JP2007336395A - Voice processor and voice communication system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a voice processor and a voice communication system which enable a listener to easily catch keywords emphasized by a speaker, without registering keywords of various languages in a keyword dictionary beforehand. <P>SOLUTION: To each region of an input voice signal where a change in acoustic feature is relatively large, the voice processor 42 applies emphasis processing for increasing the change in acoustic feature further, and outputs it. Accordingly, since each keyword being a part which the speaker wants to tell is emphasized further, even when a conversation is made between remote places without looking at a figure of each partner, even when a language used for the conversation is unfamiliar to them, or the like, the listener can surely comprehend contents which the speaker wants to tell. Moreover, keyword registration work is not required, and the voice processor and the voice communication system can be used in each country, since the voice processor and the voice communication system are so structured that a keyword dictionary can be dispensed with, and keywords can be extracted without any problems irrespective of language. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a voice processing device and a voice communication system used when talking by exchanging voice signals.

  2. Description of the Related Art Conventionally, various audio communication systems have been devised in which a plurality of points are connected via a network to perform an audio conference or chat (see Patent Document 1).

  For example, Patent Document 1 discloses a system in which each speaker individually connects a personal computer corresponding to a voice communication device to a network and performs a conference with each other in a virtual conference room. In this system, each speaker individually operates a voice communication device, adjusts the sound quality, volume, and sound of the received voice signal and emits sound, so that a conference with a sense of reality can be made for each speaker. Can do.

  However, in a multipoint communication audio conference, the volume of the voice and the conversation speed vary depending on the speaker, and therefore it may be difficult to hear a specific speaker. For example, in the voice communication system of Patent Document 1, when the voice of a specific speaker is too low, the sound quality, volume, and sound of the received voice signal can be adjusted so as to be easily heard. However, since the received voice signal for each speaker cannot be individually adjusted, not only the voice of a specific speaker but also the voices of other speakers are increased. In addition, when the speaking speed is too fast to hear the speaker's speech, the speaking speed cannot be adjusted.

On the other hand, as a device for solving such a problem, there has conventionally been a speech enhancement device that can emphasize important words (keywords) in input speech (see Patent Document 2).
JP-A-8-125761 JP-A-5-27792

  In general, regardless of the type of language, a person speaks the part (keyword) that he / she wants to convey during a conversation with a loud voice, such as speaking speed, volume, and pitch (pitch). There is a habit of changing the acoustic feature amount to be larger or smaller than other regions. For example, if a speaker says, “I love apples”, the acoustic features of any of the keywords “I”, “Apples”, or “I love” are changed to be larger or smaller. And try to convey the keyword to the listener. In addition, the content that the speaker wants to convey changes depending on which word the acoustic feature amount is increased or decreased. Therefore, if the listener can hear the keyword whose acoustic feature amount has changed, the listener can roughly grasp the content that the speaker wants to convey. However, in voice conferences where the other party's face cannot be seen, or when the listener is talking in an unfamiliar language, the listener will not be able to change the acoustic feature of the part that the speaker wants to convey. I sometimes missed out and couldn't understand what the speaker wanted.

  Further, in the conventional speech enhancement device, the keyword is emphasized, so that the listener can grasp the keyword. However, in the conventional speech enhancement device, it is necessary to register keywords in the keyword dictionary in advance, and words that are not registered in the keyword dictionary are not emphasized. In a conference or voice chat, the keyword changes depending on the conversation flow. Therefore, when a conventional speech enhancement device is applied to a conventional speech communication system, if a word that is not registered in the keyword dictionary is a keyword, the enhancement process is not performed on the word, and the listener misses the keyword. Therefore, there is a possibility that the content that the speaker wants to convey cannot be grasped.

  Furthermore, in order to make the conventional speech enhancement apparatus compatible with various languages, it is necessary to register keywords in various languages in the keyword dictionary, which is complicated.

  Accordingly, an object of the present invention is to provide a voice processing device and a voice communication system that allow a listener to easily listen to a keyword emphasized by a speaker without previously registering keywords in various languages in a keyword dictionary. To do.

  The present invention has the following configuration as means for solving the above problems.

(1) a feature amount extraction unit that analyzes an acoustic feature amount in an input audio signal and extracts a region where the acoustic feature amount is relatively larger or smaller than other regions;
The speech signal region extracted by the feature amount extraction means is subjected to an emphasis process in which the feature amount is further increased when the feature amount is relatively large, and is further decreased when the feature amount is relatively small. Emphasis means to output,
It is provided with.

  In general, the speaker can use the speed, volume, and pitch (pitch) of the keyword part regardless of the language so that the listener can easily understand the part (keyword) that the speaker wants to convey to the other party during the conversation. And so on, with the acoustic feature amount being relatively large or small relative to other regions. For example, the speaker speaks the keyword portion slowly and loudly. However, in voice conferences where the other party's face is not visible, or when the listener is talking in an unfamiliar language, the listener can change the part even if the acoustic feature of the part that the speaker wants to communicate changes. I missed it.

