JP4840082B2 - Voice communication device - Google Patents

Description

  The present invention relates to a voice communication device that transmits a collected voice signal of its own device to a counterpart device, and receives and emits a sound emission sound signal from the counterpart device.

  At present, various audio communication systems have been devised in which a plurality of points distant from each other are connected by a network to perform an audio conference or the like.

  For example, Patent Literature 1 includes a main unit at each point, a microphone, a speaker, and an RFID reader according to the number of participants, and a voice conference in which the main units are connected via a network to perform a voice conference. A system is disclosed.

By the way, in such a voice conference system, since participants with various voice qualities and voice volumes participate, it may be difficult to hear the other party's conference person depending on the voice qualities and voice volumes. As a method for solving this, in Patent Document 1, a voice correction amount is stored in advance for each user, speaker identification is performed by RFID or the like at the time of a conference, and voice correction peculiar to each identified speaker is performed. Is going. As a sound quality correction method, Patent Document 2 discloses a method of detecting a speaker formant and transforming it into a formant that is easy to hear.
JP 2005-80110 A Japanese Unexamined Patent Publication No. 1-93796

  However, in the method of Patent Document 1, it is necessary to previously store identification information and audio correction amounts of all users who can participate in the conference, and the user identification process becomes complicated. Moreover, when a new user participates in a meeting, the identification information acquisition process for every user must be performed, and the usability was bad for the user.

  Further, in the method of Patent Document 2, the formant is completely deformed, so that it is impossible for the listener to reliably identify the speaker.

  Accordingly, an object of the present invention is to enable sound that is easy to hear for the listener to be heard by correcting the speaker so as to be identifiable even if the voice is from a speaker whose voice quality or volume is difficult to hear. It is to constitute a communication device.

  The voice communication device according to the present invention is configured to collect a voice collecting unit for collecting voices of conference participants around the device, and analyze a voice feature amount of the collected voice signal from a plurality of preset voice feature amount groups. The collected sound feature detecting means for detecting the corresponding sound feature group, the collected sound correcting means for performing the sound correction processing according to the detected sound feature group, and the collected sound signal after the sound correction processing as the other party And a communication means for transmitting to the destination device.

  In this configuration, the collected sound feature detecting unit detects the sound feature amount of the collected sound signal collected by the sound collecting unit, and the corresponding sound feature group is determined. The collected sound correction unit performs sound correction according to the determined sound feature group, and the communication unit transmits the corrected collected sound signal to the counterpart device. By performing such sound correction, the correction amount is set in units of groups having appropriate widths according to sound quality and volume.

  The voice communication device according to the present invention comprises a communication unit that receives a sound signal for sound emission from a destination device, and a plurality of preset sound features by analyzing the sound feature value of the received sound signal for sound emission. A sound feature detector for sound emission that detects a corresponding sound feature group from the sound amount group, a sound corrector for sound emission that performs sound correction processing according to the detected sound feature group, and a release after sound correction processing. And sound emitting means for emitting sound based on the sound signal for sound.

  In this configuration, the sound feature quantity detecting unit detects the sound feature quantity of the sound output sound signal received by the communication means, and the corresponding sound feature group is determined. The sound emitting sound correcting means performs sound correction according to the determined sound feature group, and the sound emitting means emits sound based on the corrected sound emitting sound signal. Even with such audio correction, the correction amount is set in units of groups having appropriate widths according to the sound quality and volume.

  The voice communication device according to the present invention receives the sound output sound signal from the counterpart device by the communication means in addition to the sound collection / acquired sound correction / transmission function described above, and receives the sound output sound signal received. The sound feature detecting means for detecting a sound feature amount group from a plurality of preset sound feature groups, and sound correction according to the detected sound feature group It is characterized by comprising sound emitting sound correcting means for performing processing and sound emitting means for emitting sound based on the sound emitting sound signal after the sound correcting process.

  With this configuration, it is possible to obtain a voice communication apparatus capable of executing both the sound quality correction of the collected sound signal and the sound quality correction of the sound output sound signal. At this time, the analysis of the voice feature quantity and the identification of the voice feature quantity group can be realized by one element.

