JPH04273748A - Sound recognition synthesis device and telephone system - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech recognition and synthesis device and a telephone device using the speech recognition and synthesis device.

0002

[Prior Art] Conventionally, in an office, when someone receives a phone call while away from their desk, a person sitting nearby picks up the phone on their behalf, announces the caller's absence, and inquires about the business. It is. By the way, there is an answering machine with a message sending function that allows you to notify the person of your absence and receive a message when you receive a call while you are away. This means that if you receive a call while you are away,
A message that has been registered (recorded) in advance will be sent. This message is, for example, something like ``My seat is now vacant. If you hear a beep, please let me know what you need.''. If necessary, the caller enters the message after the beep. Now, the above-mentioned message informing that the person is absent lacks information such as ``why is the person absent?'', ``where did he go?'' ``Will he return?'' and ``When will he return?'' For this, prepare (register) multiple messages, such as "I went out today and won't be returning...", "I'm in a meeting...", etc.
This can be compensated to some extent by selecting the appropriate options. However, there are too many combinations of specific messages such as "I'll be out until 3pm...", "I'm on a business trip until the 10th...", and "I'll be back in an hour...", so I can't register all of them. , it also consumes a large amount of recording media (memory, tape, etc.). If you re-register the message every time you leave your seat, you can include any information you want, but this puts a burden on the operator, as he or she must enter the entire message carefully and worded each time.

0003

[Purpose] The present invention was made in view of the above-mentioned circumstances.
, "1 hour. Meeting", "Until 10th. Business trip",
To make it possible to compose a proper message by simply saying something like "I'm not going back. I'm going out." Furthermore, to make the message sound audible to the other party in a pleasant voice (for example, a young woman's voice). This was done with the purpose of making a good impression on the other person by doing the same thing.

0004

[Structure] In order to achieve the above objects, the present invention provides (1)
A speech recognition unit that sequentially recognizes a plurality of voices, a word and sentence selection unit that selects words and sentences corresponding to the combination of the results of the plurality of voice recognition, and synthesizes speech based on the selected words and sentences. a voice synthesis unit, or
(2) A speech recognition unit that recognizes each of a plurality of sequentially input voices, and word and sentence selection that selects a plurality of words and sentences corresponding to the recognized plurality of voice recognition results from a predetermined dictionary. a selection result holding section that holds the selected results, a speech synthesis section, a first operation section, a second operation section, and the first operation section sets the speech synthesis section on output standby. a voice output control unit that outputs the selected words and sentences from the voice synthesis unit by a second operation unit; and (3) instructing the start of voice recognition in (2) above. The method includes an operation section, a speech synthesis section that automatically enters an output standby upon completion of speech recognition, and a speech output control section that outputs the selected words and sentences from the speech synthesis section in response to an external control signal. Further, (4) in (3) above, the words and sentences are a message indicating absence, and the signal from the outside is an incoming call from another telephone. Hereinafter, the present invention will be explained based on examples.

FIG. 1 shows the voice recognition absence setting and
This is a block diagram of an answering telephone. In the figure, 1 is a receiver, 2 is a communication circuit control section, 3 is a keyboard, 4 is a control section, 5 is a feature extraction section, 6 is a pattern matching section, and 7 is a standard pattern memory for recognition. , 8 is a text memory, 9 is a result holding unit, 10 is a standard pattern for synthesis, 11 is a speech synthesis unit, 11 is a telephone line, and when setting an absence, the operator first uses the keyboard 3
Press the absence setting key 3a inside. When this key is pressed, the control section 4 controls the pattern matching section 6 so as to recognize successive input voices. Microphone 1 of receiver 1
The voice inputted from a is turned into an input pattern by the feature extraction section 5, and is compared by the pattern matching section 6 with the standard pattern stored in the recognition standard pattern memory 7, and is stored in the sentence memory 8. In this way, the voices inputted from the microphone 1a are sequentially recognized, and the result of the combination of the recognized voices is held in the result holding section 9. For example, if "until 3 o'clock" and "going out" are sequentially recognized, the sentence is stored as "I will be out until 3 o'clock today."