  In this configuration, the speech processing apparatus performs enhancement processing for further increasing or decreasing the acoustic feature amount in a region where the acoustic feature amount is relatively large or small in the input sound signal. To output. Therefore, even if you have a conversation between remote locations without seeing the other person's figure, or if you are unfamiliar with the language used for conversation, the acoustic features of the keyword part that the speaker wants to convey Because the amount is emphasized larger or smaller, the listener can reliably understand what the speaker wants to convey.

(2) The feature amount extraction unit includes a low speed region extraction unit that analyzes a speech speed in a speech signal and extracts a region having a relatively low speech speed compared to other regions,
The enhancement means performs an enhancement process for further reducing the speech speed on the region extracted by the low speed region extraction means, and outputs the result.

  In this configuration, the speech processing apparatus performs an enhancement process for further reducing the speech speed on an input speech signal in a region where the speech speed is relatively low (slow) compared to other regions, and outputs the result. To do. Therefore, since the speaking speed is further reduced for the keyword portion that the speaker wants to convey, the listener can surely hear the keyword, so that the content that the speaker wants to convey can be easily grasped.

(3) The feature amount extraction unit includes a large volume range extraction unit that analyzes a volume level of the audio signal and extracts a region having a relatively large volume compared to other regions,
The enhancement means performs an enhancement process for further increasing the volume of the region extracted by the high volume range extraction means, and outputs the result.

  In this configuration, the audio processing device performs an enhancement process for further increasing the volume of an input audio signal with respect to an area where the volume is relatively higher than other areas, and outputs the result. Accordingly, since the volume of the keyword portion that the speaker wants to convey is further increased, the listener can surely hear the keyword, so that the content that the speaker wants to convey can be easily grasped.

(4) 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
A microphone unit,
A speaker unit;
A communication means for transmitting the audio signal collected by the microphone unit to another audio communication device via the server device, and receiving the audio signal transmitted by the server device;
Speaker control means for inputting an audio signal received by the communication means to the speaker unit;
With
The server device
A voice processing device according to any one of claims 1 to 3;
Server communication that receives an audio signal transmitted from the audio communication device to another audio communication device, inputs the audio signal to the audio processing device, and transmits an audio signal output from the audio processing device to the other audio communication device Means,
It is provided with.

  In this configuration, in the voice communication system, the voice signal picked up by the voice communication device is transmitted to the server device, and the server device is applied to an area where the change in the acoustic feature amount of the voice signal is relatively large. Then, enhancement processing for further increasing the change in the feature amount is performed, and this audio signal is output to another audio communication device. Therefore, in the audio communication system, the server device performs the audio signal emphasis process, so that each audio communication apparatus can transmit and receive the audio signal without increasing the load due to the emphasis process.

(5) A voice communication system in which a plurality of voice communication devices are connected to each other,
Each voice communication device
A voice processing device according to any one of claims 1 to 3;
A microphone unit,
A speaker unit;
A communication means for transmitting the audio signal collected by the microphone unit to another audio communication device, receiving the audio signal transmitted by another audio communication device, and inputting the audio signal to the audio processing device;
Speaker control means for inputting an audio signal output from the audio processing device to the speaker unit;
It is provided with.

  In this configuration, in the voice communication system, the enhancement process is performed on the voice signal transmitted from another voice communication apparatus, and the voice subjected to the enhancement process is emitted from the speaker unit. Therefore, by using this voice communication system, the listener can easily understand what the speaker wants to convey. In addition, since it is possible to perform enhancement processing even for voice signals sent from voice communication devices that do not have voice enhancement means, the listener can surely hear the keywords and what the speaker wants to convey Can be reliably grasped.

(6) A voice communication system in which a plurality of voice communication devices are connected to each other,
Each voice communication device
A voice processing device according to any one of claims 1 to 3;
A microphone unit that picks up sound and inputs a sound signal to the sound processing device;
A speaker unit;
A communication means for transmitting the audio signal output by the audio processing device to another audio communication device and receiving the audio signal transmitted by another audio communication device;
Speaker control means for inputting an audio signal received by the communication means to the speaker unit;
It is provided with.

  In this configuration, in the audio communication system, an enhancement process that further increases the change in the feature amount relative to the region in which the change in the acoustic feature amount is relatively large with respect to the sound signal collected by the microphone unit. Then, the audio signal is transmitted to another audio communication device. Therefore, by using this voice communication system, the listener can easily understand what the speaker wants to convey.

In addition, since the voice signal subjected to the enhancement process can be transmitted to the voice communication apparatus that does not include the voice enhancement means, it is possible to reliably transmit the contents to be transmitted to the listener.

  The speech processing apparatus according to the present invention further increases or decreases the acoustic feature amount of an input speech signal with respect to a region where the acoustic feature amount is relatively large or small compared to other regions. Output with emphasis processing. Therefore, even if you have a conversation between remote locations without seeing the other person's figure, or if you are unfamiliar with the language used for the conversation, the keywords that the speaker wants to convey are more emphasized, The listener can easily understand what the speaker wants to convey.

  In the audio communication system of the present invention, when the server device and a plurality of audio communication devices are connected to each other, the server device includes an audio processing device, and the feature amount in the audio signal is relative to that of other regions. For an area that is large or small, an enhancement process that further increases or decreases the feature amount is performed and output. In the audio communication system of the present invention, in the case of a configuration in which a plurality of audio communication devices are connected to each other, each audio communication device includes an audio processing device and performs enhancement processing on the audio signal receiving side, or Emphasis processing is performed on the audio signal transmission side. Therefore, by using this voice communication system, the listener can easily understand what the speaker wants to convey.