  The voice communication apparatus according to the present invention is characterized in that the voice feature amount group is set by at least one kind or a combination of a classification by language type, a classification by gender, and a classification by generation.

  In this configuration, the speech feature groups are “Japanese” and “English”, “Woman” and “Men”, “Children”, “Adult (Adolescence)” and “Adult (Old)”, etc. , Each of them is classified according to a criterion capable of acquiring a difference in voice feature amount. Then, by providing individual voice correction characteristics for each group, correction is made so as to be easy to hear as appropriate regardless of which group's voice is collected or a voice signal is received.

  According to the present invention, the collected sound signal of the own device and the sound signal for sound emission from the partner device are classified by the sound feature amount, and the sound quality correction corresponding to each classification is performed. Any voice including difficult voice can be corrected to a voice that is easy to hear and emitted to the user. Thereby, the user can always hear a voice that is easy to hear and can perform a comfortable voice conference.

A voice communication apparatus according to an embodiment of the present invention and a voice communication system using the voice communication apparatus will be described with reference to the drawings.
FIG. 1 is a configuration diagram of a voice communication system according to the present embodiment.
FIG. 2 is a block diagram showing the main configuration of the voice communication apparatus of this embodiment.

  As shown in FIG. 1, the voice communication system according to the present embodiment has a configuration in which voice communication apparatuses 111A and 111B installed at points a and b are connected by a network 100.

  The voice communication device 111A installed at the point a picks up the speech of the conference participants A to G who are present around the device, generates a collected voice signal, converts it into voice data, and converts it into voice data. Send via. In addition, the voice communication device 111A acquires a sound signal for sound emission (sound collected voice signal of the partner device (for example, the voice communication device 111B)) from the voice data acquired via the network 100, and the participants A to G Sounds so that it can be heard. On the other hand, the voice communication device 111B installed at the point b picks up the speeches of the conference participants H to L who are present around the device, generates a collected voice signal, and converts it into voice data. Transmit via the network 100. Also, the voice communication device 111B acquires a sound signal for sound emission from the voice data received via the network 100 (a collected voice signal of the partner device (for example, the voice communication device 111A)), and the conference participants H to L Sounds so that it can be heard.

  The voice communication apparatuses 111A and 111B installed at the two points a and b have the same configuration and the structure shown in FIG. Hereinafter, functions common to the voice communication apparatuses 111A and 111B will be described as the voice communication apparatus 111.

  As shown in FIG. 2, the voice communication apparatus 111 includes a main control unit 10, a communication control unit 11, a sound emission control unit 12, a D / A converter 13, a sound emission amplifier (AMP) 14, speakers SP1 to SP16, a microphone. MIC 101 to MIC 116, MIC 201 to MIC 216, a sound collection amplifier (AMP) 15, an A / D converter 16, a sound collection control unit 17, an echo cancellation unit 18, an audio signal correction unit 19, and an operation unit 20.

  The voice communication apparatus 111 according to the present embodiment has a substantially rectangular parallelepiped shape whose casing is long in one direction. In the following description, of the four side surfaces of the housing, the long surface is referred to as a long surface, and the short surface is referred to as a short surface.

  An operation unit 20 including a plurality of buttons and a display screen is installed at one end of the upper surface of the housing in the long direction. These operation units 20 are connected to a main control unit 10 installed in the housing, receive operation inputs from conference participants, output them to the main control unit 10, and display operation contents, execution modes, and the like on a display screen. . Although not shown, various input / output interface terminals such as a network connection terminal are installed on the short surface of the housing on the side where the operation unit 20 is installed.

  Speakers SP <b> 1 to SP <b> 16 having the same shape are installed on the lower surface of the casing of the voice communication device 111. These speakers SP1 to SP16 are installed in a straight line at regular intervals along the longitudinal direction, thereby constituting a speaker array. Microphones MIC101 to MIC116 having the same shape are installed on one long surface of the housing. These microphones MIC101 to MIC116 are installed in a straight line at regular intervals along the longitudinal direction, thereby forming a microphone array. Further, microphones MIC201 to MIC216 having the same shape are also installed on the other long surface of the housing. These microphones MIC201 to MIC216 are also installed in a straight line at regular intervals along the longitudinal direction, thereby forming a microphone array. A bottom grille (not shown) formed in a shape covering the speaker array and the microphone array and punch meshed is installed on the lower surface side of the housing. In this embodiment, the number of speakers in the speaker array is 16 and the number of microphones in each microphone array is 16. However, the present invention is not limited to this, and the number of speakers and the number of microphones may be set as appropriate according to the specifications. That's fine.

