JPH06188966A - Automatic answering telephone set - Google Patents

Abstract

PURPOSE:To allow the automatic answering telephone set to respond to a call from an acquainted caller without impoliteness as much as possible in a natural sense of message talking as if the called party were on the phone in spite of the automatic answering mode. CONSTITUTION:A reserved word of a name is registered in a reserved word storage circuit 13, voice information of a specific party is registered in a caller information storage circuit 14 and a relevant reply message is registered in a reply message storage circuit 11. The entered caller voice information and the registered reserved word or voice information are matched and a reply message corresponding to the coincident word is read and sent via a voice synthesis circuit 15. When non-voice state continues, a non-voice detection circuit 10 is operated to read and sent a specific reply message corresponding to the non-voice.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an answering machine which is configured to recognize a voice of a calling party and send a response message according to the calling party.

[0002]

2. Description of the Related Art A conventional general answering machine has automatically responded to an incoming call from a calling party, and then sends a certain fixed response message to any calling party without distinction. I was recording the caller's voice. From the called party's side, there was no room to select the calling party, and the answer was always the same. For example, "Yes, I'm out. I'm out of the office right now. If you have any message, please give me a message after a beeping sound."

Even if the called party wants to convey a specific message (for example, details of absent contents) to a specific calling party, he / she cannot convey the message to the specific calling party, The inconvenience that the message would be known to those who do not want to be sent was likely to occur. In addition, depending on the content of the message, there is a tendency for the caller to be treated unfamiliarly, or on the contrary, to be rude to the caller.
In addition, the message prerecorded by the called party is a one-sided talk to the calling party during the answering machine, which is different from a normal telephone conversation. There was a tendency that the calling party could not sufficiently speak (cannot record) the message that the calling party wanted to convey to the called party due to the tension.

An answering machine having a function of alleviating the above inconvenience has been proposed (see, for example, Japanese Patent Laid-Open No. 182942/1983). It recognizes the input voice content as words and selects a corresponding response message, and a response message prepared in advance to make a voice response to the content of the voice is converted into a voice signal and output. And a voice generation circuit section.

In the answering machine having the voice recognition function, voice recognition is performed when voice is input from the caller, and a corresponding response message is selected and output based on the result of the voice recognition. Is something that can be done.

[0006]

However, even if the voice content of the caller is recognized as a word, if the voice content of "Hello" is first issued by the caller,
There is no way to determine who the caller is. By all means
When an incoming call arrives, the answering machine will first send a message such as "Yes, it's ○○" or "○○. I'm out of the office ..." . After the above first message from the called party, if the calling party announces his / her name, the voice recognition control section of the answering machine will recognize it and output a response message according to the person. Control.

However, the first message is actually a problem.

The callee's message on the answering machine is a prerecorded form of the above-mentioned typical answer message. It is a so-called live voice (raw voice, real voice)
Not mechanical or electronic voice. The same applies to the first message sent when there is an incoming call. The caller thinks that the other party is there and is calling, but when such a mechanical or electronic voice first comes out after the call is received, the caller flicks for a moment. Not only those who have never experienced an answering machine, but also those with a little experience, are hesitant to start talking. They often hang up by hanging up the handset without speaking their business, or in some cases giving their name.
Even if you are used to answering machine, if the other party is absent, you will feel disappointed and will either hang up the phone as it is, or you will be reluctant to briefly say your message. Yes, it somehow becomes inorganic and unfriendly.

This is because the message of the other party (caller) is not taken into consideration in the psychology of the other party. It would be better if the caller gave his name, told him what he wanted, and recorded an answering machine. However, the biggest problem is that the caller hangs up knowing that the answering machine is an answering machine. The called party, who is the owner of the answering machine, does not know who is calling and is uncomfortable. In addition, I am worried that the caller may feel uncomfortable. As the caller, "I made a phone call ...
I feel like "...".

In the conventional answering machine having the voice recognition function, as described above, no consideration is given to the response at the moment when the caller is hesitant to notice that the calling party is the answering machine.

The present invention was devised in view of the above circumstances and, even though the answering machine is in the state of the answering machine, it is as if the called party exists at the other party's telephone port. An answering machine that allows you to speak in a natural manner by being answered and guided, and answering as much as possible without being rude to callers you know as callees The purpose is to provide a device.