  In addition, the speech processing apparatus and speech communication system of the present invention have a configuration that does not require a keyword dictionary, and can extract keywords in any language without any problem, so no keyword registration work is required, and any country in the world Can be used in

  In the present invention, when a voice signal of a speaker is input, an acoustic feature amount (speech speed, volume, pitch (pitch), etc.) in the voice signal is analyzed and compared with other parts of the voice signal. When a region with a relatively large or small acoustic feature value is extracted and the feature value is relatively large, the feature value is further increased, and when the feature value is relatively small. Further performs enhancement processing to reduce the feature amount and outputs the result.

  Specifically, in the speech processing device of the present invention, a change in speech speed and volume is detected as an acoustic feature amount of the speech signal, and a region where the speech speed is relatively decreased or a region where the volume is relatively increased Is extracted to further reduce the speech speed of the audio signal in the area or further increase the volume. As a result, the feature quantity is further emphasized for the keyword whose feature quantity has been changed as a part that the speaker wants to convey, so that the listener can speak in a remote conference when the speaker's face cannot be seen or in an unfamiliar language. You can listen to keywords even if you are. Therefore, the listener can hear the change in the feature amount with the emphasized keyword, and can surely grasp the content that the speaker wants to convey.

  FIG. 1 is an image diagram for explaining processing performed by the speech processing apparatus of the present invention. If the speaker says, “I love apples,” the speaker changes the inflection in particular, extracts the region that spoke slowly and loudly, and further emphasizes that region. That is, when the inflection (acoustic feature amount) of the “I am” part is changed as shown in FIG. 1 (A), the speech speed of this part is relatively lower than the other parts. The volume is relatively increased compared to other parts. In the speech processing apparatus of the present invention, the speech speed of the word “I am”, which is a region in which the acoustic feature amount has changed relative to other regions, is further reduced, and the volume is further increased. Further, when the inflection of the “apple is” portion is changed as shown in FIG. 1B, the speech speed of this portion is similarly further lowered and the volume is further increased. Further, when the inflection of the “large (like)” portion is changed as shown in FIG. 1C, the speech speed of this portion is further reduced and the volume is further increased.

  FIG. 2 is a configuration diagram showing an outline of the voice communication system. When the voice processing device of the present invention is applied to a voice communication system that is a voice conference or voice chat configuration between a plurality of points, the following configuration is obtained. Here, FIG. 2 shows a state in which the audio signal is transmitted in one direction, but there is a case where the audio signal is transmitted in the opposite direction. Therefore, in the following description, the audio signal is transmitted in the opposite direction. The explanation conveyed by is written in parentheses. FIG. 2 shows an example in which a telephone is used as a voice communication device constituting the voice communication system.

  When the server device 13 and the plurality of telephones 11 and 12 are connected to each other via the network 10 as in the voice communication system 1 shown in FIG. A configuration in which a sound processing device is provided is preferable. In this configuration, the telephone 11 (12) transmits the voice signal input from the receiver 11J (12J) to the server device 13 via the network 10. Then, the enhancement processing is performed by the voice processing device 42 built in the server device 13, and the voice signal subjected to the enhancement processing is transmitted to the telephone set 12 (11) via the network 10. When the telephone 12 (11) receives the voice signal subjected to the enhancement processing, the telephone 12 (11) emits a sound from the receiver 12J (11J).

  When a plurality of telephones 14 and 15 are connected to each other via the network 10 as in the voice communication system 2 shown in FIG. To enhance the received audio signal. In this configuration, the telephone 14 (15) transmits the voice signal input from the receiver 14J (15J) to the telephone 15 (14) via the network 10. When the telephone 15 (14) receives the audio signal, the telephone 15 (14) performs enhancement processing with the voice processing device 42 built in the telephone 15 (14), and emits the voice from the receiver 15J (14J).

  When a plurality of telephones 16 and 17 are connected to each other via the network 10 as in the voice communication system 3 shown in FIG. And an enhancement process may be performed on the audio signal to be transmitted. In this configuration, the telephone 16 (17) performs an emphasis process on the voice signal input from the receiver 16 J (17 J) by the voice processing device 42 included in the telephone 16 (17), and the telephone 16 via the network 10. 17 (16). When the telephone set 17 (16) receives the voice signal, the telephone set 17J (16J) emits the voice.

  Next, a specific configuration of the voice communication system 1 will be described. FIG. 3 is a block diagram showing the configuration of the voice communication system. The voice communication system 1 has a configuration in which a telephone 11, a telephone 12, and a server device 13 are connected to a network 10. FIGS. 2A and 3 show a configuration in which two telephones 11 and 12 are connected to the network 10, but this is for the sake of simplification, and a plurality of telephones (voices). It is possible to connect a communication apparatus) to the network 10 and exchange audio signals.

  The telephone set 11 and the telephone set 12 have the same configuration. The communication unit 21, the D / A converter 22, the sound emission amplifier 23, the speaker unit SP1, the microphone unit MIC1, the sound collection amplifier 24, the A / D converter 25, the operation A unit 26, a display unit 27, and a control unit 28.