  The main control unit 10 performs overall control of the audio conference apparatus, and also performs control such as power on / off input from the operation unit 20 and other control of the signal processing system. Further, when the main control unit 10 receives the sound emission sound correction control from the conference person via the operation unit 20, the main control unit 10 gives a sound emission sound correction control signal to the sound signal correction unit 19. Further, when the main control unit 10 receives the collected sound correction control from the conference via the operation unit 20, the main control unit 10 gives the collected sound correction control signal to the sound signal correcting unit 19. Further, the main control unit 10 gives the speaker control data to the communication control unit 11 when acquiring the speaker orientation data from the sound collection control unit 17 described later.

  As a reception-side function, the communication control unit 11 is connected to the network 100 and demodulates speaker orientation data and sound emission sound signals from voice data received from other devices via the network 100, The azimuth data is output to the main control unit 10 and the sound output sound signal is output to the sound signal correction unit 19.

  When the sound signal correcting unit 19 receives a sound emitting sound correction control signal from the main control unit 10 as a reception side function, the sound signal correcting unit 19 corrects the sound emitting sound signal from the communication control unit 11 and passes through the echo canceling unit 18. The corrected sound emission sound signal is output to the sound emission control unit 12. On the other hand, if the sound signal correction unit 19 does not receive the sound output sound correction control signal from the main control unit 10, the sound signal output from the communication control unit 11 is directly used as the sound output control unit via the echo cancellation unit 18. 12 is output. The detailed configuration and operation of the audio signal correction unit 19 will be described later.

  When the sound emission control unit 12 acquires the sound emission control data set by the main control unit 10 based on the speaker orientation data from the partner device, the sound emission control unit 12 performs delay processing, amplitude processing, and the like on the input sound emission sound signal. To generate individual sound emission signals corresponding to the speakers SP1 to SP16 so as to form sound emission beams with the characteristic of sound emission equally to all the conference persons present around the audio conference apparatus.

  Each D / A converter 13 performs digital-analog conversion on the input individual sound emission signal and supplies it to each sound emission amplifier 14, and each sound emission amplifier 14 amplifies the analog individual sound emission signal, It gives to each speaker SP1-SP16. Each of the speakers SP1 to SP16 converts the input electrical individual sound emission signal into sound and emits the sound.

  The microphones MIC101 to MIC116 and MIC201 to MIC216 collect ambient sounds including utterances from speakers present around the device, convert them into electrical sound collection signals, and provide them to the sound collection amplifier 15 . The sound collection amplifier 15 amplifies the sound collection signal and applies it to the A / D converter 16, and the A / D converter 16 converts the analog sound collection signal into a digital signal and outputs it to the sound collection control unit 17.

  The sound collection control unit 17 performs a delay process on the sound collection signals of the microphones MIC101 to MIC116 and MIC201 to MIC216, and generates a sound collection beam signal having a strong directivity in a predetermined direction including the direction of each conference person. To do. The sound collecting control unit 17 compares the amplitudes of the generated sound collecting beam signals in the respective directions, selects the sound collecting beam signal having the largest amplitude, and outputs it to the echo canceling unit 18 as a sound collecting sound signal. In addition, the sound collection control unit 17 extracts a sound collection direction corresponding to the selected sound collection beam signal, and provides it to the main control unit 10 as the speaker direction data.

  The echo cancellation unit 18 includes an adaptive filter and a post processor, generates a pseudo regression sound signal based on the sound output sound signal by the adaptive filter, and generates a pseudo regression sound from the collected voice signal of the sound collection control unit 17. The signal is subtracted and output to the audio signal correction unit 19. By performing such echo cancellation processing, the wraparound sound from the speaker SP to the microphone MIC is suppressed.