[0012]

A first answering machine according to the present invention comprises an incoming call detecting means for detecting a ringing signal from a line, a loop control means for closing and opening a loop, and a reply message storage. A response message accumulating means, a voice recording means for recording the voice of the calling party, and a voice synthesizing means are provided, and a response message is automatically sent in response to an incoming call from the calling party to make a call. In an answering machine configured to record a person's voice, a reserved word accumulating means for accumulating a plurality of reserved words as words, and a response message accumulating means for accumulating a plurality of response messages respectively corresponding to the reserved words. A silence detecting means for determining whether or not the silence state continues for a predetermined time or more immediately after the incoming call is answered, and a voice recognizing means for identifying an unspecified word from voice information input from a line, Corresponding voice word detecting means for judging whether the uttered word of the caller identified by the voice recognizing means is stored in the reserved word storing means and extracting the corresponding reserved word, and the extracted reserved word. Response message or a specific response message corresponding to the silence when the silent state is detected, is read from the response message accumulating means, is sent to the line through the voice synthesizing means, and the voice of the caller is transmitted to the other party. It is characterized by comprising a main control means for causing the voice recording means to record.

Further, the second answering machine according to the present invention has an incoming call detecting means for detecting a ringing signal from a line, a loop control means for closing and opening a loop, and a reply message storage for storing a reply message. Means, a partner voice recording means for recording the voice of the calling party, and a voice synthesizing means, and automatically responds to an incoming call from the calling party and sends a response message to output the voice of the calling party. In an answering machine configured to record, caller information storage means for storing a plurality of voice information of specific persons and response message storage means for storing a plurality of response messages respectively corresponding to the voice information of each specific person. And a silence detecting means for determining whether or not the silent state continues for a predetermined time or more immediately after the incoming call response,
A voice recognition means for identifying an unspecified speaker from the voice information input from the line, and whether the voice information of the caller identified and specified by the voice recognition means is stored in the caller information storage means. Caller detection means for determining whether or not the corresponding caller voice information is extracted, and a response message corresponding to the extracted caller voice information, or a specific message corresponding to the silence when the silent state is detected. A main control means for reading a response message from the response message storage means, sending the response message to the line through the voice synthesis means, and recording the voice of the calling party in the voice recording means of the other party. Is.

[0014]

In the first aspect of the invention, a plurality of reserved words as words are stored in the reserved word storage means. In the second invention, a plurality of voice information of the specific person is stored in the caller information storage means. Then, the response message storage means stores a plurality of response messages respectively corresponding to the plurality of reserved words or the specific person voice information. The first standard response message (for example, "Yes, ○
Yes. )) Is sent. It is assumed that the caller responds to this initial response message in some way (for example, "Hello, I'm △△ ..."). In the first invention, it is judged whether or not the caller's uttered word is stored in the reserved word storage means based on the input voice information related to the caller's reaction, and the corresponding reserved word (for example, “Δ The word “△”) is extracted. In the second invention, it is judged whether or not the caller voice information is stored in the caller information storage means based on the input voice information, and the corresponding caller voice information (for example, "△△ Voice information) is extracted. And
The response message accumulating means reads out the response message corresponding to the extracted reserved word or the caller voice information and sends it to the line. This makes it possible to send an appropriate response message according to each calling party to the calling party without any discomfort. Even if the response message is not a so-called real voice, it is a message that corresponds to the caller or an interactive message, so it is a live message that is not a cleaved type. Then, while the caller is guided by the response message, the psychological resistance can be reduced, and at the same time, it is possible to request that the user speak comfortably and record the answering machine. Inorganic feeling can be reduced.

If the caller hesitates for a moment and the silence continues, the caller becomes tense by a specific response message corresponding to the silence (for example, "Which one?" Loosen and guide you to a state where you can talk easily. In other words, it is because the other party (calling party) is supposed to be in a silent state because he / she hesitates to speak, and sends a message in a timely manner considering the psychology of the other party. This makes it easier to avoid the situation where the caller suddenly hangs up. Even if the caller is recording an answering machine, the message prompts them to feel as if the called party is at the telephone and can start speaking naturally, making it relatively easy. You can say what you need to convey and finish your business. The called party will not have to miss the opportunity to know what is needed, and will have the peace of mind that they will not be rude or offensive to the other party.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of an answering machine according to the present invention will be described below in detail with reference to the drawings. FIG. 1 is a block diagram showing an electrical configuration of an answering machine according to an embodiment.