  The control unit 28 controls each unit of the telephone 11 (12).

  The communication unit 21 converts an input audio signal input from the server device 13 via the network 10 from a data format (protocol) corresponding to the network and outputs the converted signal to the D / A converter 22.

  The D / A converter 22 converts the audio signal sent from the communication unit 21 into an analog format and outputs it to the sound emission amplifier 23.

  The sound emission amplifier 23 amplifies the audio signal and gives it to the speaker unit SP1 (SP2).

  The speaker unit SP1 (SP2) converts the sound signal provided from the sound emission amplifier 23 into sound and emits the sound.

  The microphone unit MIC1 (MIC2) collects the voice uttered by the speaker and converts it into an audio signal that is an electrical signal, and outputs the audio signal to the sound collection amplifier 24.

  The sound collecting amplifier 24 amplifies the audio signal and outputs it to the A / D converter 25.

  The A / D converter 25 converts the analog audio signal input from the sound collection amplifier 24 into a digital audio signal and outputs the digital audio signal to the communication unit 21.

  The communication unit 21 converts the audio signal output from the A / D converter 25 into a data format (protocol) corresponding to the network, and transmits it to the server device 13 via the network 10. Note that destination information indicating which voice communication device is connected to the network 10 via the server device 13 is added to the voice signal.

  The server device 13 includes a communication unit 41, a voice processing device 42, and a control unit 43. In addition, the speech processing device 42 includes an intonation analysis unit 51, a feature extraction unit 52, an enhancement region determination unit 53, and an enhancement processing unit 54.

  The intonation analysis unit 51 includes a speech speed analysis unit 511 and a sound volume analysis unit 512.

  The feature extraction unit 52 includes a low speed region extraction unit 521 and a large volume region extraction unit 522.

  The control unit 43 controls each unit of the server device 13.

  The communication unit 41 converts an input voice signal from the telephone set 11 or the telephone set 12 input via the network 10 from a data format (protocol) corresponding to the network, and converts the speech speed analysis unit 511 of the inflection analysis unit 51 and Output to volume analysis unit 512.

  The speech speed analysis unit 511 analyzes the speech speed of the voice signal transmitted from the communication unit 41 and outputs a signal to the low speed region extraction unit 521 of the feature extraction unit 52 when the speech speed decreases. Specifically, the speech speed analysis unit 511 extracts a speech signal at a phoneme level, extracts a consonant or a vowel, and analyzes the length and interval of each sound. Then, the average value of the length and interval of each sound is calculated, and if the length and interval of each sound are within a certain range from the average value and the speech speed is faster than a preset speech speed threshold, Judge that the speed is stable. On the other hand, the speech speed analysis unit 511, when the length or interval of each sound temporarily becomes larger than the value from the average value to a certain range and the speech speed falls below the threshold of the speech speed, the acoustic feature It is determined that the amount is small, and a signal to that effect is output to the low speed region extraction unit 521.

  The volume analysis unit 512 analyzes the volume of the audio signal sent from the communication unit 41 and outputs a signal to the loud volume range extraction unit 522 of the feature extraction unit 52 when the speech speed volume increases. Specifically, the volume analysis unit 512 cuts out the audio signal into phoneme levels, extracts consonants or vowels, and analyzes the volume level of each sound. Then, the average value of the volume level of each sound is calculated, and if the volume level of each sound is within a certain range from the average value and the volume is lower than a preset volume threshold, the volume is stable. It is determined that On the other hand, when the volume level of each sound temporarily becomes larger than the value from the average value to a certain range, and the volume becomes larger than the volume threshold, the volume analysis unit 512 increases the acoustic feature amount. And outputs a signal to that effect to the loud volume extraction unit 522.

  While the signal is sent from the speech speed analysis unit 511, the low speed region extraction unit 521 extracts a speech signal and outputs the extracted region to the enhancement region determination unit 53 as a low speed region of the sound speed.

  The loud sound volume extraction unit 522 extracts the audio signal while the signal is sent from the sound volume analysis unit 512, and outputs the extracted region to the enhancement region determination unit 53 as the loud sound volume range.

  The enhancement region determination unit 53 determines, in the audio signal, an area where the low speed region of the sound speed input from the low speed region extraction unit 521 and the high sound volume region input from the high sound volume region extraction unit 522 overlap in the audio signal. It is determined that the portion is emphasized, and a control signal is output to the enhancement processing unit 54 so that enhancement processing is performed on the region.

  The enhancement processing unit 54 performs enhancement processing on the audio signal based on the control signal output from the enhancement region determination unit 53. That is, the speech speed is decreased by a certain amount and the sound volume is increased by a certain amount with respect to the region of the sound signal designated by the control signal. For example, the enhancement processing unit 54 increases the speech speed of the audio signal by 0.7 times and increases the volume of the audio signal by 1.5 times. In addition, the enhancement processing unit 54 does not perform processing for decreasing the speech speed or increasing the volume for a region for which no control signal is output from the enhancement region determination unit 53. Further, in the enhancement processing unit 54, in order to prevent a time lag from occurring between the voice communication devices by performing the processing for reducing the speech speed as described above, the enhancement processing unit 54 depends on the time for performing the speech speed reduction processing. Then, a known process for shortening the audio signal by cutting the silent region of the audio signal is performed. Then, the enhancement processing unit 54 outputs the audio signal subjected to these processes to the communication unit 41.

