech rate information communication unit that receives target speech rate information that is information indicating the target speech rate from the server device;
Based on the speaking speed of the speaker detected by the speaking speed detection unit and the speaking speed indicated by the target speaking speed information received from the server device by the speaking speed information communication unit, the speaking rate conversion of the speaking rate conversion unit A speech rate conversion amount determination unit for determining the amount;
A voice communication device comprising:   The speech communication apparatus according to claim 1, wherein the speech speed information communication unit transmits information indicating the speech speed of the speaker detected by the speech speed detection unit to the server device as speech speed information of the own device.   The speech speed detection unit cuts out a consonant speech segment and a vowel speech segment from a speech signal output from the microphone, and based on a ratio of time lengths of the consonant speech segment and the vowel speech segment, The voice communication device according to claim 1, wherein the voice communication device is detected. A voice communication system in which a server device and a plurality of voice communication devices are connected to each other,
Each voice communication device picks up the voice of the speaker and outputs a voice signal;
A speech speed conversion unit for converting the speech speed of the audio signal output from the microphone;
A voice signal output unit that outputs the voice signal converted by the voice speed conversion unit to another device;
A speech speed detection unit for detecting a speech speed of a speaker from an audio signal output from the microphone;
A speech rate information communication unit that transmits information indicating the speech rate of the speaker detected by the speech rate detection unit to the server device as speech rate information;
With
The server device includes target speech speed determining means for determining a target speech speed based on the speech speed information transmitted from each voice communication device and distributing the target speech speed to each voice communication device,
The speech rate information communication unit of the voice communication device further receives the target speech rate from a server device,
The voice communication apparatus determines a speech speed conversion amount of the speech speed conversion unit based on a speaker's speech speed detected by the speech speed detection unit and a speech speed included in the target speech speed amount. A speech communication system further comprising a speech speed conversion amount determination unit.   The speech speed detection unit cuts out a consonant speech segment and a vowel speech segment from a speech signal output from the microphone, and based on a ratio of time lengths of the consonant speech segment and the vowel speech segment, The voice communication system according to claim 3, wherein the voice communication system is detected.

Images