[0017] In FIG. 1, L _1, L ₂ is the station line, 1 incoming detecting circuit for detecting a call signal through the office line L _1, L _2, 2 are connected to the office line L _1, L ₂ Hook switch,
Reference numeral 3 is a loop control circuit connected in parallel to the hook switch 2 for closing and opening a loop. Reference numeral 4 is a call circuit connected to the hook switch 2 and loop control circuit 3. Reference numeral 5 is a call circuit 4. A CODEC connected via the talk terminal T ₁ and the receiving terminal R ₁ for encoding / decoding a voice signal / voice data, and 6 is a D connected to the CODEC 5.
SP (Digital Signal Processor).

The DSP 6 includes a voice recognition / analysis circuit 7 for identifying an unspecified word and an unspecified speaker from the voice information input from the line, and a reservation in which the uttered word of the identified caller is described later. The other party's voice word detection circuit 8 which judges whether or not the word is stored in the word storage circuit 13 and extracts the corresponding reserved word, and the caller information storage circuit 14 to which the voice information of the identified unspecified speaker is described later. A caller detection circuit 9 that determines whether or not the caller voice information is stored in the inside and a caller detection circuit 9 that determines whether or not the silence state continues for a predetermined time or more immediately after the incoming call is answered. Circuit 1
It is composed of 0 and 0.

Reference numeral 11 is a response message storage circuit for storing a response message, 12 is a partner voice recording circuit for recording the voice of the calling party, 13 is a plurality of reserved words such as "Tanaka" and "Ikeda" which are mainly words such as personal names. Reserved word storage circuit for storing, 14 is a caller information storage circuit for storing a plurality of voice information of a specific person such as "Tanaka is but", "Ikeda is but", etc. separately from words, 15 is a voice synthesis circuit, 16 Is a bus line, 17 is a main control circuit using a microcomputer for controlling the entire answering machine, and 18 is a bus line. A large number of reserved words are stored in the reserved word storage circuit 13, and a plurality of calling party voice information are stored in the calling party information storage circuit 14. These accumulations are performed by selecting the registration mode by the main control circuit 17.
Reserved words are stored in the reserved word storage circuit 13 via the voice synthesis circuit 15.
In addition, the caller voice information read from the other party voice recording circuit 12 is registered in the caller information storage circuit 14. The response message storage circuit 11 includes a plurality of response messages corresponding to each reserved word in the reserved word storage circuit 13 and a plurality of response messages corresponding to each caller voice information in the caller information storage circuit 14. A reply message, an initial reply message for answering an incoming call, and a reply message in the silent state are registered. The voice synthesizing circuit 15 is connected to the call circuit 4 via the transmitting terminal T ₂ .

Although the response message storage circuit 11, the partner voice recording circuit 12, the reserved word storage circuit 13 and the caller information storage circuit 14 are composed of a memory IC, the partner voice recording circuit 12 is a tape recorder. May be.

When the incoming call detection circuit 1 detects an incoming call, the main control circuit 17 reads out the first response message (for example, "Yes, it is XX") from the response message storage circuit 11 and calls from the voice synthesis circuit 15. It is sent to the calling party via the circuit 4 and the loop control circuit 3, and when there is a call from the calling party, the voice information is sent to the DSP.
6, the response message corresponding to the extracted reserved word or the caller voice information is read from the response message storage circuit 11 and sent from the voice synthesis circuit 15.
Furthermore, when the silent state continues for a predetermined time or more immediately after the incoming call response, a specific response message corresponding to the silent state is sent from the voice synthesizing circuit 15 to the caller. Then, if necessary, the voice recording circuit 1 of the other party is recorded with the voice of the caller sent through the line as a subject.
2 is configured to record.

Next, the operation of the answering machine having the above construction will be described.