  The communication unit 41 converts the audio signal output from the enhancement processing unit 54 of the audio processing device 42 into a data format (protocol) corresponding to the network, and is a telephone that is a specified audio communication device via the network 10. 12 or the telephone 11.

  Next, operations of the telephone set 11, the telephone set 12, and the server apparatus 13 of the voice communication system 1 will be described based on flowcharts. FIG. 4 is a flowchart for explaining the operation of the server apparatus.

  In the voice communication system 1, when the speaker A and the speaker B have a conversation using the telephone 11 and the telephone 12, the following processing is performed.

  First, when the speaker A starts speaking using the handset 11J of the telephone set 11 and voice input is started from the microphone unit MIC1 (s1), the microphone unit MIC1 converts the collected voice into a voice signal. The sound collecting amplifier 24 amplifies the sound signal, and the A / D converter 25 digitizes the sound signal and outputs it to the communication unit 21 (s2). The communication unit 21 sends the digitized audio signal to the network 10 for transmission to the telephone set 12 via the server device 13 (s3). Then, if the telephone set 12 has finished collecting sound and the communication unit 21 has not received an audio signal from the server device 13 (s4), the telephone set 11 performs the processing after step s1.

  When the communication unit 41 receives the voice signal from the telephone 11 (s11), the server device 13 analyzes the speech speed and the volume by the inflection analysis unit 51 of the voice processing device 42 (s12). The server device 13 has a speaking speed that is larger than the value from the average value to a certain range, and the volume level of each sound is temporarily larger than the value from the average value to the certain range. (S13), the low speed region extraction unit 521 and the high volume region extraction unit 522 extract the audio signal of the portion, and the enhancement region determination unit 53 is a portion where the audio signal is enhanced for the overlapping region. And a control signal is output to the enhancement processing unit 54 so that the enhancement process is performed on the region (s14). The enhancement processing unit 54 performs enhancement processing on the audio signal based on the control signal output from the enhancement region determination unit 53. That is, the speech speed is decreased by a certain amount and the sound volume is increased by a certain amount with respect to the region of the sound signal designated by the control signal, and is output to the communication unit 21 (s15).

  The communication unit 41 converts the audio signal output from the enhancement processing unit 54 into a data format (protocol) corresponding to the network, and transmits the data to the telephone set 12 that is a specified audio communication device via the network 10. (S16).

  When the telephone 12 does not receive an audio signal (s21) and the communication unit 21 receives an audio signal from the server device 13 (s24), the D / A converter 22 analogizes the audio signal sent from the communication unit 21. The sound is converted into a format and output to the sound output amplifier 23, and the sound output amplifier 23 amplifies the sound signal and outputs it to the speaker unit SP2 (s25). The speaker unit SP2 converts the sound signal amplified by the sound emission amplifier 23 into sound and emits the sound (s26). When the sound emission is finished, the telephone 12 performs the processing after step s21.

  Further, when the speaker B finishes listening to the speaker A, speaks using the handset 12J of the telephone 12, and starts inputting voice from the microphone unit MIC2 (s21), the microphone unit MIC2 is not stored. The sound that has been sounded is converted into a sound signal, the sound collecting amplifier 24 amplifies the sound signal, and the A / D converter 25 digitizes the sound signal and outputs it to the communication unit 21 (s22). The communication unit 21 sends the digitized audio signal to the network 10 for transmission to the telephone set 12 via the server device 13 (s23). Then, if the telephone set 12 has finished collecting sound and the communication unit 21 has not received an audio signal from the server device 13 (s24), the telephone set 11 performs the processing after step s21.

  When the communication unit 41 receives the audio signal from the telephone set 12 (s11), the server device 13 performs the above-described steps s12 to s16 in the audio processing device 42. That is, the enhancement processing is performed on the voice signal, and the voice signal is transmitted to the telephone 11 that is the designated voice communication apparatus via the network 10.

  When the telephone 11 receives no audio signal (s1) and the communication unit 21 receives the audio signal from the server device 13 (s4), the D / A converter 22 analogizes the audio signal sent from the communication unit 21. The sound is converted into a format and output to the sound output amplifier 23, and the sound output amplifier 23 amplifies the sound signal and outputs it to the speaker unit SP1 (s5). The speaker unit SP1 converts the sound signal amplified by the sound emission amplifier 23 into sound and emits the sound (s6). When the telephone 11 finishes emitting sound, the telephone 11 performs processing from step s1 onward.

  Thus, in the voice communication system 1, since the server apparatus 13 performs the enhancement process on the audio signal output from the telephone set 11 or the telephone set 12, each apparatus can efficiently perform the process.