  When the sound signal correcting unit 19 receives a sound collecting sound correction control signal from the main control unit 10 as a transmission side function, the sound signal correcting unit 19 corrects the sound collecting sound signal output from the echo canceling unit 18 and outputs it to the communication control unit 11. To do. On the other hand, if the sound signal correcting unit 19 does not receive the sound collecting sound correction control signal from the main control unit 10, the sound signal correcting unit 19 outputs the sound collecting sound signal output from the echo canceling unit 18 to the communication control unit 11 as it is. The detailed configuration and operation of the audio signal correction unit 19 will be described later.

  As a transmission side function, the communication control unit 11 attaches the speaker orientation data from the main control unit 10 and the device ID data to the collected voice signal from the voice signal correction unit 19 to form a network communication format. Are transmitted to the network 100.

Next, the configuration and operation processing of the audio signal correction unit 19 will be described more specifically.
FIG. 3 is a block diagram showing the main configuration of the audio signal correction unit 19 shown in FIG. FIG. 4 is a diagram showing an example of a voice feature group stored in the EQ table 194.

  As shown in FIG. 3, the audio signal correction unit 19 includes a correction control CPU 190, a formant detection unit 191, a sound collection EQ processing unit 192, a sound emission EQ processing unit 193, and an EQ table 194 that is a memory.

  The EQ table 194 stores speech feature amount groups classified for each combination having different formant features. For example, as shown in FIG. 4, "Japanese and English", "Women and Men", "Children, Adults (Adolescence), and Adults (Old)" are set as three different combinations of formants, Twelve types of speech feature groups G11 to G13, G21 to G23, G31 to G33, and G41 to G43 classified from these three types are stored. The contents stored corresponding to each voice feature group G are the group name G, formants that are the characteristics of each group, and EQ correction characteristics of the voice signal. For example, in the case of a group G12, a general person represents a general formant representing a combination of the group name G12, “Japanese”, “female”, and “adult (youth)” and the sound signal of the formant. An EQ correction coefficient for conversion into an audio signal having a flat frequency characteristic that is easy to hear is stored. The combination criterion is not limited to this example, and further finer criteria such as a criterion based on a combination of “fast mouth” and “slow” may be additionally set. Moreover, you may make it set audio | voice feature parameters, such as fundamental frequencies other than a formant, a pitch, intonation, and a cepstrum.

<When correcting sound output for sound emission>
When the formant detection unit 191 receives the sound emission correction control signal from the main control unit 10, the formant detection unit 191 detects the formant by dividing the sound emission sound signal input from the communication control unit 11 by a predetermined time length at every predetermined timing. To the correction control CPU 190.
When the correction control CPU 190 receives a sound emission sound correction control signal from the main control unit 10, it performs sound emission sound correction processing. Specifically, the correction control CPU 190 detects the similarity between the formants detected by the formant detection unit 191 and the formants of each group G stored in the EQ table 194, and selects the group with the highest similarity. To do. The correction control CPU 190 reads the EQ correction coefficient of the selected group G and provides it to the sound emission EQ processing unit 193.

  The sound emission EQ processing unit 193 performs an equalizer process on the sound emission sound signal from the communication control unit 11 using the given EQ correction coefficient, and converts the sound emission sound signal into, for example, a substantially flat frequency characteristic. And output to the sound emission control unit 12.

<When correcting the collected sound signal>
When the formant detection unit 191 receives the collected sound correction control signal from the main control unit 10, the formant detection unit 191 detects the formant by dividing the collected sound signal output from the echo cancellation unit 18 by a predetermined time length at every predetermined timing, and corrects it. This is given to the control CPU 190.
When the CPU 190 for correction control receives a sound collection sound correction control signal from the main control unit 10, it performs sound collection sound correction processing. Specifically, the correction control CPU 190 detects the similarity between the formants detected by the formant detection unit 191 and the formants of each group G stored in the EQ table 194, and selects the group with the highest similarity. To do. The correction control CPU 190 reads the EQ correction coefficient of the selected group G and supplies it to the sound collection EQ processing unit 192.