When an incoming signal is sent from the caller side (telephone of the other party) through the exchange to the answering machine, the incoming call detection circuit 1 connected to the office lines L ₁ and L ₂ operates. The incoming signal is detected and the incoming detection signal is output to the main control circuit 17 via the bus line 18. The main control circuit 17 composed of a microcomputer closes the loop control circuit 3 after a lapse of a predetermined time from the input timing of the incoming call detection signal. As a result, the station line L ₁ ,
The L ₂ and the communication circuit 4 are connected.

Furthermore, after a lapse of a certain time, the voice synthesis circuit 15 is connected to the speech circuit 4 via the transmission terminal T _2, and the first fixed response message (for example, the first fixed response message registered in the response message storage circuit 11) (for example, After reading the code "Yes, this is Suzuki.") And synthesizing the voice, it is sent to the telephone line via the call circuit 4 and the loop control circuit 3.

After sending the first response message, the main control circuit 17 connects the CODEC 5 and the DSP 6 to the communication circuit 4 via the receiving terminal R ₁ . The caller receiving the first response message ("Yes, this is Suzuki.")
Immediately respond (for example, "Hello, Tanaka, but ...
… ”) Or, some kind of reaction occurs, such as staying silent for a while. The voice information of the caller that has entered the voice recognition / analysis circuit 7 of the DSP 6 via the call circuit 4 and the CODEC 5 is transferred to the voice recording circuit 12 of the other party and recorded, and at the same time, the voice information is recognized and analyzed. Is done.

In order to recognize and analyze the voice information of the calling party, it is necessary to cut out the voice information. As one method, it is assumed that the audio information is taken out every predetermined time starting from the start of recording. Further, the silent section is constantly monitored, and one word (or one phrase may be used) or silence is determined according to its length. In parallel with the indexing as a word, voice recognition as voice information itself is also performed. That is, the voice recognition analysis circuit 7 identifies the unspecified word from the voice information of the caller input from the call circuit 4 and also identifies the unspecified speaker.

The calling party uttered word (for example, "Tanaka") thus identified is sent to the partner speech word detection circuit 8. The partner voice word detection circuit 8 determines whether or not a reserved word corresponding to the caller's uttered word is registered in the reserved word storage circuit 13, and if there is a corresponding word, extracts the reserved word. It is sent to the main control circuit 17. The main control circuit 17 searches the response message storage circuit 11 based on the reserved word and reads the response message corresponding to the reserved word. On the other hand, the caller voice information identified by the voice recognition analysis circuit 7 (for example, "Tanaka is ...") is sent to the caller detection circuit 9. The caller detection circuit 9 determines whether or not the voice information corresponding to the caller voice information is registered in the caller information storage circuit 14, and if there is a corresponding one, the voice information is extracted. And sends it to the main control circuit 17. The main control circuit 17 searches the response message storage circuit 11 based on the voice information and reads the response message corresponding to the voice information. When the unspecified word is indexed and the unspecified speaker is indexed simultaneously, the response message storage circuit 11 by the main control circuit 17 is executed.
The search for is done with the highest priority. It is possible to freely decide which is to be prioritized, but normally, the one in which the calling party voice information is the same is prioritized. The read response message is transferred to the voice synthesizing circuit 15, where it is subjected to voice synthesizing and then sent from the call circuit 4 to the telephone line.

When the silence state continues for a predetermined time or more immediately after the response at the time of incoming call, the silence detection circuit 10 operates and sends a silence detection signal to the main control circuit 17. In this case, the main control circuit 17 reads the response message corresponding to the silent state from the response message storage circuit 11,
It is sent from the voice synthesis circuit 15 to the communication circuit 4.

The above-described recognition / analysis of caller voice information, extraction of the corresponding information, reading and sending of a response message are repeated a necessary number of times according to the reaction from the caller. When the voice input from the caller is interrupted for a predetermined time or longer, or when the scheduled sequence execution is completed by repeating the above, the main control circuit 17 opens the loop control circuit 3 and ends the message recording.

A specific example of sending a response message is shown in FIG. There is an incoming call at step n1, the loop is closed at step n2, and the first response message "Yes, this is Suzuki." Is sent at step n3. When the caller's response to this first response message is silent for more than the predetermined time, the process proceeds to step n4 and a message "Which one?" Is asked from the called party to the caller. send. In other words, the atmosphere should be such that the caller can be naturally prompted to answer the question. The situation in which the caller suddenly hangs up when he / she knows that the answering machine is an answering machine should be greatly reduced. If the silence continues, the message "Which one?" May be repeated again.