  Next, a specific configuration of the voice communication system 2 will be described. FIG. 5 is a block diagram showing a different configuration of the voice communication system from FIG. The voice communication system 2 has a configuration in which a telephone 14 and a telephone 15 are connected to a network 10. FIGS. 2B and 5 show a configuration in which two telephones 14 and 15 are connected to the network 10, but this is for the sake of simplification of explanation and a plurality of telephones (voices). It is possible to connect a communication apparatus) to the network 10 and exchange audio signals. Moreover, in the following description, the same code | symbol is attached | subjected to the structure similar to the audio | voice communication system 1, and detailed description is abbreviate | omitted.

  The telephone 14 and the telephone 15 have a configuration in which an audio processing device 42 is provided between the communication unit 21 of the telephone 11 and the D / A converter 22. When the telephone 14 (15) receives the voice signal transmitted from the telephone 15 (14) by the communication unit 21, the voice processor 42 analyzes the voice speed and volume of the voice signal as described above, and the voice signal If the speech speed is relatively decreased and the sound volume is relatively increased, an emphasis process for further decreasing the sound speed of the audio signal and further increasing the sound volume is performed for that region. Then, the audio signal is analogized and amplified to emit sound from the speaker unit SP1 (SP2).

  Therefore, in the voice communication system 2, the listener can easily understand the content that the speaker wants to convey. In addition, since the emphasis process can be applied to a voice signal transmitted from a voice communication device that does not include the voice processing device 42, the listener can surely hear the keyword and the speaker wants to convey it. The contents can be grasped reliably.

  Next, a specific configuration of the voice communication system 3 will be described. FIG. 6 is a block diagram showing a configuration different from FIGS. 3 and 5 of the voice communication system. The voice communication system 3 has a configuration in which a telephone set 16 and a telephone set 17 are connected to a network 10. FIGS. 2C and 6 show a configuration in which two telephones 16 and 17 are connected to the network 10, but this is for simplifying the description, and a plurality of telephones (voices) are also shown. It is possible to connect a communication apparatus) to the network 10 and exchange audio signals. Moreover, in the following description, the same code | symbol is attached | subjected to the structure similar to the audio | voice communication system 1, and detailed description is abbreviate | omitted.

  The telephone set 16 and the telephone set 17 are configured such that a voice processing device 42 is provided between the A / D converter 25 of the telephone set 11 and the communication unit 21. The telephone 16 (17) has a relatively low voice speed and a relatively high volume of the voice signal input from the microphone unit MIC1 (MIC2) and digitized by the A / D converter 25. The speech processing device 42 performs enhancement processing on the existing area. And the communication part 21 transmits to the telephone set 17 (16) via the network 10. FIG. Upon receiving the audio signal, the telephone set 17 (16) performs analogization and amplification, and emits audio from the speaker unit SP2 (SP1).

  Therefore, in the voice communication system 3, the listener can easily understand the content that the speaker wants to convey. Further, since the emphasized voice signal can be transmitted to the voice communication apparatus that does not include the voice processing apparatus 42, the speaker can reliably transmit the contents to be transmitted to the listener.

  FIG. 7 is an external view of the audio conference apparatus. FIG. 8 is a functional block diagram of the audio conference apparatus. FIG. 9 is a block diagram showing a case where an audio conference apparatus is applied as the audio communication apparatus of the audio communication system shown in FIG.

  Next, in the voice communication systems 1 to 3 shown in FIG. 2, an example in which a telephone is used as the voice communication apparatus has been shown. However, the present invention is not limited to this and may be another voice communication apparatus. . For example, instead of the telephones 11 and 12 shown in FIG. 2A, voice conference apparatuses 61 and 62 including a microphone array and a speaker array can be used as voice communication apparatuses. As shown in FIG. 7, the audio conference apparatuses 61 and 62 include a microphone array MA1 including microphone units MIC101 to MIC116 on a front side 72M of a casing 72, and microphone units MIC201 to MIC216 on a rear side 72U of the casing 72. A microphone array MA2 is provided. In addition, a speaker array SA1 including speaker units SP301 to SP312 is provided on the lower surface 72K of the casing 72.

  As shown in FIG. 8, the audio conference apparatuses 61 and 62 include an operation unit 74, a display unit 77, a control unit 110, an input / output connector 81, an input / output I / F (interface) 112, a sound emission directivity control unit 113, D / A converter 114 (114-1 to 114-12), sound emission amplifier 115 (115-1 to 115-12), speaker units SP301 to SP312 constituting speaker array SA1, and microphone arrays MA1 and MA2 are constituted. Microphone units MIC101 to MIC116 and microphone units MIC201 to MIC216, sound collecting amplifier 116 (116-1 to 116-32), A / D converter 117 (117-1 to 117-32), sound collecting beam generating unit 118, sound collecting A sound beam selection unit 119 and an echo cancellation unit 120 are provided.

  The control unit 110 of the audio conference apparatuses 61 and 62 converts the audio signal from the partner apparatus input from the input / output I / F 112 from a network format data to a general audio signal, and passes through the echo cancellation unit 120. While outputting to the sound emission directivity control part 113, the azimuth | direction data attached to the input audio | voice signal is acquired, and sound emission control with respect to the sound emission directivity control part 113 is performed.