[0006] To explain in more detail, for example, when it is recognized that ``until 3 o'clock'' has been input, the control unit 4 limits the type of sentence to be composed with the next input voice. For example, ■I'll be out until 3 o'clock today. ■I am on a business trip "until 3 o'clock". ■We are having a meeting "until 3 o'clock" today. Limit sentences such as Next, when it is recognized that "going out" has been entered, the text numbered ■ is selected, and when it is recognized that "business trip" has been entered, the text numbered "■" is selected, and "meeting" is recognized.
When is recognized, the sentence numbered ■ is selected. Similarly, “1
If you enter "until the 0th," then ■ "I will be away until the 10th." ■I will be on a business trip until the 10th. ■“1
We are holding a meeting until the 0th. etc., and confirm with the following input. Also, if you enter "I will not return", ■I will not return even if I go out. ■I'm on a business trip. ■I can't return to the meeting. Select and confirm with the next input.

[0007] After this, the telephone set 1 automatically enters a standby mode for answering an absent response. When a call comes in, connect the phone line,
The message, for example, "I'll be out until 3 o'clock today." is output from the speech synthesis section 11 and transmitted to the other party. In this way, the operator can create a detailed out-of-office message by saying simple words. As mentioned above, the vocabulary to be recognized may include time, reason for absence, whether to return to the seat, etc., and the response message may be appropriately selected by combining these words. The synthesized sound is not the operator's own voice,
By using a pleasant voice (for example, a young woman's voice), you can make a good impression on the other person. Note that in the above description, descriptions of parts other than those related to the present invention (telephone lines, etc.) have been omitted. It would also be good to use this invention with an answering machine function. The present invention is not limited to the speech recognition and synthesis algorithms described above, nor is there any limitation to the distinction between specific and unspecified speakers and the registration method.

[0008]

[Effects] As is clear from the above explanation, according to the present invention, when registering a message, you can simply say something like "I'm out until 3 o'clock today." You can compose a neat message such as "I'm looking forward to hearing from you.", and you can convey this neat message to the other party when you receive a call while you're away. Furthermore, by making the message sound audible to the other party in a pleasant voice (for example, a young woman's voice), it is possible to improve the impression of the other party.

[Brief explanation of the drawing]

FIG. 1 is a block diagram for explaining one embodiment of the present invention.

[Explanation of symbols]

1...Telephone, 2...Telephone line control unit, 3...Keyboard, 4
...Control unit, 5...Feature extraction unit, 6...Pattern matching unit, 7...
Standard pattern memory for recognition, 8...Text memory, 9...Result holding unit, 10...Standard pattern memory for synthesis, 11...Speech synthesis unit.

Claims (4)

[Claims] Claim 1: a speech recognition unit that sequentially recognizes a plurality of speech sounds; a word and text selection unit that selects words and sentences corresponding to a combination of the results of the plurality of speech recognition; 1. A speech recognition and synthesis device, comprising: a speech synthesis section that synthesizes speech based on the speech recognition unit. 2. A speech recognition unit that recognizes each of a plurality of sequentially input voices, and a word recognition unit that selects a plurality of words and sentences corresponding to the plurality of recognized speech recognition results from a predetermined dictionary. a sentence selection section, a selection result holding section holding the selected result, a speech synthesis section, and a first
an operation section, a second operation section, and a voice output device in which the first operation section causes the speech synthesis section to stand by for output, and the second operation section outputs the selected words and sentences from the speech synthesis section. 1. A speech recognition and synthesis device, comprising: a control section. 3. The speech recognition and synthesis device according to claim 2, further comprising: an operation section for instructing the start of speech recognition; a speech synthesis section that automatically enters output standby when speech recognition is finished; A speech recognition and synthesis device comprising: a speech output control section that outputs selected words and sentences from the speech synthesis section. 4. The speech recognition and synthesis apparatus according to claim 3, wherein the words and sentences are messages representing absence, and the external signal is an incoming call from another telephone. Telephone equipment with device.