It is assumed that a caller whose tension has been reduced by this inquiry receives a reaction voice, for example, "I'm Sato ...". The DSP 6 performs matching of reserved words or caller voice information. That is, it is determined whether the reserved word "Sato" is registered in the reserved word storage circuit 13 and the caller voice information "Sato is" is registered in the caller information storage circuit 14. . If no match is found, it is judged that the call is from a party unknown to the called party, and the process proceeds to step n5, and a standard message such as "Please call again because you are not near now" is sent and step n1
Go to 1 to open the loop and end the call. Do not record. If there is a match, it is judged that the call is from a person who is familiar to the called party, and the process proceeds to step n6, and a message prompting an answering machine recording, such as "I am absent. Please have a message." To send.
Then, the recording is continued, and the recording is continued until the call is ended, the voice of the calling party is interrupted for a predetermined time or more, or a time-out occurs. In the figure, the rice mark (*) means the above-mentioned predetermined recording operation.

On the other hand, after sending the first response message,
It is assumed that there is a response from the caller within the stipulated time, and a voice such as "Tanaka is ..." or "Yamamoto is ..." or "Sawaguchi is ..." comes in. Similar to the above, it is checked whether or not there is a match, and if there is no match, the process of step n5 is performed. If matched, move to another process. Which response message is read depends on the matched content. For example,
If it is Tanaka ..., go to step n7,
The message "Thank you for taking care of me the other day" is sent out, and the process proceeds to step n6. If "I'm Yamamoto, but ...", step n
Proceed to step 8 and continue to the message "I'm shopping nearby. I'll call you back later." And then proceed to step n9 to send the message "Please give me a phone number." At this time, record (see *).
Also, when it is "I am Sawaguchi ...", proceed to step n10 and say, "I am at Mr. Yamada today.
Please contact ○○○○. Message is sent.
Do not record at this time.

As described above, it is possible to send an appropriate or necessary response message corresponding to each caller to the caller without any discomfort. Even if the response message is not a so-called real voice, it becomes a message that corresponds to the caller or an interactive message, so it becomes a familiar and live message. For the caller, it is less likely that the caller will feel at ease and will record an answering machine, or will feel an inorganic feeling even when not answering.

FIG. 2 illustrates a specific flow of message contents, and the technical operation will be described below with reference to the flowchart of FIG.

When an incoming call is detected, the process proceeds to step S1 to send the first response message. In step S2, the input caller voice information is recognized and analyzed. At this time, as a rule, recording will also be performed. In step S3, it is determined whether or not there is a voice input. If not, the process proceeds to step S4 to determine whether or not the silent state lasts for a predetermined time or longer. If the predetermined time has not been exceeded, the process returns to step S3, but if the silent state continues for more than the predetermined time, for example, the message "How is it?" Is repeated again, or the next step shown in step n5 is performed. After sending the message, the process proceeds to step n11 to open the loop.

If there is a voice input in the determination of step S3, the process proceeds from step S5 to step S6, and it is determined whether the time for word cutout has timed out. If the time-out has not yet occurred, the process returns to step S5 to determine whether or not the voice input is continued. If the voice input is continued, the process proceeds to step S6.
Proceed to 7. If the time is determined to have timed out in step S6, the process proceeds to step S7.

In step S7, the input caller voice information is compared with the registered voice information,
The process proceeds to a predetermined step according to which voice information matches, and the response message corresponding to each voice information is read and transmitted. If the caller voice information does not match any of the registered voice information, the process proceeds to step S8, the input voice word is compared with the registered reserved word, and it matches with any reserved word. Depending on whether or not it proceeds to a predetermined step, the response message corresponding to each reserved word is read out and transmitted. When the voice word does not match any of the registered reserved words, another standard response message (for example, "Please have a message") is read out and sent.

If only a part of the matching is found in the matching, a message "Please give me your name again." Is sent, or if it is judged that the input voice level is too low, "A slightly louder voice." Please send me again. " In addition, as an introduction when asking the caller's requirement, for example, a message such as "Hello" or "Who do you want to use?" Or "What is your need?" Furthermore, it is possible to consider whether or not the call is a prank call, and if so, a measure to automatically delete the recorded contents can be considered.