  The sound emission directivity control unit 113 generates sound emission sound signals for the speaker units SP301 to SP312 according to the sound emission control contents. The sound output sound signal for the speaker units SP301 to SP312 is formed by performing signal control processing such as delay control and amplitude control on the input sound signal. The D / A converter 114 (114-1 to 114-12) converts the sound output sound signal in a digital format into an analog format, and the sound output amplifier 115 (115-1 to 115-12) amplifies the sound output sound signal. The speaker units SP301 to SP312 convert the sound emission sound signal into sound and emit the sound. Thereby, the voice of the conference person of the other party apparatus connected with the network is emitted to the conference person of the own apparatus.

  The microphone units MIC101 to MIC116 and the microphone units MIC201 to MIC216 collect ambient sounds including the utterances of the conference participants of their own devices and convert them into electrical signals, and generate collected audio signals.

  The collected sound beam generator 118 performs delay processing on the collected signals of the microphone units MIC101 to MIC116 and the microphone units MIC201 to MIC216, and generates the collected beam sound signals MB1 to MB8 having strong directivity in a predetermined direction. To do. The collected sound beam audio signals MB1 to MB8 are set so as to have strong directivity in different directions.

  As shown in FIG. 9, in the audio conference apparatus 61, MB1 is set to the direction Dir11, MB2 is set to the direction Dir12, MB3 is set to the direction Dir13, MB4 is set to the direction Dir14, MB5 is set to the direction Dir15, MB6 is set to the direction Dir16, and MB7 is set. In the direction Dir17, MB8 is set in the direction Dir18. As shown in FIG. 9, in the audio conference apparatus 62, MB1 is set to the direction Dir21, MB2 is set to the direction Dir22, MB3 is set to the direction Dir23, MB4 is set to the direction Dir24, MB5 is set to the direction Dir25, MB6 is set to the direction Dir26, MB7 is set in the direction Dir27, and MB8 is set in the direction Dir28.

  The collected sound beam selecting unit 119 compares the signal strengths of the collected sound beam sound signals MB1 to MB8, selects the collected sound beam sound signal having the highest intensity, and outputs it to the echo canceling unit 120 as the collected sound beam sound signal MB. To do. The sound collection beam selection unit 119 detects the direction Dir corresponding to the selected sound collection beam sound signal MB and supplies the detected direction to the control unit 110. The input / output I / F 112 converts the collected sound beam audio signal MB from the echo cancel unit 120 into an audio signal having a predetermined data length in a network format, and attaches the azimuth data and the collected sound time data obtained from the control unit 110. And output to the network 10.

  As described above, in the configuration shown in FIG. 9, the audio conference apparatus 61 (62) transmits the audio signal input from the microphone array MA <b> 1 or the microphone array MA <b> 2 to the server apparatus 13 via the network 10. Then, the enhancement processing is performed by the voice processing device 42 built in the server device 13, and the voice signal subjected to the enhancement processing is transmitted to the voice conference device 62 (61) via the network 10. When the voice conference device 62 (61) receives the voice signal subjected to the enhancement processing, the voice conference device 62 (61) emits the beamed voice from the speaker array SA1 to each speaker.

  Next, in the voice communication system 2 shown in FIGS. 2B and 5, the voice conference apparatuses 64 and 65 can be used as voice communication apparatuses instead of the telephones 14 and 15. However, in this case, like the telephones 14 and 15, it is necessary to provide the audio processing device 42 in the audio conference devices 64 and 65 so that the received audio signal is enhanced. Therefore, as shown in FIG. 8, in the audio conference apparatuses 64 and 65, the audio processing apparatus 42 is provided between the echo cancellation unit 120 and the sound emission directivity control unit 113.

  In this configuration, the audio conference apparatus 64 (65) transmits the audio signal input from the microphone array MA1 or the microphone array MA2 to the audio conference apparatus 65 (64) via the network 10. When an audio signal is received, enhancement processing is performed by the built-in audio processing device 42, and beamed audio is emitted from the speaker array SA1 to each speaker.

  Next, in the voice communication system 3 shown in FIG. 2C and FIG. 6, it is possible to use the voice conference apparatuses 66 and 67 as voice communication apparatuses instead of the telephones 16 and 17. However, in this case, similar to the telephones 16 and 17, it is necessary to provide the audio conference devices 66 and 67 with the audio processing device 42 so as to perform enhancement processing on the collected audio signals. is there. Therefore, as shown in FIG. 8, in the audio conference device 66 (67), the audio processing device 42 is provided between the collected sound beam selection unit 119 and the echo cancellation unit 120.

  In this configuration, the audio signal collected by the audio conference device 66 (67) by the microphone array MA1 or the microphone array MA2 is emphasized by the audio processing device 42 incorporated in the audio conference device 66 (67). The data is transmitted to the voice conference device 67 (66) via the network 10. When the voice conference device 67 (66) receives the voice signal, the voice conference device 67 (66) emits the beamed voice from the speaker array SA1 to each speaker.

  As described above, in the voice communication system of the present invention, since the voice processing device 42 is provided in the server device or the voice communication device (voice conference device), the keyword portion that the speaker wants to convey is extracted and emphasized. The listener can surely understand the content that the speaker wants to convey. In addition, the speech processing apparatus and speech communication system of the present invention have a configuration that does not require a keyword dictionary, and can extract keywords in any language without any problem, so no keyword registration work is required, and any country in the world Can be used in

  In the above description, the example in which the voice processing apparatus of the present invention is applied to a voice communication system has been described. However, the present invention is not limited to this, and a system or apparatus that exchanges voice signals to realize a conversation. Any other configuration may be used. For example, a small speaker unit and a small microphone unit are connected to the audio processing device, and an audio signal picked up by the microphone unit is emphasized, and the sound is emitted from the speaker unit. By configuring, it can be used as a hearing aid.