  The sound collection EQ processing unit 192 performs an equalizer process on the collected sound signal from the echo canceling unit 18 using the given EQ correction coefficient, and converts it into, for example, a sound emission sound signal having a substantially flat frequency characteristic. And output to the communication control unit 11.

  By correcting the emitted sound signal by such an equalizer process, it is possible to clearly hear the sound even when it is difficult to hear the original sound. Furthermore, since the reverberation of the original speech can be left rather than directly correcting the formant as in the prior art, the listener can easily identify the speaker.

  In addition, it is possible to correct the sound quality by a simple process, compared to the case where speaker identification is performed as in the prior art, and then correction characteristics are read and corrected for each speaker. Thereby, sound quality correction can be performed more reliably and quickly.

  In the above-described apparatus, a configuration is shown in which a functional unit that performs correction processing of a sound emission sound signal and a functional unit that performs correction processing of a collected sound signal are simultaneously provided. However, in a case where the sound collection side voice communication device and the sound emission side voice communication device are separated, the sound collection side voice communication device includes only a function unit that performs correction processing of the collected sound signal. The sound communication device on the sound emission side only needs to have a function of correcting the sound signal for sound emission.

  In the above-described apparatus, the sound quality of sound emission sound signals is collectively corrected, and the sound is emitted with the same sound quality to all the conference participants. However, since not all of the conference participants have the same auditory characteristics, sound quality correction characteristics may be set for each conference participant. In this case, in addition to the above-described EQ correction characteristics, by executing EQ correction based on the additional correction characteristics input from the conference through the operation unit 20 or the like, individual sound quality correction can be performed for each conference. . By adopting such a configuration, each conferee can further listen to the remarks of the conferee of the counterpart device with sound quality suitable for him.

It is a block diagram of the audio | voice communication system of this invention. It is a block diagram which shows the main structures of the audio | voice communication apparatus of embodiment of this invention. It is a block diagram which shows the main structures of the audio | voice signal correction | amendment part 19 shown in FIG. It is a figure which shows the example of the audio | voice feature-value group memorize | stored in EQ table 194. FIG.

Explanation of symbols

100-network, 111, 111A, 111B-voice communication device, 10-main control unit, 11-communication control unit, 12-sound emission control unit, 13-D / A converter, 14-sound emission amplifier, 15-collection Sound amplifier, 16-A / D converter, 17-sound collection control unit, 18-echo cancellation unit, 19-sound signal correction unit, 190-correction control CPU, 191-formant detection unit, 192-sound collection EQ processing Section, 193-sound emission EQ processing section, 194-EQ table, 20-operation section

Claims (4)

Sound collection means for collecting the utterances of conference participants around the device;
A formant detection means for detecting a formant of collected sound signals,
Storage means for storing correction characteristics for each of a plurality of preset formants;
Control means for referring to the storage means with the formants detected by the formant detection means, and for reading the correction characteristics of the corresponding formants ;
Sound collection sound correction means for performing correction processing of the sound collected by the sound collection means using the correction characteristic read by the control means ;
And a communication means for transmitting the collected sound signal after the sound correction processing to the counterpart device. A communication means for receiving a sound signal for sound emission from the counterpart device;
A formant detection means for detecting a formant of the received sound emitting signal,
Storage means for storing correction characteristics for each of a plurality of preset formants;
Control means for referring to the storage means with the formants detected by the formant detection means, and for reading the correction characteristics of the corresponding formants ;
A sound emission correcting means for correcting the received sound output sound signal using the correction characteristic read by the control means ;
A sound communication device comprising: sound emitting means for emitting sound based on a sound emission sound signal after the sound correction processing. The communication means receives the sound output sound signal from the counterpart device,
The formant detection means detects the formant of the received sound emitting signal,
The control means refers to the storage means with the formant detected by the formant detection means , reads the correction characteristic of the corresponding formant ,
A sound emission correcting means for correcting the received sound output sound signal using the correction characteristic read by the control means ;
The voice communication device according to claim 1, further comprising: a sound emitting unit that emits sound based on the sound output sound signal after the sound correction processing. The voice communication according to any one of claims 1 to 3, wherein the storage means stores the correction characteristic for each of at least one kind or a combination of classification by language type, classification by gender, and classification by generation. apparatus.

Images