[0039]

As described above, in the first invention,
When the caller's uttered word is accumulated in the reserved word accumulating means based on the judgment based on the input voice information inputted from the caller, the corresponding reserved word is extracted, and in the second invention, When the caller voice information is stored in the caller information storage means based on the judgment based on the input voice information, the corresponding caller voice information is extracted, and the extracted reserved word or caller voice is extracted. Since the response message appropriately corresponding to the information is read out and sent to the calling party, an appropriate response message according to each calling party can be sent to the calling party without a feeling of discomfort, instead of the cut-out type. While guiding the caller and reducing psychological resistance, it is possible to easily talk about the subject and record the answering machine.

In particular, if the caller becomes hesitant for a moment and becomes silent, a specific response message corresponding to the silence is sent to ease the caller's tension and make the caller easy to talk to. Therefore, it is possible to avoid a situation where the caller suddenly hangs up. As a result, even if the answering machine is recorded, it is possible for the caller to spontaneously urge them to talk to them and tell them what to do.
Moreover, it is possible to have a sense of security that the person has not been rude or offended.

[Brief description of drawings]

FIG. 1 is a block diagram showing an electrical configuration of an answering machine according to an embodiment of the present invention.

FIG. 2 is a flowchart illustrating a specific flow of a response message in the example.

FIG. 3 is a flowchart for explaining the technical operation in the embodiment.

[Explanation of symbols]

1 ... incoming call detection circuit, 2 ... hook switch, 3 ... loop control circuit, 4 ... communication circuit, 5 ... CODEC, 6
...... DSP (Digital Signal Processor), 7 ……
Speech recognition analysis circuit, 8 ... Other party's speech word detection circuit, 9 ...
... Caller detection circuit, 10 ... Silence detection circuit, 11 ... Response message storage circuit, 12 ... Other party voice recording circuit, 1
3 ... Reserved word storage circuit, 14 ... Caller information storage circuit,
15 ... Voice synthesis circuit, 17 ... Main control circuit

Claims (2)

[Claims] 1. An incoming call detection means for detecting a ringing signal from a line, a loop control means for closing and opening a loop, a response message storage means for storing a response message, and a party for recording a voice of a calling party. In an answering machine configured to include a voice recording means and a voice synthesizing means, which is configured to automatically respond to an incoming call from a caller, send a response message, and record the voice of the caller, Reserved word storage means for storing a plurality of reserved words as words, response message storage means for storing a plurality of response messages respectively corresponding to each reserved word, and whether or not the silent state continues for a predetermined time or longer immediately after an incoming call is answered. Silence detecting means for determining whether or not, voice recognition means for identifying an unspecified word from voice information input from the line, and the voiced word of the caller identified by this voice recognition means is the reservation A partner voice word detecting means for judging whether or not the word is stored in the word storing means and extracting a corresponding reserved word, and a response message corresponding to the extracted reserved word or a silence when the silent state is detected. And a main control unit for reading out the specified response message from the response message storage unit, sending it to the line via the voice synthesis unit, and recording the voice of the calling party on the partner voice recording unit. Answering machine. 2. An incoming call detection means for detecting a ringing signal from a line, a loop control means for closing and opening a loop, a response message storage means for storing a response message, and a party for recording a voice of a caller. In an answering machine configured to include a voice recording means and a voice synthesizing means, which is configured to automatically respond to an incoming call from a caller, send a response message, and record the voice of the caller, Caller information accumulating means for accumulating a plurality of voice information of specific persons, response message accumulating means for accumulating a plurality of response messages corresponding to the voice information of each specific person, and a silent state for a predetermined time immediately after an incoming response. Silence detection means for determining whether or not to continue, voice recognition means for identifying an unspecified speaker from voice information input from the line, and caller identification and identification by the voice recognition means Caller detection means for determining whether or not voice information is stored in the caller information storage means and extracting the corresponding caller voice information, and a response message corresponding to the extracted caller voice information. Alternatively, when the silent state is detected, a specific response message corresponding to the silent state is read from the response message accumulating means and sent to the line via the voice synthesizing means, and the voice of the calling party is recorded by the voice recording means of the other party. An answering machine which is provided with a main control means for recording on the telephone.