  In the above description, the speech processing apparatus is configured to analyze the speech speed and the volume as the acoustic feature amount in the speech signal. However, the present invention is not limited to this, and for example, the speech speed A configuration for analyzing the pitch (pitch) or a configuration for analyzing the volume and pitch (pitch) may be used to emphasize these feature amounts. It is also possible to analyze the three feature quantities of speech speed, volume, and pitch (pitch) and to emphasize these feature quantities. Further, it is possible to analyze three feature amounts of speech speed, volume, and pitch (pitch) and emphasize any one of these feature amounts.

It is an image figure for demonstrating the process which the audio | voice processing apparatus of this invention performs. It is a block diagram which shows the outline of an audio | voice communication system. It is a block diagram which shows the structure of an audio | voice communication system. It is a flowchart for demonstrating operation | movement of a server apparatus. It is a block diagram which shows the structure different from FIG. 3 of an audio | voice communication system. It is a block diagram which shows the structure different from FIG. It is an external view of an audio conference apparatus. It is a functional block diagram of an audio conference apparatus. It is a block diagram which shows the case where an audio conference apparatus is applied as an audio | voice communication apparatus of the audio | voice communication system shown to FIG. 2 (A).

Explanation of symbols

1, 2, 3-voice communication system 10-network 11J, 12J, 14J, 15J, 16J, 17J-receiver 11, 12, 14, 15, 16, 17-telephone (voice communication device) 13-server device 21-communication Unit 22-D / A converter 23-sound emission amplifier 24-sound pickup amplifier 25-A / D converter 26-operation unit 27-display unit 28, 43, 110-control unit 41-communication unit 42-voice processing device 51-Intonation analysis unit 52-Feature extraction unit 53-Enhancement region determination unit 54-Enhancement processing unit 61, 62, 64, 65, 66, 67-Audio conference device 72-Housing 74-Operation unit 77-Display unit 81- Input / output connector 112-Input / output I / F 113-Sound emission directivity control unit 114-D / A converter 115-Sound emission Amplifier 116-Sound collecting amplifier 117-A / D converter 118-Sound collecting beam generating unit 119-Sound collecting beam selecting unit 120-Echo canceling unit 511-Speech rate analyzing unit 512-Sound volume analyzing unit 521-Low speed region extracting unit 522 -Large volume extraction unit

Claims (6)

Analyzing the acoustic feature quantity in the input audio signal and extracting a region where the acoustic feature quantity is relatively larger or smaller than other areas;
The speech signal region extracted by the feature amount extraction means is subjected to an emphasis process in which the feature amount is further increased when the feature amount is relatively large, and is further decreased when the feature amount is relatively small. Emphasis means to output,
A voice processing apparatus. The feature amount extraction unit includes a low speed region extraction unit that analyzes a speech speed in a speech signal and extracts a region having a relatively low speech speed compared to other regions,
The speech processing apparatus according to claim 1, wherein the enhancement unit performs an enhancement process for further reducing a speech speed on the region extracted by the low-speed region extraction unit and outputs the region. The feature amount extraction unit includes a large volume region extraction unit that analyzes a volume of the audio signal and extracts a region having a relatively large volume compared to other regions,
The speech processing apparatus according to claim 1, wherein the enhancement unit performs an enhancement process for further increasing the volume of the region extracted by the high volume range extraction unit and outputs the region. 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
A microphone unit,
A speaker unit;
A communication means for transmitting the audio signal collected by the microphone unit to another audio communication device via the server device, and receiving the audio signal transmitted by the server device;
Speaker control means for inputting an audio signal received by the communication means to the speaker unit;
With
The server device
A voice processing device according to any one of claims 1 to 3;
Server communication that receives an audio signal transmitted from the audio communication device to another audio communication device, inputs the audio signal to the audio processing device, and transmits an audio signal output from the audio processing device to the other audio communication device Means,
A voice communication system. A voice communication system in which a plurality of voice communication devices are connected to each other,
Each voice communication device
A voice processing device according to any one of claims 1 to 3;
A microphone unit,
A speaker unit;
A communication means for transmitting the audio signal collected by the microphone unit to another audio communication device, receiving the audio signal transmitted by another audio communication device, and inputting the audio signal to the audio processing device;
Speaker control means for inputting an audio signal output from the audio processing device to the speaker unit;
A voice communication system. A voice communication system in which a plurality of voice communication devices are connected to each other,
Each voice communication device
A voice processing device according to any one of claims 1 to 3;
A microphone unit that picks up sound and inputs a sound signal to the sound processing device;
A speaker unit;
A communication means for transmitting the audio signal output by the audio processing device to another audio communication device and receiving the audio signal transmitted by another audio communication device;
Speaker control means for inputting an audio signal received by the communication means to the speaker unit;
A voice communication system.

Images